BERL - July 2009 - Costs of Harmful Alcohol and Other Drug Use-1

Report to:
Ministry of Health and ACC
COSTS OF HARMFUL ALCOHOL AND OTHER DRUG USE
FINAL REPORT
Prepared by Adrian Slack Dr Ganesh Nana Michael Webster Fiona Stokes Jiani Wu
July 2009
Copyright BERL BERL ref #4577
ACKNOWLEDGEMENTS
BERL wishes to acknowledge the assistance of: ACC Peter Larking, Peter Roscoe, Agnes Guevara, and Wen Jhe Lee Department of Corrections Peter Johnston Health Outcomes International Ltd Jim Hales and Jane Manser Ministry of Health Susan Joy, Chris Laurenson, Fiona Julian, Chris Lewis (NZHIS) and Miranda Devlin (H&DIU) Ministry of Transport Wayne Jones New Zealand Police Jonathan Lyall, Rebecca Stevenson and Virginia Andersen St Johns Ambulance Andrew Cratchley Statistics New Zealand Lynne Mackie University of Otago Des ODea and Richard Edwards (Wellington School of Medicine), and Susan Dovey (Dunedin School of Medicine)
We are grateful to both the wider project team (Des, Richard and Susan) for comments received during the course of the project and external reviewers (Professor David Collins and Professor Helen Lapsley) on the final draft of the report. All suggestions were carefully considered and were incorporated as appropriate.
The views expressed in this report are not necessarily those of the New Zealand Ministry of Health or the Accident Compensation Corporation.
ii
Costs of Harmful Alcohol and Other Drug Use Ministry of Health and ACC July 2009
Costs of Harmful Alcohol and Other Drug Use

1 Executive Summary ......................................................................... 1 2 Introduction ...................................................................................... 6
2.1 Research scope ....................................................................................... 6 2.2 Report structure ..................................................................................... 13
3 Literature Review ........................................................................... 14

3.1 Methodological issues ........................................................................... 14 3.2 Analytical perspectives .......................................................................... 15 3.3 Cost categories ...................................................................................... 17
4 Method and Calculation Summary ............................................... 29

4.1 Method ................................................................................................... 29 4.2 Population patterns and impacts related to AOD use ........................... 30 4.3 Calculations of the costs of harmful drug use ....................................... 34
5 Results: the Costs of Harmful AOD Use ...................................... 56

5.1 Costs of harmful alcohol and other drug use overall .......................... 56 5.2 Costs of harmful alcohol use ................................................................. 61 5.3 Costs of harmful other drug use ............................................................ 64
6 Additional Analytical Focuses ...................................................... 67

6.1 Avoidable costs ..................................................................................... 67 6.2 Injury costs ............................................................................................ 70 6.3 Costs to the government ....................................................................... 76
7 References ..................................................................................... 78 8 Glossary ......................................................................................... 86 9 Appendix Method and Calculation Detail .................................. 88
9.1 Method ................................................................................................... 88 9.2 Methodological issues ........................................................................... 89 9.3 Population patterns and impacts related to drug use ............................ 91 9.4 Cost calculations ................................................................................... 98
10 Appendix Additional Tables ..................................................... 131 11 Appendix Sensitivity Analysis ................................................. 158
11.1 Harmful drug use and consumption decisions .................................... 172
12 Appendix Alternative GP cost estimates ................................ 174
iii
Tables
Table 2.1 Drinking pattern thresholds by gender, grams of alcohol per day ..................... 9 Table 4.1 Harmful drug users by sex and age group, 2005/06 ....................................... 31 Table 4.2 New Zealand (2005/06) and Australia (2004/05) alcohol use prevalence by drinking pattern ............................................................................................................... 31 Table 4.3 New Zealand (2006) and Australian (1998) illegal drug use prevalence by age group .............................................................................................................................. 32 Table 4.4 Average daily alcohol consumption by sex and drinking pattern (grams of alcohol per day), 2005/06 ............................................................................................... 33 Table 4.5 Police activity by offence category, 2005/06 ................................................... 42 Table 4.6 AOD-related apprehensions by offence category, 2005/06 ............................ 43 Table 4.7 AOD-related Police expenditure ($m) by offence category, 2005/06 .............. 44 Table 5.1 Social costs of harmful drug use ($m), 2005/06.............................................. 56 Table 5.2 Tangible costs of harmful drug use ($m), 2005/06 .......................................... 57 Table 5.3 Intangible costs of harmful drug use ($m), 2005/06 ........................................ 57 Table 5.4 Social costs of harmful alcohol use ($m), 2005/06 ......................................... 61 Table 5.5 Tangible costs of harmful alcohol use ($m), 2005/06 ..................................... 61 Table 5.6 Social costs of harmful other drug use ($m), 2005/06 .................................... 64 Table 5.7 Tangible costs of harmful other drug use ($m), 2005/06 ................................ 64 Table 6.1 Potential avoidable alcohol consumption and mortality in Australia, 2004/05 . 68 Table 6.2 Potential avoidable costs of harmful drug use in New Zealand, 2005/06 ........ 69 Table 6.3 Injury costs of harmful drug use ($m), 2005/06............................................... 71 Table 6.4 Total tangible and intangible costs of road crash injuries ($m), 2005/06 ........ 72 Table 6.5 ACC claim numbers and costs due to harmful drug use ($m), 2005/06 .......... 74 Table 6.6 Proportion of AOD injury and social costs borne by ACC, 2005/06 ................ 74 Table 6.7 Costs of harmful drug use government perspective ($m), 2005/06 ............. 76 Table 9.1 Alcohol consumption by drinking pattern, sex and age group, 2005/06 .......... 93 Table 9.2 Total alcohol and other drug use by sex and age group, 2005/06 .................. 93 Table 9.3 Harmful drug users by sex and age group, 2005/06 ....................................... 94 Table 9.4 New Zealand (2005/06) and Australian (2004/05) alcohol use prevalence by drinking risk .................................................................................................................... 95 Table 9.5 New Zealand (2006) and Australian (1998) illegal drug use prevalence by age group .............................................................................................................................. 96 Table 9.6 Workforce status of the additional population with no harmful AOD use, 2005/06......................................................................................................................... 100 Table 9.7 Workforce status of working-age harmful AOD users by drug type, 2005/06 100 Table 9.8 Police activity by offence category, 2005/06 ................................................. 108 Table 9.9 AOD-related apprehensions by offence category, 2005/06 .......................... 111 Table 9.10 AOD-related Police expenditure ($m) by offence category, 2005/06 .......... 111 Table 9.11 Comparison of NZ-ADAM and Alco-Link offence and apprehension rates with alcohol involvement by offence category, 2005/06 ....................................................... 113 Table 9.12 Case-weight multipliers, 1998/99 2007/08 ............................................... 120 Table 9.13 Estimated hospital costs of AOD-caused cases 2001 to 2006.................... 120
iv
Figures
Figure 3.1 Cost categories, components and analytical focuses .................................... 17
Appendix Tables
Appendix Table 1 Harmful drug use cost inclusions and exclusions ............................ 131 Appendix Table 2 Tangible costs of harmful drug use ($m), 2005/06 detail .............. 132 Appendix Table 3 Intangible costs of harmful drug use ($m), 2005/06 detail ............ 133 Appendix Table 4 Drug use prevalence 13+ year olds, 2005/06 .................................. 133 Appendix Table 5 Total alcohol caused deaths by nature of cause, 2001-2005 ........... 134 Appendix Table 6 Alcohol caused deaths and age-gender mortality rates, 2001-2005 135 Appendix Table 7 Total other drug caused deaths by nature of cause, 2001-2005 ...... 135 Appendix Table 8 Other drug deaths and age-gender mortality rates, 2001-2005 ....... 136 Appendix Table 9 Counterfactual population estimates males, 2005/06 ................... 137 Appendix Table 10 Counterfactual population estimates females, 2005/06 .............. 138 Appendix Table 11 Counterfactual population estimates total, 2005/06 .................... 139 Appendix Table 12 NZ-ADAM distribution of crime by offence category and drug type, 2005/06......................................................................................................................... 139 Appendix Table 13 Crime multipliers to estimate actual crime from recorded crime, 2003/04......................................................................................................................... 140 Appendix Table 14 NZP offence codes and offence categories ................................... 140 Appendix Table 15 NZ-ADAM offence categories (HOI)............................................... 141 Appendix Table 16 Lost output due to harmful drug use ($m), 2005/06 - detail ........... 142 Appendix Table 17 Justice sector costs of harmful drug use ($m), 2005/06 - detail ..... 143 Appendix Table 18 Health sector costs of harmful drug use ($m), 2005/06 - detail...... 144 Appendix Table 19 Hospital costs due to alcohol use by category, 2001-2006 ............ 145 Appendix Table 20 Hospital costs due to other drug use by category, 2001-2006 ....... 146 Appendix Table 21 Road crash costs due to harmful drug use ($m), 2005/06 - detail . 147 Appendix Table 22 Drug-attributable morbidity and mortality health conditions ........... 148 Appendix Table 23 Alcohol-attributable morbidity and mortality conditions .................. 154 Appendix Table 24 Cost to business of lost output ($m), 2005/06 ................................ 156 Appendix Table 25 Cost to government of lost output ($m), 2005/06 ........................... 156 Appendix Table 26 Costs to government of harmful drug use ($m) detail, 2005/06 .. 157 Appendix Table 27 Sensitivity analysis of general assumptions, 2005/06 .................... 159 Appendix Table 28 Sensitivity analysis of lost output assumptions, 2005/06 ............... 160 Appendix Table 29 Sensitivity analysis of drug production assumptions, 2005/06 ....... 162 Appendix Table 30 Sensitivity analysis of crime assumptions, 2005/06 ....................... 167 Appendix Table 31 Sensitivity analysis of health assumptions, 2005/06 ...................... 170 Appendix Table 32 Sensitivity analysis of intangible cost assumptions, 2005/06 ......... 171 Appendix Table 33 Sensitivity analysis of harmful AOD use assumptions, 2005/06 .... 173
Executive Summary
The New Zealand Ministry of Health and the Accident Compensation Corporation engaged BERL to estimate the social cost of harmful drug use in New Zealand. Harms related to drug use include a wide range of crime, lost output, health service use and other diverted resources. Drug harm may be avoided via interventions that interrupt supply, reduce demand or encourage safe drug use. The study analyses two categories of drugs: alcohol and other drugs, where other drugs include both illegal and misused legal drugs. It does not cover tobacco. The focus of the study is on the harmful effects of drug use, that is, use that results in a net social cost. This reflects that society, as a whole, has fewer resources and less welfare than in the absence of harmful use. Given this focus, the study covers a broad range of personal, economic, and social impacts, which we denote collectively as social costs. This report provides four broad answers. First, it estimates the total social costs from harmful drug use in the 2005/06 year. Second, it uses these estimates to characterise the potential level of costs that are avoidable. Third, it estimates the social cost stemming from injuries as a result of alcohol and other drug use. Fourth, it provides an estimate of the social costs from harmful drug use borne by the government. The study shows that harmful drug use imposed a substantial cost on New Zealand in 2005/06. Overall, harmful drug use in 2005/06 caused an estimated $6,525 million of social costs. This is equivalent to the GDP of New Zealands agricultural industry ($6,701 million) or finance industry ($6,982 million). The total was made up of $4,562 million of tangible resource costs and $1,963 million of intangible welfare costs.
1
Harmful alcohol use in 2005/06 cost New Zealand an estimated $4,437 million of diverted resources and lost welfare. To put this figure in perspective, the social cost across all cost categories was equivalent to almost two fifths of Vote Health in 2005/06; and the tangible costs alone to over one quarter of Vote Health.
2
GDP does not include intangible costs according to its definition in the system of national accounts (SNA). This may suggest that a comparison between social costs (that include intangible costs) and GDP may not be useful, as they have different conceptual bases. However, Easton (1997) argues that some form of benchmarking is useful for informed decision-making. It states that the magnitudes shed light on the enormity of the problem, and the significance of its various components. In this report, comparisons made with GDP figures are used as orders of magnitudes, to provide an indicator of size rather than a precise measurement of proportion of GDP.
2
The term Vote refers to funding approved by parliament for a specified range of outputs and that is the responsibility of a particular government Minister (the Vote Minister).
Harmful other drug use was estimated to cost $1,427 million, of which $1,034 million were tangible costs. The total is equivalent to over half of the justice sector Vote funding (Justice, Customs, Police, Courts, Corrections), and the tangible costs were equivalent to almost two fifths of the Vote funding.
Joint alcohol and other drug use that could not be separately allocated to one drug category cost a further $661 million. If the joint costs are split proportionately, total alcohol and total other drug costs equate to $4,939 million (over three quarters) and $1,585 million (just under one quarter).
Using estimates from international research, this study suggests that up to 50 percent ($3,260 million) of the social costs of harmful drug use may be avoidable.
The research indicated that 29.9 percent (or $1,951 million) of the social costs of harmful drug use result from injury. This equates to $2,900 per harmful drug user per annum.
The costs of harmful drug use from a government perspective amount to an estimated $1,602 million, or almost one third (35.1 percent) of the total tangible costs to society. Summary Table 1 Total social costs of harmful drug use ($m), 2005/06
($m) Tangible costs Intangible costs Total social costs % of social costs
Source: BERL
Alcohol 2,875.1 1,561.9 4,437.1 68.0%
Other drugs Joint AOD 1,034.2 392.4 1,426.7 21.9% 652.1 8.7 660.8 10.13%
Total 4,561.5 1,963.1 6,524.6 100.0%
Summary Figure 1 Tangible costs of harmful drug use by cost type ($m), 2005/06
Health 428.2 9%
Road 209.4 5%
Drug prod'n 861.0 19%

Crime 1111.4 24%
Lost output 1951.6 43%
Source: BERL
Tangible costs reflect productive resources diverted due to harmful drug use and totalled $4,562 million in 2005/06. This was equivalent to 2.9 percent of GDP in 2005/06.
3
Lost output ($1,952 million), crime costs not included in other components ($1,111 million) and drug production ($861.0 million) were the largest resource drains overall. Drugattributable health care and road crashes not included elsewhere cost a further $638 million. Drug users and victims suffered a further $1,963 million of intangible costs. The three largest tangible cost drivers for alcohol were labour costs, justice sector costs and drug production, which accounted for 84 percent of the tangible costs of alcohol. Similarly, drug production, crime and labour costs accounted for 92 percent of the tangible costs of other drug use. Given an estimated 513,000 harmful alcohol users, 27,000 other drug users and 127,000 joint alcohol and other drug users, harmful drug use cost approximately $9,800 per user, where over 70 percent of these impacts represented tangible resource costs. The research indicates that there is substantial scope to avoid costs resulting from harmful drug use via interventions that target supply and demand and that aim to reduce harmful use. The research did not specifically examine the cost-effectiveness of alcohol and other drug prevention and treatment interventions. Summary Table 2 Avoidable costs of harmful drug use ($m), 2005/06
Avoidable costs ($m) Tangible costs Intangible costs Total avoidable costs % of avoidable costs % of social costs
Source: BERL
Alcohol 1,440 780 2,220 70% 34%
Other drugs Joint AOD 520 200 710 20% 11% 330 0 330 10% 5%
Total 2,280 980 3,260 100% 50%
International studies suggest that potentially up to 50 percent of social costs can be avoided. Applying this proportion to this studys main estimates indicates that $3,260 million of these social costs of harmful alcohol and other drug use are avoidable. However, this figure should be viewed as providing an order of magnitude on potential avoidable costs, rather than an accurate estimate based on New Zealand evidence. At this stage, further analysis cannot be made, and this is an area where we recommend further research.
The estimates are GST exclusive figures. GDP, however, is measured at 'market prices', which includes indirect taxes such as GST. Therefore the estimates are not directly comparable with GDP. The percentage figure is indicative of magnitude, but is not precise.
Summary Table 3 Injury costs of harmful drug use ($m), 2005/06

Injury costs ($m) Tangible costs Intangible costs Total injury costs % of injury costs % of social costs
Source: BERL
Alcohol 624.4 967.5 1,591.9 81.6% 24.4%
Other drugs Joint AOD 82.7 254.3 337.1 17.3% 5.2% 17.0 0.5 17.5 0.9% 0.3%
Total 729.1 1,222.4 1,951.4 100.0% 29.9%
Total injury costs due to harmful alcohol and other drug use were estimated to be $1,951 million, or just over one quarter of the total social costs of harmful alcohol and other drug use. According to the injury cost analysis, tangible costs associated with harmful alcohol and other drug use totalled $729 million, and intangible costs was estimated to be around $1,222 million. Harmful alcohol use was estimated to cost our community approximately $1,592 million in total in 2005/06. This equates to injury costs of approximately $3,100 per harmful drinker, of which $1,200 is tangible and $1,900 is intangible. Other drug use accounted for a relatively smaller proportion of total injury costs (17.3 percent); it had an estimated impact of $337 million, the majority of which were intangible costs ($254 million in 2005/06). Approximately 27,000 people (aged 13 years old plus) were estimated to use illegal drugs and not alcohol, implying costs of $12,300 per drug user, of which $3,000 is tangible and $9,300 is intangible. The research also investigated the impacts of injuries resulting from joint alcohol and other drug use. The tangible costs of injury due to harmful joint alcohol and other drug use were estimated to be around $17.5 million. Almost all of these costs (97.0 percent) were estimated to be tangible costs associated with health care.
5 4
The costs of harmful drug use to the government amount to an estimated $1,602 million, or over one third (35.1 percent) of the total tangible costs to society. Reallocating the joint costs, just under 70 percent of the costs are due to alcohol and just over 30 percent are attributable to other drugs. Justice sector costs related to harmful drug use impose the largest burden on the government, accounting for just under half (49.4 percent) of the
Intangible costs make up a relatively larger proportion of injury costs than other costs. This is because many of the tangible costs were excluded as they were not injury related, or in some cases, for example the lost output for victims, it was impossible to separate out injury and non-injury related costs. This latter issue also means that the relativity between tangible cost components, such as lost output versus healthcare costs, will differ from the main estimates, partly as a result of the data issues rather than underlying behaviours and the consequent costs.
5
See footnote 4 for discussion on the implications of data availability for the relativity of tangible injury cost components.
estimated cost to the government. Over one third (38.2 percent) of the governments costs were due to lost tax revenue from reduced output, while a substantial 20.6 percent were borne by the health sector. Estimated costs in this report drew on a range of data and working assumptions. To give a measure of confidence in the robustness of results derived from these assumptions, sensitivity analyses were performed on several key factors. These sensitivity analyses suggest that the estimates of harmful alcohol and other drug use in 2005/06 are robust. On average, a one percent increase in the factors analysed leads to 0.1 percent increase in estimated costs (for positive changes) and a -0.09 percent reduction in estimated costs (for negative changes). The results are most sensitive to the assumptions about mortality rates; a 10 percent increase in mortality rates leads to a 3.3 percent increase in total social costs. A 10 percent increase in the proportion of the supply of illegal drugs that is imported reduces the total social costs by 1.0 percent. Other sensitive results include harmful alcohol and drug use prevalence, and the value of a statistical life year.
Introduction
The Ministry of Health and the Accident Compensation Commission (ACC) commissioned BERL to estimate the social costs of harmful alcohol and other drug (AOD) use in New Zealand. The Ministry of Health funds a substantial proportion of health care for New Zealanders and has a policy interest in the health and other impacts of harmful drug use. This research estimates the social costs of harmful drug use across a range of sectors, and includes a particular focus on these costs from a government perspective. ACC is a public insurer charged with providing cover for both workplace and non-workplace accidents. As part of this research, ACC asked for an analysis of the social costs of harmful drug use related injuries. In addition, we were able to access ACC data and provide a further analysis of the implications of harm ful drug use for ACCs expenditure. 2.1 Research scope
This report separately identifies, where possible, the social cost from harmful alcohol use and other drug use borne by the country in 2005/06. The other drug category in this study primarily covers illegal drugs including cannabis, opioids, stimulants, and hallucinogens. Where possible, the social costs of legal drug use were also included, such as the health care for harmful legal drug use (for example, legal anabolic steroid poisoning) and the cost of providing treatment using legal drugs (for example, methadone and naltrexone use for people receiving treatment for substance dependence). The study also carried out three sub-analyses of harmful drug use: avoidable costs, injury costs, and costs from a government perspective. Avoidable costs are the portion of total social costs that may be avoided by reducing harmful use via government intervention or changes in user behaviour. The second analysis focused on costs stemming from injuries as a result of AOD misuse. The third analysis uses the main estimates to determine the impacts of harmful AOD use on government expenditure and revenue. 2.1.1 Drugs: alcohol and others The study analyses two categories of psychoactive substances: alcohol and other drugs. The research specifically excludes consideration of tobacco.
6
The social costs of tobacco were recently updated in two pieces of work by Des ODea and co -researchers, ODea et al (2007a) and ODea (2007b).
The alcohol category includes both home and commercially produced alcohol.
The other drugs category refers to both illegal drugs and medicines or other legal products diverted from legitimate use to be used for their psychoactive effects . The study mainly found suitable evidence on the impacts of illegal drug use, but evidence on the use and impact of legal, but misused drugs, such as some party pills or solvents, tend to be limited. As such, the results are likely to underestimate the impact of misused, legal drugs. 2.1.2 Harmful use We define harmful AOD use as use that results in a net social cost. That is, society as a whole has fewer resources and less welfare than it would in the absence of harmful use. This approach focuses on observed negative impacts of rather than on the level of consumption. For example, we include the costs of road crashes where alcohol is a causal factor regardless of the drivers level of consumption. The World Health Organisation has a lexicon of terms related to harmful alcohol and drug use. Several of these terms are used in the articles, books, and journals reviewed and the media often uses these terms when discussing alcohol. These terms include, for example: substance misuse, substance abuse, addictive substance use alcohol use, alcohol misuse, hazardous drinking, heavy drinking, binge drinking, abnormal drinking behaviour illegal/non-medical drug use, harmful drug use, dependent/ habitual drug use.
9
Some authors reserve the term abuse for illegal substances, such as illegal drugs, while harmful use of legal drugs, such as alcohol, is called misuse. The phrasing of use, misuse and abuse is complicated by the possibility of beneficial substance use, particularly in the case of moderate alcohol consumption. This possibility is based on epidemiological studies
While all alcohol is considered in the characterisation of usage patterns and its harmful impacts, it was not possible to determine the share of overall harmful consumption that resulted from home-made alcohol. As such, the estimate of resources diverted by alcohol production is likely to be underestimated. According to the Ministry of Health's (2007) report "Alcohol use in New Zealand 2004", 1.8 percent of New Zealanders reported producing home-made alcohol in 2004. The proportion that home-made alcohol makes up of the total volume of alcohol consumed is likely to be smaller than the proportion of the population making home-make alcohol. Therefore, we believe that this omission will have a minimal impact on the studys results and will result in a conservative estimate.
8
That is, drugs classified as controlled drugs under the Misuse of Drugs Act 1975 and its subsequent amendments. However, although benzylpiperazine (BZP) was reclassified as a class C drug in April 2008 under the Misuse of Drugs (Classification of BZP) Amendment Act, there was insufficient information on its impacts to robustly include it in this study.
9
Epidemiological literature and previous drug abuse cost studies were used to determine thresholds for harmful AOD use (English et al 1995, Ezzati et al 2004, Rehm et al 2004, Connor et al 2005). This study defines harmful alcohol use as average daily consumption of alcohol per day over 20 grams for women and 40 grams for men. Any illegal drug use is classified as harmful for the range of impacts investigated in this study.
showing reduced risk of certain diseases, such as ischaemic heart disease among light to moderate drinkers.
10
So, while alcohol use may be benign or harmful in some cases, it is
possible that it may be beneficial in other cases. Aside from medical drug use, other drug consumption is routinely presented in the literature as misuse or having harmful impacts only. This tends to reflect the absence of evidence for the non-medical health benefits from the consumption of other drugs (Ridolfo and Stevenson 2001). Collins and Lapsley (2008), for example, have no problem in using the term abuse when referring to the consumption of illegal drugs. The authors argue, in the case of illegal drugs, by definition, society has decided to proscribe their consumption, with the implication that any consumption is abuse. This study uses the term harmful drug use instead of abuse. This term is: less judgmental than abuse recognises the complicated relationship between substance use and its impacts allows for the possibility that some use may be benign or beneficial.
This term is consistent with the requested focus for this project on drug abuse: where society, including the substance user, incurs extra costs as a result of the drug use. This study concentrates on the economic costs of harmful use. It does not explicitly estimate the social impacts of non-harmful use, nor the private costs associated with such use. Literature on beneficial use was not specifically reviewed, and estimated impacts of beneficial drug use were not included in this study.
11
However, there are intangible benefits,
for example, to consumers from non-harmful consumption of alcohol. As these impacts are benefits, however, they do not fall within the scope of this study on the social costs of harmful drug use.
12
10
The British Medical Association Board of Science (2008) argues that alcohol consumption is linked to long -term health and social consequences through three main causal pathways: intoxication, dependence and toxic (and beneficial) biological effects. WHO (2002) has also used this schema, but only with reference to harmful alcohol consumption. These arguments are reinforced by recent epidemiological work that argues that firm conclusions on potential health benefits of moderate alcohol consumption cannot be made on the evidence that is available (Lindberg and Amsterdam 2008, Fillmore et al 2007, 2006).
11
In the case of alcohol, there is substantial and on-going epidemiological debate about the existence and magnitude of any health benefits from any level of alcohol consumption. For example, Begg et al (2007) and Connor et al (2005) estimate some positive impacts of alcohol consumption for particular age groups and health conditions. But Lindberg and Amsterdam (2008), Fillmore et al (2007) and Fillmore et al (2006) contest the evidence base of the health benefits of alcohol, and suggest that it is not currently possible to conclude that alcohol is a causal factor for good health.
12
We use Collins and Lapsleys (2008) attributable fractions in our estimates of AOD-related hospital use and mortality rates. These fractions indicate some alcohol use may be beneficial but any other drug use is harmful. To
Our study does not assume that the legal status has any necessary implication for the harmfulness of the substance. For example, legality does not imply the substance is harmless nor does illegality imply it is harmful. The basis of whether use is harmful is determined on available epidemiological and other relevant evidence. This study defines harmful alcohol use as a hazardous or high risk drinking pattern (English et al 1995, Rehm et al 2004, Connor et al 2005).
13,14
Table 2.1 below reports the
drinking pattern thresholds, in grams per day, used in this study. They allow for different impacts by gender, and are based on population average levels. Table 2.1 Drinking pattern thresholds by gender, grams of alcohol per day
Grams of alcohol per day

Abstinent Low risk Hazardous High risk
Source: Connor et al (2005)
Women 0.0 0-19.99 20-39.99 40+
Men 0.0 0-39.99 40-59.99 60+
Low risk drinkers, such as social drinkers, are assumed to have no harmful alcohol use, unless specific information to the contrary was found. For example, the analysis includes harms resulting from road crashes, hospitalisations, workplace absenteeism or criminal offences involving low-risk drinkers, as these incidents are captured in the data sources used for this study. Any illegal drug use is assumed to be harmful. This reflects an absence of evidence for the non-medical health benefits from the consumption of illegal drugs (Ridolfo and Stevenson 2001). This approach is also consistent with the approach used in recent Australian social cost estimates (Collins and Lapsley 2002, 2008).
concentrate on harmful drug use, zero fractions were applied to conditions for which alcohol provided a net benefit, that is, for conditions with negative attributable fractions. This approach is likely to underestimate the harmful impacts of drug use. Although the net beneficial impact was removed, the harmful component for those conditions could not be estimated. However, Collins and Lapsley advise that the harmful impact for beneficial conditions is minute.
13
The average daily consumption ranges are consistent with the WHO categories (Rehm et al 2004), the Australian alcohol guidelines (NHRMC 1992), Australian epidemiological and substance abuse studies (English et al 1995, Pidd et al 2006, Collins and Lapsley et al 2008) and a recent New Zealand study on the burden of death, disease and disability due to alcohol (Connor et al 2005).
14
These patterns are notional concepts that are derived from aggregated population information and used in a wide variety of social cost estimation studies. However, these levels should not be interpreted as individual consumption guidelines. ALAC provides guidance and advice on individual alcohol consumption levels that are likely to minimise risk. http://www.alac.org.nz/LowRiskDrinking.aspx?PostingID=963
2.1.3 Costs of harmful drug use This study focuses on a broad range of costs covering personal, economic, and wider social impacts. These costs are collectively denoted by the term social costs in this report. This focus is consistent with that presented in Collins and Lapsley (2008). Collins and Lapsley gives a comprehensive economic definition of harmful drug use costs: The value of the net resources which in a given year are unavailable to the community for consumption or investment purposes as a result of the effects of past and present drug abuse, plus the intangible costs imposed by this abuse. Our definition assumes a counterfactual situation in which no harmful drug use has occurred. The range of costs included in this study is detailed in Appendix Table 1. The inclusions and exclusions are compared to the range of costs found in Collins and Lapsley (2008), BERL (2008a) and other drug misuse cost studies. The study aimed to estimate net social costs, rather than gross social costs of harmful drug use.
15
That is, drug use may offset some costs as users reduce the burden on societys
scarce resources. For example, while drug use may impose costs on the health system, premature death reduces the health care that users might otherwise have required if they had lived longer. Net costs are conceptually distinct from avoidable costs as they reflect impacts from consumption. Avoidable costs refer to the potential reduction in net costs due to effective policy or clinical interventions that reduce harmful substance use or minimise harm from substance use. The present study, however, does not extend to analysing the costeffectiveness of specific drug interventions. This study takes a conventional approach for economic cost studies, which do not attempt to fully consider the economic benefits of alcohol and other drugs, and should not be confused with cost-benefit or cost-effectiveness analyses (Single et al, 2003: 14).
15
A related, but separate, issue is that of the beneficial consequences of drug consumption. This report concentrates specifically on the social costs of harmful use. It does not analyse the impacts from non-harmful use, such as any protective health effects of alcohol consumption. That is, beneficial impacts of alcohol use are not included as cost offsets. Aside from medical drug use, illegal drug consumption is routinely presented in the literature as misuse or having harmful impacts only. This tends to reflect the absence of evidence for the nonmedical health benefits from the consumption of illegal drugs (Ridolfo and Stevenson 2001). Therefore, to be consistent with the focus on harmful drug use and the available evidence base, all illegal drug use is deemed harmful. A focus on the total impacts of drug use, rather than the harmful impacts, would allow for such beneficial impacts and result in lower net harmful effects.
10
2.1.4 The counterfactual The literature presents two conventional approaches to evaluate the costs of harmful substance use: prevalence and incidence. This cost study uses a prevalence approach as prevalence-based studies are considered useful for planning and budget decisions. Both prevalence and incidence approaches value a range of cost components that result from misuse and compare these costs to a hypothetical scenario. This scenario is known as the counterfactual. In this study, the counterfactual reflects the costs that would not occur if there was no past or present harmful drug use. The main difference between these approaches relates to whether the focus is on when the costs occur (prevalence) or when the use occurs (incidence). 2.1.5 Caveats on interpretation The study has a number of limits. Where possible, it uses New Zealand data and research. Where appropriate, we draw on Australian and American research to support our assumptions. Attributable fractions It was not feasible to re-calibrate attributable fractions using New Zealand prevalence statistics due to data constraints. In particular, it was not possible to get recent drug use prevalence statistics in a form that would suitably match up by drinking pattern with the relevant epidemiological literature on the relative risks associated with harmful drug use. Therefore, we use Collins and Lapsleys (2008) attributable fractions to estimate the proportions of mortalities and morbidities caused by AODs. This approach is likely to result in reasonably robust estimates given that the population and policy parameters in Australia and New Zealand are similar. However, New Zealand has a higher prevalence of harmful alcohol use than Australia, and a lower prevalence of other drug use. This means that this study is likely to underestimate the social costs of alcohol, while overestimating those for other drugs, where the estimated components draw on the Australian attributable fractions. Mortality rates We calculated AOD-related mortality rates for the 2001 to 2005 period using NZHIS data and attributable fractions. These rates are projected backwards to calculate AOD-related deaths from 1951 to 2006, and the likely survival of these people to 2005/06. That is, we estimate the number of people in the past who would have survived to live in 2005/06 if there was no harmful alcohol or drug use.
11
One assumption used in this estimation process is that the prevalence of drug use has remained constant over the estimation timeframe. The prevalence of hazardous drinking has remained reasonably constant over the past decade at a population level, although this experience may differ for specific groups within the population (Ministry of Health 2008). It is possible that the prevalence of harmful drinking in earlier periods was lower (Easton 1997). However, other drug use patterns and trends have changed over time (Wilkins and Sweetsur 2007). The mortality rates are based on data from the 2001 to 2005 period. But these rates are likely to overstate mortality from heroin and cocaine use in the 1950s or 1960s when use of these drugs in New Zealand was likely to be lower. Similarly, the data is not sufficient to include changes in recent trends, such as an increase in amphetamine use (and the related specific health conditions) over the last decade and a fall in cocaine use. Therefore our estimates for the additional population in the absence of harmful drug use are likely to be over-stated. Exclusions due to limited data In some cases, the local and international research was not sufficient to estimate some components. For example, this study does not provide an estimate of the intangible costs that result from poor health caused by harmful AOD use. This is likely to underestimate overall intangible costs. General equilibrium impacts Drug producers provide employment, income and output. This research does not, however, examine the general equilibrium (economy-wide) impacts of reducing harmful drug use. It is beyond the scope of this report to examine the alternatives to which these resources could be put. Such an evaluation would require industry (microeconomic) and countrywide (macroeconomic) analyses of the relative productivity of these resources in their current and alternative uses.
16
In the case of alcohol production, this analysis would be complicated by
impacts on the cost of production, i.e. economies of scale, from substantial reductions in the size of the industry.
16
The size of the illegal drug industry poses a social and economic policy issue in New Zealand, particularly as some drug production has a strong regional concentration. Drug laws and enforcement lead to high prices for illegal drugs. While these prices may encourage some drug production, the prices give a distorted signal about the social value of that activity. That is, drug production may be profitable but producers fail to (fully) account for the harmful impact of their output. The harmful impacts that drug use imposes on that and other regions should be set against drug profits. Furthermore, the drug industry may trap resources and stop them from moving to better alternatives, such as innovation or education. Reforming the drug industry and moving resources to other industries may cut income in the short term but strengthen a regions economic base and deliver higher, sustainable growth in the future.
12
2.2
Report structure
The remainder of the report is divided into three main parts. Sections 3 and 4 provide background to the study. Section 3 provides a summary of a brief literature review completed for this project.
17
Section 4 briefly sets out the broad methods
and definitions BERL used in this study; appendix 8 provides more detail on our methods and calculations. Section 5 presents the studys main results. It analyses the overall impacts of harmful AOD use, including estimated costs for alcohol, other drugs and joint AOD costs, which are costs that could not be attributed to a specific drug given the available data. Separate subsections examine those costs identified as relating specifically to alcohol or to other drugs. Section 6 reports three sub-analyses of harmful drug use: avoidable costs, injury costs and costs from a government perspective. Avoidable costs are the portion of total social costs that may be avoided by reducing harmful use via government intervention or changes in user behaviour. The second analysis focused on what part of total social costs results from AODrelated injuries. The third analysis estimates the costs to government of harmful AOD use. The report ends with literature references and a glossary (sections 7 and 8). Additional tables and materials are appended in sections 9 through 12.
17
The literature review is available in a separate report, BERL (2008) AOD cost analysis literature review.
13
Literature Review
18
This section summarises a literature review carried out as an initial step in this project.
The literature review surveys major international and New Zealand literature on harmful substance use cost estimation and drug-related research. It explicitly excludes research on the costs of tobacco. It initially informed key methodological decisions and later provided a context to interpret the studys results. As such the literature review is organised around particular decisions that have been made in the course of the research, works from the general to the particular, and considers various analyses that could build on the main cost estimates. 3.1 Methodological issues
The cost of harmful substance use literature presents two conventional approaches to evaluate these costs: prevalence and incidence approaches (BERL 2008, Collins and Lapsley 2008, Johansson et al 2006, Rehm et al 2006, UK Cabinet Office 2003, Catalyst Health Economics 2001, ONDCP 2001, Devlin 1997, Easton 1997).
19
International
guidelines for harmful substance use cost estimation recognise both approaches as legitimate (Single et al 2003), but they address different research questions. Both approaches value a range of misuse cost components and compare these costs to a hypothetical or counterfactual scenario where there is no misuse. The main difference relates to whether the focus is on estimating costs of impacts occurring in the current period, which are attributable to past and current drug use (prevalence), or of estimating the current and future impacts of current drug use (incidence). Single et al (2003) discusses prevalence studies as those that estimate the number of deaths and hospitalisations attributable to harmful substance use in a given year, and the costs associated with these deaths or hospitalisations. These costs also take into account harmful substance use prevention and intervention, and law enforcement costs in the same period. The study argues that prevalence is commonly used to refer to the number of cases of a particular disease or disorder that occurs in the general population during a specified period. Using a prevalence approach is therefore useful to estimate the social costs of
18 19
The summary is based on a separate literature review report, BERL (2008)
The terms incidence and prevalence may have different interpretations in other contexts. For example, in the justice sector, the New Zealand Crime and Safety Survey (Mayhew and Reilly 2007a) notes, "Risks can be measured in terms of incidence rates the number of offences per 100 households or adults. Incidence rates are used to estimate the full volume of crime taking into account that some people are victimised more than once. Risks can also be measured in terms of prevalence rates, which show the percentage of households, or adults who have been victimised once or more.
14
alcohol misuse as you are able to estimate the costs based on the number of cases of a particular disease or disorder that has occurred in a given year. The alternative incidence-based approach aims to estimate current and future costs resulting from harmful substance use by people in a given year. For example, the value of lost output from a person who dies prematurely as a result of harmful substance use is based on their estimated lifetime earnings rather than the single year of output lost in the year of their death. This requires projections into the future and the use of discount rates to establish the potential loss in output incurred due to morbidity and mortality resulting from harmful use. This approach is more complicated than the prevalence approach, as it involves estimating a lifetime profile of earnings (or other impacts) and choosing an appropriate valuation method so that present and future costs are measured in commensurate terms. A prevalence approach focuses on the impacts due to current and past AOD use that occur in a given year. Prevalence-based studies are useful for planning and budget decisions. It is for these reasons that this study uses a prevalence approach. 3.2 Analytical perspectives
The primary analytical perspective of this study is the social costs of harmful AOD use; this is the most common perspective used in the major harmful substance use cost studies. The analytical perspective of a study determines what costs are relevant and may be captured in the analysis. Possible viewpoints in the general health economics literature include private, business, government and social (Drummond et al 1994). Single et al (2003) makes a distinction between private costs and costs borne by others such as businesses or governments. The report argues a key distinction is that private decision costs are knowingly borne by an individual, while social costs are not knowingly or freely borne by the user or are borne by others such as businesses or government. Accurately valuing private benefits is likely to be complicated by the role of addiction, information issues (Collins and Lapsley 2008, Easton 1997) and the hidden nature of illegal drug use. Social costs are defined in various ways, but a common thread is a version of the economic idea of negative externalities. Conventionally, an externality is an impact positive or negative that is borne by a third party and for which there is no compensation, for example, alcohol-related crime. Social costs are the costs imposed on New Zealand society by harmful drug use, excluding purely private impacts. Markandya and Pearce (1989) extends this idea to include third-party costs plus costs unknowingly borne by the user, for example, where the actual cost is greater than the
15
perceived cost. Under this definition, poorly informed decisions generate social costs even though these costs are borne by the user. Such costs may be referred to as internal costs in economic terms (Single et al 2003). Single et al (2003) argues that from the point of view of public policy, social costs are the most relevant as they determine the costs that an activity such as drug misuse imposes on users and the rest of a community. Further, the study states: Social costs may be incurred by other persons in the private sector (e.g. when private insurance premiums are increased due to payouts to smokers) as well as by public sector expenditure. Thus, in the context of COI [cost of injury] studies, social is not a synonym for public, nor private for private sector. Thus the focus on social costs often appears to be driven by impacts that are relevant to policy or emerging policy issues. The Strategy Unit of the UK Cabinet Office (2003) argues that estimating the cost of alcohol misuse is a valuable source of information for policy makers. It serves a variety of functions such as justifying, or otherwise, resources spent on reducing the harm associated with alcohol misuse; appropriately targeting specific problems; providing insight into future policy appraisal and evaluations; providing baseline measures to determine the efficiency of alcohol policies and programmes; and helping to identify information gaps, research needs and desirable refinements to national statistical reporting systems. As social costs are the costs imposed on society, they are often relevant to policy or emerging policy issues. As such, harmful substance use studies concentrate on those impacts that may justify government intervention, such as poorly informed decisions or decisions that harm others. Eastons (1997) study of the costs of alcohol and tobacco use in New Zealand is an exception to the focus on social costs with the exclusion of known private costs and benefits of substance use. That study incorporates an estimate of the private benefit that nonaddicted drinkers derive from alcohol consumption. It is included as a negative item in the counterfactual, that is, a benefit that would be lost if alcohol misuse stopped. A final note on analytical perspective in the literature concerns the social value placed on resources freed by reduced drug misuse. This issue is complicated in the literature by distributional issues and economy-wide (or general equilibrium) impacts. For example, eliminating cannabis production in New Zealand would be likely to impact on some people and regions more than others. However, harmful substance use studies tend to focus on quantifying the magnitude of harms, rather than the economic impact of reducing or
16
eliminating harmful substance use. Further, Collins and Lapsley (2008) argues that it would be speculative to estimate the macroeconomic impact of reduced drug misuse as the uses to which these resources would have been put would be largely determined by government macroeconomic and microeconomic policies. 3.3 Cost categories
The broad categorisation of costs below is common across the major harmful substance use cost studies, and Single et al (2001, 2003) acts as a landmark reference for it. The two broad cost categories are tangible (or productive resource) costs and intangible (welfare) costs. These include crime, inputs diverted to drug production, health care costs, road crash costs, lost output (which is sometimes referred to as lost productivity) and selected dimensions of quality of life and loss of life. Despite the wide range of costs included in the major harmful substance use cost studies, there are a number of components for which estimates are not provided due to a lack of data about the distribution of risk factors or the association between a risk factor and an outcome. These include costs such as environmental damage or a broad concept of lost wellbeing. Figure 3.1 Cost categories, components and analytical focuses
Total social costs of substance misuse
Cost categories Tangible costs Cost components Intangible costs
Crime
Diverted inputs
Health
Road crashes
Lost output
Morbidity
Mortality
The colours suggest possible alternative analytical focuses, for example, a focus on total costs (light blue plus purple) or on avoidable costs (purple only). Avoidable costs are the portion of total costs that may be avoided by reduced harmful substance use through treatment and preventive interventions. The remaining costs, shown in light blue, represent costs that are likely to persist in spite of policy interventions. Other focuses might concentrate on who bears the costs, or costs due to particular types of harms such as injury. Accurate categorisation of the costs of harmful substance use also needs to take into account two issues. The first relates to co-morbidity costs, that is, costs which are not primarily caused by the misuse of a substance but may be exacerbated by it. The second issue relates to jointly attributable costs, costs that are associated with the use of multiple substances.
17
3.3.1 Tangible costs A tangible cost can be either a direct cost or an indirect cost, depending on whether it is an explicit cost or an opportunity lost due to harmful substance use. The largest direct costs due to the misuse of alcohol or other drugs in dollar terms (Collins and Lapsley 2008, BERL 2008) include:
20
crime costs caused by harmful drug use resources diverted from beneficial consumption or investment to drug production road crashes health care costs.
Conceptually, the direct cost category also includes the unpaid time given up by family and friends to take care of those who are ill as a result of harmful drug use, as well as time spent seeking or participating in treatment by persons affected by harmful drug use. Estimation of these costs would require information on the quantum and value of time involved. This study does not estimate this cost. While the literature is not explicit on this point, some direct cost components may be more relevant for some substances than others. For example, property damage due to fires may be a greater issue for alcohol misusers than injecting drug users. Indirect costs refer to potential resources or output that is not generated as a result of misuse. These costs may be borne by individuals or third parties such as employers. The primary indirect costs of AOD are: production lost to the economy as a result of premature death of users of AODs production lost to the economy as a result of an injury or disability to users of AODs reduced production by those who are disabled, for example, due to the ill effects of harmful drug use or AOD-related road crashes reduced production by family members and friends who take care of those who are ill as a result of harmful drug use. However, estimation of these costs would require information on the quantum and value of time involved. This study does not estimate this cost.
20
These components may also involve indirect and intangible costs. For example, in addition to health care related to road crashes, time off work would be counted as an indirect cost, while lost quality of life or loss of life would be measured as an intangible cost.
18
3.3.2 Intangible costs Intangible costs are welfare impacts borne by individuals that cannot be shifted (Collins and Lapsley 2008). Intangible costs harm the individual but any reduction in harm cannot be transferred to other members of society. In the case of AOD, intangible costs include: premature death among users as a result of AOD misuse reductions in the quality of life among users due to pain, disability and lost wellbeing caused by AOD misuse. Most AOD misuse studies estimating intangible costs analyse loss of life and assign a monetary value to it. A UK Cabinet study (2003) argues that a monetary value of premature death can be estimated by the Willingness to Pay (WTP) approach. In this approach life is valued according to what individuals would be willing to pay for a change that reduces the probability of illness or death. Single et al (2003) also discusses how studies should consider the loss of income and quality of life due to premature mortality. One issue is how to account for the age at which death occurs and the impacts of this on lost output and the quantum of life years lost. The age at death affects both the persons potential level of output, the remaining number of productive years and the number of life years sacrificed to premature mortality. Evaluating future losses is primarily an issue for cost studies that use an incidence approach or a human capital approach to valuing lost output. Those approaches aim to capture the current and future impacts, which are affected by the age of death and consequently the number of future years affected. In contrast, a prevalence approach focuses on costs borne in a particular year as a result of past and current behaviours. This approach avoids having to estimate the implications of premature mortality that extend out to the future. Summary measures of the lifelong impact of disease on loss of wellbeing include years of life lost (YLLs), years of life lost due to disability (YLD), disability-adjusted life years (DALYs), or quality-adjusted life years (QALYs). These measures are briefly defined below. YLLs relate to what reduction in mortality is possible, and measure th e amount of time a person would have lived had he or she not died prematurely the (potential) years of life lost (Ministry of Health 1999). The YLL method implicitly values death at a younger age greater than death at an older age, and YLLs may be calculated to allow for different assumptions about potential life expectancy. However, differential life expectancies, based on the populations current life expectancies, raise equity concerns due to existing differences between Mori and non-Mori life expectancy.
19
YLDs are the number of years a person has lived with a health condition multiplied by the disability weight for that condition. Disability weights are based on expert opinion rather than individual preference-based estimates.
DALYs measure disability or lost health in the current population compared to a counterfactual situation where everyone lives to old age in full health. DALYS can be calculated for specific health conditions and are the sum of YLLs and YLDs.
QALYs are also a summary measure of health-related morbidity and mortality, but a higher QALY measure represents greater wellbeing. QALYs value different health states based on five health dimensions (mobility, self-care, usual activities, pain/discomfort, and anxiety/depression). The conventional health state values used are based on preference-based evaluations developed by the EuroQol Group (Brooks 1996). Some New Zealand research has looked at developing culturally-specific health state preferences (Devlin et al 2000).
A further step is to convert the total intangible cost measured in natural units (e.g. YLLs or QALYS) to dollar terms using an appropriate value statistic, such as a value of statistical life (VOSL).
21
Collins and Lapsley (2008) argues that pain and disability attributable to road accidents can also be given a monetary value. But neither that study nor its earlier editions find sufficient data to estimate other morbidity costs. Easton (1997) argues that intangible morbidity costs in New Zealand were of a similar magnitude to mortality costs. However, he notes that it is not easy to summarise [alcohol-related loss of wellbeing] in dollar terms. This is because there is sparse information about the incidence of morbidity from alcohol misuse and there is no measured (let alone agreed) valuation of the reduced quality of life. However, more recent transport sector research in New Zealand has generated a WTP-based VOSL, which is adapted for use in our study. Even though well established methodologies exist for this process, there is controversy about what an appropriate VOSL is (BERL 2007b). Issues around VOSLs include whether: a persons relative productivity is significant, for example, placing a higher value on more productive individuals an allowance for a persons age, for example, whether the deaths of young people and old people should be attributed the same value
21
For more information, refer to section 9.4.6 on intangible costs and footnote 133.
20
the mode of death is influential, for example, whether death by road traffic accident is psychologically more or less traumatic than death in a house fire.
3.3.3 Co-morbidities and jointly attributable costs Two further methodological issues involve how to account for co-morbidities and impacts that are jointly attributable to multiple substance use. Co-morbidity costs are not primarily caused by the misuse of a substance but may be exacerbated by it. In addition, co-morbidities may be exacerbated by joint AOD use when the substances interact. Where these costs are ignored, they will lead to an underestimate. Jointly attributable costs are the product of multiple substance use. Where this joint attribution is ignored, it could lead to an overestimate due to double counting of a single cost when single substance estimates are inappropriately summed. The literature concentrates on health-related co-morbidities, for example, where alcohol affects the general condition of a patient whose main diagnosis is non-alcohol related. In this case, alcohol use may worsen pre-existing conditions and increase the cost of health care (Johansson et al 2006). Johansson et al calculates co-morbidity using an extended version of that used by Single et al (1998). In this method, Johansson et al compares patients with alcohol-related secondary diagnosis with patients with no alcohol-related secondary diagnoses. The co-morbidity cost is then estimated by taking the average difference in length of stay times the number of cases with an alcohol-related co-morbidity. Ezzati et al (2004) canvasses the epidemiological issues and research on estimating the joint effects of two or more risk factors and the calculation of joint population attributable fractions (PAFs).
22
A joint PAF gives the fraction of a disease collectively due to the factors
examined. Following a technical exposition on the calculation of a joint PAF, Ezzati et al briefly surveys examples of joint effects for selected major risk factors, including smoking but not AOD. English et al (1995) provides an accessible examination of the issues with joint attribution in an appendix and sets out how its estimates of drug-caused attributable fractions account for
22
An attributable fraction measures the proportion of a disease or mortality in a population due to exposure to a defined environmental risk, such as drinking alcohol (WHO 2002). A negative fraction indicates that the drug in question has a protective effect against the medical condition under study (Collins and Lapsley 2008).
21
joint attribution. The study reviews the ideas of sufficient, component and necessary causes.
23
It notes (p 603),
while hazardous/harmful alcohol use could well be a component cause of some deaths and morbid events ascribed to opiate and other illegal drug poisoning, such deaths would have been ascribed only to illegal drugs. Thus, the question of double counting does not arise, although such events represent an underestimation of mortality and morbidity caused by hazardous/harmful alcohol intake. The report specifies 10 conditions for which sufficient evidence of causation was found, and which are common to at least two out of the three broad categories of drugs (alcohol, tobacco, illegal drugs). Only two of these 10 relate to potential interaction effects between alcohol and illegal drugs: low birth weight and suicide. However, there are more than 10 conditions for which attributable fractions have been derived, and many of these conditions could also have interaction effects. Begg et al (2007) is based on the earlier work of English et al (1995) and Ridolfo and Stevenson (2001) and estimates joint effects for key health risks. It draws on the WHOs Comparative Risk Assessment (CRA) project on health risk exposure prevalence and the relationship with health outcomes.
24
Begg et als introduction to joint risk attribution makes
an over-simplification that health risks are biologically independent and uncorrelated. However, its quantitative analysis and calculation of joint PAFs does attempt to take into account risk factor interactions. As for English et al (1995), the main conditions where joint attribution is an issue are suicide/self-harm and a set of minor conditions grouped into a broad Other categor y. However, it does not report joint PAFs for specific conditions which are jointly related to alcohol and illegal drugs. Collins and Lapsley (2008), the recent Australian substance abuse cost study, recognises the risk of double counting due to causal interactions between drugs. To account for this possible effect, its aggregate estimates of these types of costs are reduced by 2.18 percent. Its estimates appear to be based on English et als (1995) calculations and the authors analysis of drug-related Australian mortality in 2004/05.
25
As noted in the authors earlier
23
A sufficient cause inevitably produces the effect; a component cause does not cause the effect alone but may become a sufficient cause in combination with other factors; a necessary cause is a component cause that is a member of every sufficient cause (English et al 1995).
24
The CRA used international panels of experts to collect the up-to-date information for a range of countries and health risks on the prevalence of exposure to the selected health risks and the relationship between these exposures and health outcomes. The results were published in the WHOs (2002) annual world health report and by Ezzati et al (2007).
25
The authors are aware of Begg et als (2007) work. But as Begg et al (2007) did not report joint PAFs for specific substances and conditions, Collins and Lapsley may not have been able to use that research in their latest analysis.
22
study (Collins and Lapsley 2002), the quantified interactions are between alcohol and tobacco, not between alcohol and illegal drugs. For non-health crime and justice sector costs, Collins and Lapsley (2008) provides a combined estimate when the study is unable to sensibly allocate joint costs to a single substance. For example, some crime costs are jointly attributed to alcohol and illegal drug use, and a portion of these costs could not be robustly allocated to either substance independently. With regards to violent crime, the independent and joint estimates are based on Police detainee interviews. The interviews list either alcohol, illegal drugs or both as contributing factors to a crime, with the result that the joint portion of violent crimes cannot be disaggregated by substance. 3.3.4 Avoidable costs One focus that has seen increasing research over the past decade is what costs are susceptible to policy intervention. These costs are known as avoidable costs of substance misuse (Rehm et al 2008, Collins and Lapsley 2008). Avoidable costs are the proportion of total costs, or the burden of misuse, that could in principle be changed given the implementation of appropriate public policies. The International Guidelines for the Estimation of the Avoidable Costs of Substance Abuse (Collins et al 2006) argue that identifying the social costs of substance misuse involves estimating the relevant avoidable proportion of each cost category and applying these proportions of avoidable cost to the relevant aggregate cost estimates. That is, estimates of the total costs of drug misuse comprise avoidable and unavoidable costs. Further, Collins et al argues that avoidable costs are the potential economic benefits (i.e. costs avoided) from substance misuse harm minimisation strategies. These estimates can be used in policy to determine the appropriate level of resources that should be devoted to these strategies. Single et al (2003) argues that estimates of avoidable costs do not indicate how these cost reductions might be achieved or whether the social benefits that result from these programmes exceed their social costs. To do this, Single et al (2003) states that project appraisals must be undertaken that evaluate the efficiency of alternative policies or interventions and/or treatments. These appraisals could use cost-effectiveness analysis to compare the cost of alternative policies or interventions and/or treatments, but they should be undertaken from the viewpoint of the community as a whole, not just the Government.
23
The level of alcohol misuse will determine the treatment goal from reducing alcohol consumption to abstaining from alcohol consumption (Collins and Lapsley 2008, Room et al 2005). Room et al (2005) divides alcohol treatment into three general categories: brief intervention specialised treatment programmes mutual help groups.
Collins and Lapsley (2008) defines brief interventions as advice and information provided to at risk drinkers in the context of a primary care physician consultation. This information is usually conveyed verbally but may be accompanied by additional support such as follow-up telephone calls and printed material. Gibson et al (2007) also includes opportunistic interventions in primary healthcare settings as part of these interventions. This is where general practitioners screen patients on the quantity and frequency of their alcohol consumption levels and encourages patients who are drinking at harmful levels to decrease their consumption. Gibson et al (2007) divides interventions into brief interventions, psychosocial interventions (such as motivational approaches, cognitive-behavioural approaches and self-guided material), and pharmacotherapies. Room et al (2005) states brief interventions provide prophylactic treatment before or soon after the onset of alcohol-related problems. The study recommends motivating high-risk drinkers to moderate their alcohol consumption rather than promoting total abstinence. Various studies argue that brief interventions are better suited to individuals with mild drinking problems while specialised treatment programmes are better suited to individuals who are heavily dependent on alcohol. However, Collins and Lapsley (2008) argues there is a role for brief interventions for those individuals who are heavily dependent on alcohol alongside pharmacotherapies. Specialised treatment programmes involve interventions that manage the withdrawal of alcohol, prevent relapse, and help with social and psychological rehabilitation of the problem drinker. They include detoxification, rehabilitation treatment, therapeutic approaches, and pharmacotherapy using alcohol-sensitising drugs. Room et al (2005) states that mutual help groups, including groups such as Alcoholics Anonymous, are not considered formal treatment. These groups, the study argues, are often used as a substitute, alternative, or adjunct to treatment. In addition, Room et al states
24
that several studies suggest Alcoholics Anonymous can have an incremental effect when combined with formal treatment and attendance at the group is better than no intervention. Gibson et al (2007) terms these interventions motivational interventions as they involve a client-centred non-confrontational style of counselling. Gibson et al also discusses selfguided interventions, which may fall loosely within this group as they involve self-guided material provided in booklets and electronic format. This material is initially provided in a primary healthcare setting but little contact is made with medical staff. Economic literature has argued that the cost of treatment is worthwhile from a societal perspective due to the reduction of harms from alcohol consumption including crime, court costs, and productivity changes. However, the literature comparing the cost-effectiveness of different types of treatment is limited (Gibson et al 2007, Collins and Lapsley 2008). In addition, only a small number of studies have been completed on some alcohol treatments and some of these studies have lacked a control group (Gibson et al). Chisholm et al (2006) argues that only personal interventions aimed at hazardous drinking have been subjected to economic evaluation, and provides an example from the United States and Australia on the cost-effectiveness of these strategies. In addition, there are limited economic evaluations of drug treatment. Those that have been evaluated use observational studies of treatment outcomes in samples of patients with mixed substance misuse problems including opioids. Using examples from the United States again, Chisholm et al discusses the substitutive maintenance treatments for opioid dependence where buprenorphine maintenance treatment (BMT) provides a viable and cost-effective alternative to methadone in the treatment of opioid dependence. To evaluate interventions, Gibson et al (2007) uses two treatment outcomes in the analysis; percentage change in alcohol consumed and percentage change in the proportion of abstinent days. These outcomes were chosen as not all the treatment outcomes were consistently expressed as a single measurable unit. As such, Gibson et al (2007) also argues that it illustrates the difficulties of using research studies with non-comparable outcomes to inform policy on the cost-effectiveness of different treatments. Cost-effectiveness analysis can be used to understand the relative costs and outcomes of different treatments. Collins and Lapsley (2008) states that this type of analysis uses a single outcome unit such as value per life years saved or deaths prevented. However, a single outcome unit is not commonly available in substance-dependence research. As such, some studies such as Chisholm et al (2004) use DALYs to compare tax, brief interventions, random breath testing, restricting sales access, and advertising bans in 12 global regions. Gibson et al (2007) states that this form of analysis, which uses DALYs, is
25
often criticised as it ignores the context in which people experience disease and assumes the alleviation of a disease has the same impact in a poor country as it does in a rich one. As such, Gibson et al uses a cost-consequences study to present the costs and outcomes of interventions. This method does not directly compare costs and outcomes but estimates incremental cost-effectiveness ratios. 3.3.5 Injury costs Another analytical focus of particular interest in this project is costs due to injury. A New Zealand study, Connor et al (2005), found that injury was a major contributor to alcoholrelated mortality, and in particular, most alcohol-related deaths before middle age were due to injury. None of the comprehensive substance misuse cost studies (Collins and Lapsley 2008, Johansson et al 2006, Rehm 2006, Easton 1997) specifically took an injury cost focus across the full range of components, although some injury costs may be included in specific components such as health care for fall injuries. Rather, there was a wide range of separate studies that focus on specific cost categories, cost components or types of injury (Lee and Snape 2008, Cremonte et al 2006, Miller et al 2001, Snively 1994). For example, the literature may relate to intangible costs due to alcohol-related injuries, alcohol-related road accidents and injury, or factors underlying injuries caused by falls. A general finding in the above literature is that alcohol increases the risk of a wide range of injuries. However, in some cases inexperienced drinkers put themselves at greater shortterm risk of injury than experienced drinkers (Borges et al 2006). There is little evidence on the increased injury risk of illegal drug users. While this review did not aim to provide a comprehensive survey of studies focused on AOD injury costs, these studies highlight the significant role alcohol has in causing a wide range of injuries, and of illegal drugs in traffic accidents. Connor et al (2005) estimates the burden of death, disease and disability attributable to alcohol consumption in New Zealand, using mortality data from the NZHIS and DALY estimates from the WHOs CRA study (Ezzati et al 2004). Connor et al argues that most alcohol-related deaths before middle age were due to injury. The authors argue that frequency of heavy drinking and intoxication were both associated with injury and death due to injury. These injuries were often related to patterns of drinking and the average volume drunk, and while people who were binge drinkers were the highest injury risk, even among low volume regular drinkers there was an increased risk of injury compared to non-drinkers.
26
To decrease injuries associated with alcohol, Connor et al (2005) suggests the context and pattern of drinking needs to change. One way to do this is through policy that focuses on preventing alcohol-related injuries. The study argues that prevention policies should be implemented in the workplace as well as brief interventions in emergency rooms. Comprehensive long-term studies by Lee and Snape (2008) and Buchanan et al (2005) for Christchurch and Waikato Hospitals show there is a strong link between alcohol use and maxillofacial fractures caused by inter-personal violence. In the Christchurch Hospital study, 49 percent of all facial fractures were caused by alcohol-related incidents, while in the Waikato study, 34 percent were noted as being related to alcohol. Alcohol involvement was noted if a patient had two standard drinks or more within the hour of injury or when a patient was injured by another person who met that criterion. Alcohol reporting usually relies on patient self-reporting or clinical reporting by an emergency physician due to a lack of a routine blood alcohol level checks (BAC). The Lee and Snape (2008) study shows that 65 percent of alcohol related fractures required hospital admission, and 58 percent required surgery. It also discusses the psychological harm caused by such events. It notes research (Bull and Rumsey 1998) that found permanent physical disfigurement could have a serious effect on social and emotional functioning. Humphrey et al (2003) interviews a random sample of patients with injuries from an Auckland Emergency Department in 2000 and correlates the results with tested BAC of the patients; participant self report of alcohol intake; and clinical observation. 35 percent of injured patients reported having consumed alcohol. Out of the findings, the risk of injury is significantly increased with alcohol consumption. Borges et al (2006) considers the role of drinking experience and accident risk. It concludes that while people who misuse alcohol may have a higher risk of accident and injury, inexperienced drinkers put themselves at greater short-term risk of injury. Sethi et al (2006) looks at the relationship between income inequality and injury. In high income countries, including New Zealand, the study found that on average for every death from injury, there are an estimated 30 hospital admissions, 300 emergency department attendances, and many more people who seek help from their family doctor or self-treat. However, people living in low- and middle-income countries are 3.6 times more likely, on average, to die from an injury than those living in high-income countries. The report argues an emphasis in public health should be placed on injury prevention as proportionally higher returns can be expected from prevention.
27
Of note is the absence of almost any literature on illegal drug caused injuries. One exception is a longitudinal study by Fergusson et al (2008). It establishes a link between cannabis, alcohol and motor vehicle collisions and shows that cannabis use was a greater risk for the driver than alcohol use. However, this study has been debated on various grounds since its release. Jones et al (2006) is another study of drugged driving. However, the study involved purposive interviewing of recent cannabis users presented with hypothetical scenarios around random drug testing and increased penalties for driving under the influence of cannabis. Its analysis of accident risk from drugged driving was a subsidiary aim.
28
Method and Calculation Summary
This section provides a brief summary of the method used in the study and calculations relating to population patterns and drug use. Additional detail is provided in section 9. The methodology and estimation methods for a number of components in this study are the same or similar to those used in BERLs (2008a) illegal drug harm study. That report provides some additional and background detail for this report, and should be read alongside this one. 4.1 Method
This study uses a prevalence approach to calculate the current impact of harmful drug use. The prevalence approach estimates resource diverted in a given year due to the impacts of past and present harmful drug use. This approach compares the costs estimated to a counterfactual situation, in this case where there was no past or present harmful drug use. That is, in order to determine the harm avoided by reducing drug consumption we compare the current situation with drug use to a hypothetical case where there is no harmful drug use. More detail on the prevalence approach in outlined in Appendix section 9.1. The prevalence approach focuses on the impacts of both past and present drug use, and is likely to give an informative picture of the total impacts of harmful drug use in a given year. The study aimed to estimate net social costs, rather than gross social costs of drug use. That is, the approach takes into account that drug use may offset some costs as users reduce the burden on societys scarce resources. For example, premature death reduces the health care that users might otherwise have required if they had lived longer. However, health benefits due to drug use are explicitly excluded from the scope of the study. There is an absence of rigorous evidence for health benefits from the consumption of illegal drugs, and contentious studies regarding the health benefits of alcohol (Lindberg and Amsterdam 2008, Fillmore et al 2007, Fillmore et al 2006). The costs are presented in terms of tangible and intangible costs. Tangible costs relate to resources used (or diverted) due to the harmful use of drugs in New Zealand. Tangible costs can be further split into direct costs (resources diverted from an alternative use) and indirect costs (potential resources or output that is not generated as a result of harmful use). Intangible costs are psychological or welfare impacts borne by individuals that cannot be shifted (Collins and Lapsley 2008). These do not have (productive) re source implications for society.
29
We follow Collins and Lapsleys conclusions that there is currently no robust epidemiological evidence to quantify potential causal interactions between alcohol and other drugs (see section 3.3.3). As such, we assume that the impact and cost estimates for alcohol and other drugs are independent, and no adjustment for double counting is required. We also follow their method, for comparability, and do not attempt to account for co-morbidities. 4.2 Population patterns and impacts related to AOD use
The sub-sections below set out some of the key population and drug use parameters used in the studys calculations. 4.2.1 Prevalence of AOD use and harmful AOD use The prevalence of harmful AOD use in 2005/06 was estimated using several New Zealand studies.
26
Box 1 Harmful AOD use

Harmful alcohol use covers both hazardous or high risk drinking patterns. Estimates of the impacts of harmful use for low risk drinkers are made only where specific information was available. For example, the cost of inputs diverted for the production of alcohol that is used harmfully is estimated for the hazardous and high risk drinking pattern groups only. Costs
These
studies cover a wide range of gender, age and ethnicity groups. In particular, Connor et al (2005) provides a separate analysis of Mori and non-Mori drinking patterns. We use the total population parameter estimates, and therefore the figures implicitly allow for differences in drinking patterns by age, gender and ethnicity. Table 4.1 summarises the results of these estimates, giving the number of harmful users by age and gender according to the type of drug(s) used: alcohol only, other drugs only and joint AOD. The latter group captures polydrug users, that is, people who use both alcohol and other drugs. The other drug group also includes people who use multiple substances, but only where this does not include alcohol. BERL (2008a) describes how other drug users were allocated to mutually exclusive categories so a poly-drug using person was counted only once.
are estimated for all people where alcohol was a contributing factor where suitable information was available, such as for road crashes. Any illegal drug use is assumed to be harmful, reflecting the absence of evidence for the non-medical health benefits from the consumption of illegal drugs (Ridolfo and Stevenson 2001).
26
Prevalence data was estimated using data from a New Zealand study into the burden of disease due to alcohol (Connor et al 2005) and two national drug use surveys by SHORE (Wilkins and Sweetsur 2003, 2007). A frequent drug user survey (Wilkins et al 2006) was also used to estimate illegal drug use.
30
Table 4.1 Harmful drug users by sex and age group, 2005/06
Alcohol only Age 13-14 15-19 20-24 25-29 30-34 35-39 40-44 45-49 50-64 65+ Total Male 11,600 27,900 22,500 29,900 24,100 25,900 29,100 21,700 46,900 20,000 259,600 Female 5,900 28,400 25,000 28,700 21,100 26,700 26,900 24,300 48,100 18,000 253,300
OD only Male 0 1,000 4,900 1,100 0 3,400 1,900 1,800 4,200 2,800 21,200 Female 0 1,300 0 0 0 800 700 700 1,500 1,100 6,100
Joint AOD Male 1,100 16,600 19,500 10,100 7,700 4,900 4,800 3,600 7,700 3,300 79,000 Female 500 10,600 16,100 7,400 3,700 700 2,000 1,800 3,600 1,300 47,700
Total Both 19,100 85,800 88,000 77,200 56,600 62,400 65,400 53,900 112,000 46,500 666,900
Source: BERL, StatsNZ
The estimates indicate that approximately 667,000 people aged 13 plus engaged in harmful drug use in 2005/06. The most common drug used harmfully was alcohol, representing about 95 percent of harmful users. A further 2.27 million people used alcohol in a low risk way; this use may on occasion result in harm. Australian and New Zealand drug use prevalence Table 4.2 and Table 4.3 compare New Zealand and Australian drug use prevalence. Table 4.2 New Zealand (2005/06) and Australia (2004/05) alcohol use prevalence by 27 drinking pattern
Low risk Hazardous Australia New Zealand

Source: BERL, ABS
High risk 5.6% 8.7%
Any use 63.2% 87.1%
49.5% 68.7%
8.0% 9.7%
Table 4.2 suggests that New Zealand has a substantially higher prevalence of alcohol consumption than Australia, with 18.4 percent of people aged 18 plus drinking in a harmful manner versus 13.6 percent in Australia. Table 9.4 in the appendix shows that this pattern is
27
This table is based on the New Zealand figures estimated for this study and the ABS (2006) National Health Survey 2004-05. For comparability with the Australian figures, the table is based on people age 18+ years only. The main study estimates are based on prevalence rates for people aged 13+ years. On this basis, approximately 18.8 percent of the general population aged 13+ drink harmfully, and 22 percent of the population who drink aged 13+ do so harmfully.
31
consistent across all age groups and drinking patterns, bar high risk drinking in later middle age groups, which is lower for New Zealanders. Table 4.3 New Zealand (2006) and Australian (1998) illegal drug use prevalence by age 28 group
New Zealand Illegal drug Cannabis Opioids Cocaine Amphetamines (excl CM and MDMA) Crystal methamphetamine (CM) Ecstasy (MDMA) LSD
Source: BERL
Australia 15-45 y.o. 24.2% 1.0% 1.8% 5.0% n.a. 3.2% 4.0% 14 y.o.+ 17.9% 0.8% 1.4% 3.7% n.a. 2.4% 3.0%
15-45 y.o. 17.9% 0.2%-0.4% 1.1% 3.4% 0.8% 3.9% 1.8%
Table 4.3 indicates that New Zealand generally has a lower illegal drug use prevalence than Australia for all drug types except ecstasy. 4.2.2 Drug consumption in New Zealand Alcohol consumption in 2005/06 was based on the estimated number of drinkers by drinking pattern and average daily consumption per person by drinking pattern. Average daily consumption is used as a technical measure. But as an average it may understate actual harmful consumption patterns on a typical drinking day (Ministry of Health 2008). Table 4.4 below lists our assumptions about average daily consumption, in grams of alcohol per day by drinking pattern. The assumptions for the low and hazardous patterns use the midpoint of the drinking pattern range. With no upper limit on the high risk pattern, we could not calculate a midpoint, so it was necessary to assume an average daily consumption for the high risk group. We used an average of double the lower limit for hazardous drinkers and the lower limit for the high risk group (see Table 9.1. in the appendix for the average daily consumption ranges).
29
28
This table is based on the 2006 figures for New Zealand (Wilkins et al 2006) and 1998 Australian figures reported in McFadden (2006) based on figures from the Australian Institute of Health and Welfare (AIHW, 1999). For comparability we have estimated the Australian 15-45 year old prevalence rates, as well as the rates for all people aged 14 years old and over as reported in AIHW (1999).
29
The average daily consumption ranges are consistent with the WHO categories (Rehm et al 2004), the National Health and Medical Research Council (NHMRC) Australian alcohol guidelines (Pidd et al 2006), and a recent New Zealand study on the burden of death, disease and disability due to alcohol (Connor et al 2005).
32
Table 4.4 Average daily alcohol consumption by sex and drinking pattern (grams of alcohol per day), 2005/06
Risk category Women Range Low risk Hazardous High risk 0-19.99 20-39.99 40+
Grams of alcohol per day Men Range 0-39.99 40-59.99 60+ Midpoint 20 50 110
Midpoint 10 30 70
Source: BERL, based on SHORE data, Connor et al (2005)
Based on the estimated number of estimated harmful drinkers in Table 4.1 and the total estimated volume of alcohol consumed, we calculate that approximately 50 percent of all alcohol consumed was estimated to be harmful.
30
The estimate of the proportion of alcohol consumed in a harmful manner was applied to estimated annual expenditure of $684.5 million on alcohol to determine the value of resources diverted to harmful alcohol consumption. The estimate of total annual expenditure on alcohol for 2005/06 was based on an average of the corresponding figures from the 2003/04 and 2006/07 Household Economic Surveys (HES) conducted by Statistics New Zealand (StatsNZ). Other drug consumption has two parts: illegal and legal. Illegal drug consumption was based on how often the illegal-drug-using sub-populations used illegal drugs in the last year and how much they consumed on average per occasion. This process is fully described in section 4.2 (illegal drug consumption in New Zealand) of BERLs (2008) report on the social costs of illegal drug use. The process allowed for different consumption patterns for social and frequent drug users, and for poly-illegal drug use. We do not explicitly estimate the harmful consumption of legal drugs due to insufficient information. However, legal drug costs are included where they are used to treat alcoholand-illegal-drug-related conditions, for example, tranquilisers for treating depression, or methadone placements for addiction treatment. In these instances, costs relate to the resources being diverted for these treatments that could otherwise have been used elsewhere. Similarly, the hospital cost analysis includes categories for anabolic steroid and general categories for drug poisoning. This analysis will therefore include hospital-related costs for some harmful legal drug use.
30
Converting total estimated alcohol consumption to litres, based on a standard of 1.25 millilitres per gram of alcohol, we calculated that New Zealanders consume 30.98 million litres of alcohol per annum. This lines up closely with StatsNZs measure of 29.998 million litres of alcohol available for consumption in 2005/06. Allowing for home made liquor, we would expect total alcohol consumed to be higher than the recorded figure.
33
4.2.3 Drug-attributable mortality Alcohol and other drug attributable mortality was calculated from a dataset of all deaths in New Zealand over a five year period. Australian attributable fractions were used to estimate the number and proportion of deaths that were caused by harmful alcohol use, and other drug use among diseases and conditions that were wholly or partly attributable to AODs (Ridolfo and Stevenson 2001, Collins and Lapsley 2008). Using this estimate of drug-attributable deaths, we derived mortality rates that could be used to estimate the additional population who would be alive today had it not been for their deaths in the past due to harmful drug use. We used average mortality rates for a five year period rather than mortality rates based on the 2005/06 year. This aims to provide a more stable estimate of mortality rates, and partly to reflect past changes in drug prevalence. These rates were applied to past populations to estimate past deaths, less the population that would have died from normal causes during the intervening period. The difference between the current and estimated population gives the additional population if harmful drug use had not occurred. Our findings show that 14,250 people would be alive in 2005/2006 if not for harmful alcohol use in the past, and 3,608 people would be alive if not for other drug use. 4.3 Calculations of the costs of harmful drug use
The sections below provide a brief summary of how we calculated the costs of harmful drug use in 2005/06 by component. The main methodological description is included in an appendix in section 9.4. The calculations primarily draw on cost data for the 2005/06 financial year, and all figures are reported in 2005/06 dollar terms, and are exclusive of Goods and Services Tax (GST). 4.3.1 Lost output A societys capacity to produce output depends on its labour force as a factor of production and the non-market resources in households.
32 31
This study considers four forms of production loss due to the effects of harmful drug use on the labour force and households:
33
31 32
Dollar figures were adjusted where necessary using an appropriate New Zealand GDP deflator index.
Premature death and illness of those not in paid employment reduce a countrys capacity to support itself. That is, although unpaid, the non-market activities people do for their own household and those around them are valuable. We note, however, that there is no agreed substitute non-market valuation for the contribution of these people. Furthermore, non-market contributions are not counted as part of a countrys Gross Domestic Product (GDP), and it would be inconsistent to include this when expressing aggregate harm as a proportion of GDP.
34
Premature mortality Excess unemployment Absenteeism due to sickness or injury Reduced productivity
The distribution of additional population in 2005/06 by workforce status was based on the additional population calculations and population statistics on workforce participation and employment characteristics, for example, whether the person was employed full time or part time. The participation rate and workforce structure by age group mirror the corresponding population figures in 2005/06, which are based on StatsNZs Household Labour Force Survey data. These figures are used to calculate the labour costs due to premature mortality and morbidity. Section 9.4.1 details the methods for calculating lost output. Table 9.6 and Table 9.7 in that section show the distribution of additional population and working-age harmful drug users in 2005/06 by drug type and workforce status, for example, whether the person was employed full time or part time.
34,35
The impact of drug use on mortality and labour market outcomes differs depending on the type(s) of drug used. The hypothetical situation where a person never engaged in harmful drug use, however, assumes that all people would have an equal likelihood of mortality or illness and a similar pattern of labour market outcomes. Therefore, we use population average earnings and output figures. Earnings were calculated using data on median earnings per hour and the average number of hours worked per week by age, sex and employment status (part time and full time). Earnings statistics were sourced from the StatsNZ New Zealand Income Survey for the June 2006 quarter. The Household Labour Force Survey (2004) provided hours worked per week. The value to society of lost output is considerably larger than lost earnings alone, for example, in addition to lost wages there is also lost profit. As such, the earnings profiles were scaled up to reflect the difference between wages and residual value added. The resulting output profiles were based on the assumption that the average GDP per FTE
33 34
BERLs drug harm project considered illegal-drug related premature mortality and absenteeism impacts only.
The workforce structure of people who use drugs harmfully is based on the working age population, that is, people aged 15-64 years old. The working-age population total differs from the estimated total number of harmful users in Table 4.1, as that table includes people aged under 15 and over 65 years old.
35
We prepare separate estimates for the lost output of people in prison for drug-related offences. This number of people (by drug type, age and gender) is deducted from the workforce figures in Table 9.7 before calculating the main lost output estimate, in order to avoid double counting.
35
(BERL Forecast Database) is 1.87 times the average wage income (StatsNZ). These output profiles by age, gender and workforce status are used in the calculations below. Workforce losses due to premature mortality Drug-attributable mortality causes a reduction in societys productive capacity that society could have benefited from in the counterfactual case (a world without harmful drug use). This cost is a function of how many people die prematurely due to drug use and what those people could have earned. The first element is the profile of working-age groups for the estimated additional population by drug type. We estimate that the population would have had an additional 17,800 working age people in the absence of harmful drug use. The earning capacity of these groups was determined by matching them by age and gender to give labour force characteristics.
36
This included estimating the proportion of people
engaged in the labour market and of those people how many were unemployed, in part time employment and in full time employment. The output profiles by age, gender and workforce status were then applied to the lost workforce profiles to give drug-attributable production losses due to premature mortality. The losses by drug type were aggregated to give gross mortality related production losses. Premature mortality has an offsetting effect by reducing the demand on societys scarce resources through reduced consumption. That is, were these people to have lived and increased output, we net out the resources they would also have consumed. Consumption resources released by the lost population were based on average private and public consumption expenditure per person. Private expenditure was derived from HES figures on average household expenditure and the average number of persons per household. Public expenditure per capita (excluding transfers such as social welfare benefits) was based on a study of population-related expenditure completed by BERL for the Department of Labour (BERL 2007a).
36
We assume that the labour force characteristics of current drug users are the same as the general public in the counterfactual, as in the counterfactual there is, and has never been, harmful drug use. Therefore, we assume that people who would otherwise have used drugs have equivalent employability and productivity to people who do not use drugs. It is possible that there are background, or latent, characteristics associated with drug use that also affect a persons work potential, for example, the motivation to seek higher education. Beyond accounting for age and gender characteristics of drug users versus non-drug users, we are not aware of work on such potential latent characteristics that would allow us to construct an alternative earning profile.
36
Workforce losses due to excess unemployment Rayner, Chetwynd and Alexander (1984, p47) suggest that harmful alcohol use increases unemployment in New Zealand by 10 percent. Although this study was based in New Zealand, it was a preliminary estimate and only for alcohol-related costs. Therefore, no estimate is made for excess unemployment attributable to other drug use. We take a conservative approach and apply the 10 percent proportion only to identified high risk drinkers, rather than including the wider category of harmful drinkers, which includes hazardous drinkers. We estimated this involved 14,600 full time and 3,900 part time high risk drinkers and other drug users. These people are removed from the estimated labour force below, to avoid double counting. We estimate the value of the lost output due to excess unemployment by applying the output profiles to the estimated workforce profiles for the three drug categories. Workplace losses due to absenteeism To estimate the number of people absent from work due to alcohol-attributable illness we used recent Australian research (Pidd et al 2008).
37
This research indicated that 4.2 percent
of male drinkers and 2.5 percent of female drinkers have time off work due to alcohol-related illness. This study also reported on the average number of days off due to alcohol-related illness for low risk, risky and high risk drinkers. For our study, these figures were applied to the estimated number of all drinkers by drinking risk pattern to estimate the number, and value, of days lost to alcohol-related illness. Other drug attributable absenteeism is estimated in a similar fashion to Collins and Lapsley (2008). It begins by drawing on probability estimates from Bush and Wooden (1994) on the impact of substance smoking and alcohol on absences from the workplace. As in Collins and Lapsley we assume that the probability of absenteeism is the same for tobacco use and other drug use. Overall, male drug users are assumed to be absent 70 percent more days and females 20 percent more days a year than abstainers. Australian Bureau of Statistics (2003) survey data on employee absences were used in the absence of robust New Zealand data. The survey provided information on how many hours a person taking sick leave took per week and the proportion of workers taking sick leave. These weekly figures were used to estimate the number of hours of sick leave per worker per annum.
37
Assumptions based on an Australian study were used in the absence of a comparable New Zealand study with an appropriate drinking-pattern breakdown.
37
Estimates of the other drug-using population by age and gender were used to develop a workforce profile as for harmful alcohol users. To ensure poly-drug users were not counted twice, separate estimates were made for people who use alcohol only, other drugs only and both alcohol and other drugs. Combining the estimates of hours of sick leave, hourly output and the number of workers affected generated the value of drug-related absenteeism by drug type and the total cost of absenteeism. ACC data on workers compensation for injury was used to calculate injury-related absenteeism due to harmful AOD use.
38
This process is described in section 4.3.8.
Workforce losses due to reduced productivity We estimate output lost as a result of harmful alcohol use impairing worker productivity using information on the number of days affected by drinking pattern and an estimate of reduced work efficiency on such days. As for the excess unemployment estimates, we do not estimate reduced productivity losses due to other drugs, and we take a conservative approach and estimate this loss for identified high risk drinkers only. That is we focus on the estimated high risk group of 146,100 full time, 39,000 part time and 103,400 UE/NiLF high risk drinkers and other drug users.
39
Jones et al (1995) quantified absenteeism and reduced productivity costs associated with alcohol consumption using New Zealand survey data. This study showed that the top 10 percent (heavy) of drinkers performance was reduced.
40
experienced 3.93 days per person on average when work We calculate the total number of days affected by using Jones
41
et als figures by workforce status and applying them to the age and gender output profiles for the high risk drinking group discussed above. We assume that there is a 25 percent reduction in workplace productivity on days affected by harmful alcohol use. This is the same estimate used by Jones et al (1995) and Rayner et al (1984). However, this estimate is based on dated American research that uses expert opinion on the reduction in efficiency by alcoholics (United States General Accounting
38
We assume that Pidd et als (2006) alcohol-related absenteeism estimates relate to illness only. Their estimates of absenteeism due to both illness or injury attributable to alcohol are almost three times that of the alcohol-related absenteeism estimate. Our estimates of non-injury and injury-related absenteeism and reduced productivity broadly mirror these relativities.
39 40
The abbreviation UE/NiLF refers to people who are unemployed or not in the labour force.
This corresponds very closely to the estimated proportion of drinkers in the high risk drinking group (9.99 percent) used in this study.
41
The detailed method appendix in section 9.4.1 also discusses the impacts for the next highest drinking group (the 10th to 25th percentile of drinkers).
38
Office 1970). As such, this estimate illustrates the potential magnitude of lost output due to reduced productivity. Contemporary New Zealand research would be desirable to validate this estimate. Non-market output losses Losses in the non-market sector of the economy were calculated along broadly similar lines to the market sector calculations outlined above. Two elements that differed for these estimates were the amount and value of time spent on unpaid activity. The number of hours per person spent on unpaid activities in the home was based on StatsNZs Time Use Survey (2001). This survey found that women spent an average of 4.8 hours per day on unpaid work and men 2.8 hours per day. Grossed up to an annual figure, this equates to a woman spending approximately 1,750 hours on unpaid work and men 1,020 hours. The Time Use Survey is based on a sample of people inside and outside the labour force. If non-employed people do more unpaid work than employed people, then these annual estimates would tend to under-estimate household losses due to premature mortality. Ratcliffe et al (1996) have examined the value placed on activities in the household. Based on this research, we conservatively assume that lost output in households was valued at half the median hourly wage for unskilled occupations. The value of an hour of unpaid activity was assumed to be the same for men and women. 4.3.2 Drug production Single et al (2003) discuss the value that should be placed on inputs diverted from legitimate uses to drugs produced for harmful use. This study assumes that resources used in such drug production have alternative uses, that is, their opportunity cost is not zero. Therefore, resources involved in the production of drugs that are used harmfully represent a cost to society, and a reduction in harmful consumption would release resources that could be used for other consumption or investment uses.
42
In the case of illegal drugs, we drew on the estimate calculated in BERLs (2008) drug harm study (the calculations are described more fully in section 5.3 of that report). The estimate covered the value of domestic production plus imported drugs. The volume of drugs, by drug type, were typically valued at a fraction of their street price, following Collins and
42
This is similar to how other cost components are treated within this study, for example, it is assumed that AOD treatment professionals could work productively in another profession were there no harmful alcohol or other drug use, hence their use is a cost to society.
39
Lapsleys approach. However, the value of domestically sourced pharmaceuticals diverted for opioid production was based on work by Sheerin (2004), and we allowed for differences in cost of imported and domestically sourced resources used in the production of methamphetamine. In the case of alcohol, we draw on the consumption estimates in section 4.2.2 as one component in the calculation of resources diverted for harmfully used alcohol. That is, we assume that 50 percent of all alcohol consumed is done so in a harmful manner.
43
We
applied this proportion to an estimate of New Zealanders annual aggregate expenditure on alcohol. New Zealanders spent an estimated $684.5 million (excluding taxes) in the 2005/06 year on alcohol, based on an average of figures taken from StatsNZs 2003/04 and 2006/07 Household Economic Surveys. Applying the 50 percent proportion to the annual aggregate expenditure, we estimate that $342.2 million of resources were diverted to produce alcohol that was then used harmfully. 4.3.3 Justice sector The cost of crime is compiled from several sub-components. The sub-components are: customs, community costs (victims of crime), Police, courts, prisons and community sentences, including home detention. Customs The cost of harmful drug use to the Customs Service comes in two parts: its drug enforcement programme, and alcohol excise and customs duty operations. The cost of drug enforcement was derived using estimates of total Customs Service budget and drug enforcement expenditure, and estimated budget priorities.
44
The cost of collecting alcohol
duties was estimated using Customs Service expenditure for duty collection, and weighting this by the proportion of collected duties that were related to alcohol in 2005/06. Community Costs Victims of Crime Alcohol-and-drug-related costs borne by victims of crime include preventative expenditure, property losses, lost output, health service use and intangible costs. Community costs were calculated as the number of alcohol-and-drug-attributable crimes multiplied by the average cost per offence. This involved applying crime multipliers from a
43 44
Section 11 explores the implications of this assumption for the cost estimates.
This method is similar to that used in the Drug Harm Index, and was developed with specific advice from the New Zealand Customs Service.
40
Treasury study to the offences recorded in 2005/06 to estimate actual crime levels (see Appendix Table 13).
45
The fraction of total criminal offences related to drug use was based
on information from the New Zealand Arrestee Drug Abuse Monitoring (NZ-ADAM) programme (see Table 4.6 below). The NZ-ADAM programme is discussed briefly in the next sub-section, and further detail is provided in Appendix section 9.4.3. Police New Zealand Police (NZP) resources diverted due to harmful alcohol and drug consumption were estimated in three steps: identify total NZP expenditure net of Crown revenue and receipts. Treasurys Budget documents show that this was a net expenditure of $942.6 million. allocate this expenditure to offence categories using proportions based on the time the NZP spent on dealing with various types of offence (see Table 4.5) based on data sourced from the NZP. determine the proportion of offences due to AOD use (see Table 4.6), based on data sourced from HOIs NZ-ADAM study and NZP advice.
47 46
This method is detailed in Appendix section 9.4.3, and is summarised below. Police time by offence category The time Police spent dealing with various types of offence is based on data the NZP provided BERL. The data recorded activity for over 7.63 million hours of Police time.
48
BERL aggregated the NZP data to match the offence categories to those from other data sources, such as the NZ-ADAM data and crime multipliers. The offence categories broadly map to the New Zealand Offence Hierarchy (NZOH) system. Appendix Table 14 details the groups that the NZP supplied the data in and the corresponding eight offence categories
45
Roper and Thompson (2006: 8) note that multipliers derived from the New Zealand National Survey of Crime Victims 2001 are not readily convertible to the particular crime sub-categories covered in this study. A similar caveat holds here with respect to the more recent 2006 New Zealand Crime and Safety Survey (Mayhew and Reilly 2007a).
46
Hales et al (2007) and Hales and Manser (2007) detail the NZ-ADAM programme, its participants and analytical methods.
47
A subsection in Appendix section 9.4.3 discusses an alternative New Zealand data source on the proportion of offences where alcohol is involved the NZPs Alco-link database. While the data are different, our analysis of the alternative data indicates that it confirms the general pattern shown by the NZ-ADAM data, and reinforces the validity of using this source.
48
This equates to just over half the NZPs activities coded to specific offences. A further 7.15 million hours of time was spent on incident and service activity; this time is not coded to specific offence categories. We assume instead that total Police time is distributed by offence according to the recorded statistics, as shown in Table 4.5.
41
used in this study. For example, sexual offences (the 2000 series codes in the NZOH) and dishonesty offences (the 4000 series codes) were reclassified into the Other offence category. Similarly, the Property category includes both property damage offences (the 5000 series codes) and property abuse (the 6000 series codes). Table 4.5 shows the proportion of time NZP spent on various offences in 2005/06. Table 4.5 Police activity by offence category, 2005/06
Offence category % of hours Violent 15.7% Property 5.8% Illegal drugs 7.2% Traffic 23.5% Drink/ Breaches Disorder Drugged Driving 1.7% 6.1% 7.5% Other 32.5% Total - all offences 100.0%
Source: BERL, based on NZP data
The Other offences category, for example, accounted for just under 2.48 million hours (32.5 percent) of Police time in 2005/06, while Property offences absorbed approximately 5.8 percent of Police time. Offences attributable to harmful alcohol or other drug use We estimate the proportion of offences attributable to harmful alcohol or other drug use. The estimates are primarily based on information from the NZ-ADAM programme.
49
The NZ-
ADAM programme measures drug and alcohol use among people who have recently been apprehended and detained in watch houses by police (Hales and Manser, 2008). The NZP fund Health Outcomes International (HOI) Limited to conduct this research. Data are collected from four Police watch houses around New Zealand (Whangarei, Henderson, Hamilton and Dunedin). Detainees at these watch houses who meet specific criteria and who consent are interviewed. More than 2,000 detainees a year are available to participate, and more than 37 percent completed the entire interview process. BERL commissioned HOI to provide summary data for the first three waves of data, from July 2005 to June 2008, by offence categories appropriate to our study (see Appendix Table 15). The detainee data were cross-tabulated by the number of offenders, offence type and the degree to which alcohol or other drugs contributed to the commission of the alleged offence. BERL used this data to estimate the fraction of offences (by offence category) where participants believed their alcohol or other drug use had contributed significantly to the commission of their alleged offence.
49
The NZ-ADAM data do not separately identify drink/drugged driving offences from other traffic offences. Therefore, we estimate drink/drugged driving offences separately from other traffic offences. We assume that all drink/drugged driving offences are attributable to alcohol or other drugs, so the total for this category sums to 100 percent. The distribution of this offence category by drug type is based on advice from the NZP that drug driving accounts for only about 0.26 percent of drink/drugged driving offences.
42
In the case of drug offences, however, while 28.8 percent of detainees said alcohol or drug use significantly contributed to their drug offences, we assume that all drug offending is due to drug or joint drug and alcohol use, and we have scaled the reported proportions up so they sum to 100 percent. On average, across all the NZ-ADAM offence categories, alcohol or other drug use was a significant factor for more than one quarter (26.6 percent) of detainees. In the case of drink/drugged driving offences, which is not separately reported in NZ-ADAM, we assume that all such offences are attributable to alcohol or other drugs, so the total for this category sums to 100 percent. The traffic offences category by drug type, which exclude drink/drugged driving offences, are scaled to ensure that drink/drugged driving offences and other traffic offences related to harmful AOD use are not double counted. base the distribution across drug types on advice from the NZP on the proportion of apprehensions leading to a charge for drink driving (almost all cases) versus drugged driving (about 2.6 in 1,000 cases). Table 4.6 shows the fractions of a particular offence category attributable to alcohol use only, other drug use only and joint alcohol and drug use. The offence categories are based on Appendix Table 14 and Appendix Table 15. Table 4.6 AOD-related apprehensions by offence category, 2005/06
Offence category % offences attributable to Drugs only Alcohol only A&OD only Total AOD Violent 1.5% 20.9% 3.4% 25.8% Property 2.9% 23.9% 8.1% 34.9% Illegal drugs 89.2% 0.0% 10.8% 100.0% Traffic 2.6% 0.0% 4.5% 7.1% Drink/ Breaches Disorder Drugged Driving 4.8% 17.3% 4.8% 26.9% 0.7% 33.1% 2.7% 36.5% 0.26% 99.74% 0.0% 100.0% Other 6.2% 11.8% 4.7% 22.7%
50
We
Source: BERL, based on HOI data and NZP advice
The fractions are based on participants who believed that their alcohol or drug use significantly contributed to the commission of their offence.
51
For example, just over one
50
The (non-drink/drugged driving offence) traffic offences attributable to harmful AOD use is based on NZ-ADAM data, which does not separate out drink/drugged driving offences. To ensure that drink/drugged driving offences are not double counted in the (other) traffic offence category, we scale down the fractions for all AOD-related traffic offences according to the estimated proportion of traffic offences that are due to drink/drugged driving. That is, the traffic category is for all other traffic offences bar drink/driving offences, which are separately estimated. The Police sub-section of section 9.4.3 details this process.
51
While the attribution is based on self-report, Hales and Manser (2007) note that there is a high degree of correspondence between self-reported drug use and positive urinalysis for a range of the illegal drugs analysed. The correlation ranged from 69 percent to 90 percent for use within the past 30 days and 40 percent to 63 percent for those reporting use in the previous 48 hours.
43
third (34.9 percent) of offenders felt that some AOD use contributed significantly to their property offending, with almost one quarter (23.9 percent) being due to alcohol use alone. Table 4.7 shows the estimated levels of Police expenditure by offence category related to harmful AOD use. Table 4.7 AOD-related Police expenditure ($m) by offence category, 2005/06
Offence category Police expenditure ($m) Drugs only Alcohol only A&OD only Total AOD Violent 2.2 31.0 5.0 38.2 Property 1.6 13.2 4.4 19.2 Illegal drugs 60.7 0.0 7.3 68.1 Traffic 5.7 0.0 10.1 15.8 Drink/ Breaches Disorder Drugged Driving 0.8 2.8 0.8 4.3 0.4 18.9 1.6 20.9 0.2 70.2 0.0 70.4 Other 19.1 36.2 14.3 69.6 Total - all offences 90.7 172.2 43.5 306.3
Source: BERL, based on HOI data and NZP advice
Based on the information above, we calculate, for example, that the NZP spent 15.7 percent of their time dealing with violent offences (see Table 4.5), and correspondingly net Police expenditure on violent offences amounted to $148.2 million. Of these offences, 25.8 percent were attributable to some AOD use (see Table 4.6). Applying this percentage to the $148.2 million related to policing violent offences, we estimate AOD-related violent offences diverted $38.2 million of NZP resources (see Table 4.7). As 20.9 percent of violent offences were attributed to situations where the person had used alcohol but not other drugs, the majority of the AOD-related violent offences estimate is due to alcohol misuse alone. Defence force expenditure In addition to expenditure by the NZP, we include expenditure by the New Zealand Defence Force for the RNZAFs No 3 Squadron, which supported the NZPs drug eradication programme. This amounted to an estimated $553,000 in 2005/06 dollar terms. Anti-drink driving advertising campaigns We note that the New Zealand Transport Agency (NZTA) spent $2.8 million and ACC $515,000 for anti-drink driving advertising campaigns. These expenditures are not included in the social cost estimates as they are deemed discretionary policy costs, rather than an intervention that is directly related to drug misuse, as per the recommended approach in the International Guidelines (Single et al 2003).
44
Courts Court related expenditure was estimated using net court related expenditure for 2005/06, and allocating this expenditure to the offence categories using the Police time proportions from the NZ-ADAM programme. Prisons Prison expenditure was estimated for the cost of incarcerating people due to alcohol-anddrug-related crime in 2005/06 and providing specialist drug treatment units in prisons. The first element was calculated by estimating the number of annual inmate equivalents (AIEs) times average cost per prisoner in 2005/06. AIEs were calculated using inmate numbers and average sentence lengths based on the Department of Corrections convictions and sentencing information (Ministry of Justice 2006, 2007c). We applied fractions from the Australian Institute of Criminologys Drug Use Careers of Offenders (DUCO) survey to derive the number of AIEs incarcerated for AOD-attributable crime. The DUCO fractions were adjusted to reflect New Zealand illegal drug and alcohol prevalence rates. That is, the proportions attributed to illegal drug and alcohol use were scaled down and up, respectively. The Department of Corrections provided information on the costs of specialist drug treatment programmes, excluding regional, national and corporate office overhead costs. Incarceration poses a further cost due to the lost output of inmates. Lost output estimates were calculated based on the number of AIEs incarcerated due to their drug or alcohol use. We applied an age, gender and offence profile of inmates based on the prison muster for 2005/06 provided by the Department of Corrections to the estimated number of AIEs. Section 4.3.1 describes the general process for estimating lost output. Community sentences Expenditure on alcohol-and-drug-related community sentences was based on the number of people serving community-based sentences and orders in 2005/06.
52
The number of alcohol-and-drug-attributable community sentences was derived using Ministry of Justice sentencing data and the DUCO conviction fractions used for the prison
52
As the study focuses on the costs of harmful drug use, we do not estimate the benefits to the community of the work performed by people serving community sentences.
45
estimates. These attributable numbers were then applied to total community sentences expenditure. Alcohol-and-drug-related home detention numbers were based on the proportion of corrective services population on home detention and the overall proportion of crime leading to a criminal sentence. The cost of home detentions was calculated by applying an estimated operational cost per home detention to attributable home detention numbers. 4.3.4 Health care Health care costs are compiled from several components. The components are: pharmaceuticals, hospitals and other medical costs such as primary care, ambulances, and, accident and emergency services. These costs are calculated as gross costs net of reductions in health care services stemming from drug-related premature mortality. Pharmaceuticals Data was sourced from Pharmac on pharmaceuticals involved in AOD-related treatment, plus pharmaceuticals for co-morbid depression in illegal drug users estimated as per the BERL (2008) drug harm study, using information from Wilkins, Girling and Sweetsur (2006).
53
Primary care Data outputted from the New Zealand Health Behaviours Survey for this project indicated no excess use above the general population average of primary care services due to harmful alcohol consumption. However, there are several services for people with alcohol-related problems, such as specialist consultations, that require GP consults. As such, we include an
53
An equivalent source with the necessary underlying parameters to estimate co-morbid depression for harmful alcohol users was not available. To provide a sense of the magnitude of this omission we examined the prevalence of mood disorders and treatment for these disorders in the general population based on Ministry of Health (2008) figures and the relative prevalence of depression amongst other drug users related to their drug use. Other drug users are estimated to have depression as a result of their drug use that is double the population average prevalence. We begin by assuming a similar magnitude of alcohol-related depression for harmful alcohol users aged 15 years and over. That is, we assume this group has a prevalence of mood disorders as a result of their alcohol use of 21.8 percent. Assuming these people receive treatment in a similar proportion as the general population (48.6 percent), the cost of pharmaceutical treatment as a result of alcohol-related co-morbid depression would be approximately just under $1.1 million. We note that the Ministry of Healths (2006) mental health survey reports a prevalence of 2.6 percent for alcohol abuse disorders, and that this is just over double the rate of 1.2 percent for drug dependence disorders. If the prevalence of mood disorders for alcohol users was double the rate of other drug users, that is approximately 50 percent, then the cost of treatment for co-morbid depression (assuming they received it at the population average rate) would roughly equal just over $2.5 million in 2005/06.
46
estimate based total GP funding and the ratio of people admitted to hospitals with AODrelated conditions, following Collins and Lapsleys (2008) method and advice.
54
We also estimated the cost of providing the services of drug and alcohol workers for harmful alcohol users. The method used to calculate this component is the same as for illegal drug users (see below). We assume the proportion of harmful alcohol users engaging these services was the same as for illegal drug users. Illegal drug-related primary care use was based on reported health service use by frequent drug users in Wilkins, Girling and Sweetsur (2006), and detailed in the BERL (2008) drug harm report. We include estimates for the use of GPs, counsellors, drug and alcohol workers, social workers, psychologists, and psychiatrists. The net cost of GP services takes into account savings from premature AOD-related mortality. Data from the General Practice Computer Databases were used to develop an age-gender profile GP use.
55
We apply this profile to the estimated additional population in
the absence of harmful drug use by age and gender to calculate a cost offset due to lower GP use as a consequence of premature mortality. Alcohol and other drug treatment The AOD treatment category covers specialist treatment units in the community, residential beds, inpatient medical detox beds and preventative services, such as methadone treatment.
56
Total national expenditure on alcohol and drug programs were obtained from
the Addiction Treatment Services division within the Ministry of Health. This was $96.7 million for the 2006/2007 year. This figure was adjusted to $92.9 million in 2005/06 dollar terms using a New Zealand GDP deflator index. Data was not able to be separated into alcohol programs and drug programs. This is because many of the programs were interrelated and dealt with both harmful alcohol and
54 55 56
This process is described in more detail in Appendix section 9.4.4. The databases draw on the Royal New Zealand College of General Practitioners Computer Research Network.
Our study does not estimate nursing home expenditure attributable to harmful drug consumption. Although the contribution to overall drug harm may be substantial, it was excluded due to data limitations and in order to ensure a robust, but conservative, estimate. Collins and Lapsleys Australian estimates may provide a rough order of magnitude of nursing home expenditure related to harmful AOD use for New Zealand. (Collins and Lapsley 2008), however, do not estimate treatment costs. To do this, we have adjusted their figures for differences in population size, drug use prevalence, timing and currency (our figure is reported in NZ$2005/06 terms). This estimate assumes behaviour, harm and nursing home use is otherwise the same in both countries. Alcohol-related nursing home expenditure could amount to $124.6 million and $1.9 million for illegal drugs. Adding this estimate of possible AOD-related nursing home expenditure could increase tangible costs by just over 3.6 percent. Alcohol and other drug treatment costs, which are estimated here but are not included in Collins and Lapsley, equate to approximately 1.8 percent of tangible costs. Therefore, although the range of costs included in the health component differs between the studies, the differences are likely to broadly balance out.
47
drug use. Therefore the costs for community units, residential beds, and medical detox beds were allocated to the joint AOD category. The methadone placement program costs ($11.7 million) were solely attributed to other drugs. Ambulance services Estimates of other drug-related ambulance service use were based on information provided by the St John Ambulance Service.
57
Data on illegal drugs was determined from records on
drug-related overdose, ingestion and poisoning. They do not cover other drug-related callouts such as falls or violent assaults. Therefore, the estimates are likely to underestimate actual ambulance service use. St John Ambulance Service also carried out a text-based search of callout records on alcohol-related ambulance callouts specifically for this project. However, the way the data is recorded means that the identified records underestimate the number of callouts due to alcohol-related incidents that St John Ambulance Service attended. Instead, ACC data on ambulance callouts attributed to alcohol-related injuries were used. However, this approach is also likely to result in an underestimate, as it only captures ambulance callouts where an injury was sustained and reported to ACC. We used the St John Ambulance Service data to generate a conservative estimate of the number of people that would not usually have used the ambulance service (and A&E) due to premature mortality. This yields a cost offset to the gross cost of ambulance (and A&E) services. Accident & emergency care A&E use was based on the number of people transported to hospital for drug-related treatment. The unit costs were based on Fielden et al (2005), using costs inflated to 2005/06 dollar terms. However, this approach does not capture A&E presentations that do not appear in the ACC or St John Ambulance Service records. Therefore, this approach is likely to underestimate the true cost. Hospital care Hospital care costs were calculated for inpatient treatment using a dataset of publicly-funded hospital discharges in New Zealand over the six years from 2001 to 2006. Discharges that
57
The St John Ambulance Service is not a national provider; it serves approximately 85 percent of New Zealands population (personal communication). However, no data is available for other providers. We scale up the estimates based on the St John Ambulance Service figures to allow for this omission. We have no reason to expect the pattern of ambulance service use in places such as Wellington, where the Wellington Free Ambulance service is the major provider, to differ substantially from those in areas served by the St John Ambulance Service.
48
were wholly or partly attributable to either the harmful use of alcohol, or other drug use, were based on information sourced from Collins and Lapsley (2008).
58,59
Each discharge has a case-weight attributed to it, giving the relative complexity of each case. To these case-weights were applied attributable fractions, from which numbers of cases attributable to AODs were derived. The average cost of a case for each year was applied to these numbers to give an approximation of total treatment cost. Some health care services for drug-related conditions are provided in an outpatient setting. The NZHIS has some data on outpatient treatment in its National Non-Admitted Patient Collection dataset. However, this data is organised by purchaser code rather than by diagnosis, so it is not possible to directly use this data to estimate drug-attributable outpatient costs. As such, the hospital cost estimates are likely to be conservative. In section 5.1, we note a rough estimate of outpatient costs similar to the proportion of total inpatient care that is AOD-related. This figure provides a sense of the possible magnitude of AOD-related outpatient costs, but we have not included it in the main estimates as we are concerned about its reliability. In the case of fetal alcohol syndrome disorder (FASD), we have included the direct costs of inpatient care in 2005/06 based on data from NZHIS. Although there is evidence to suggest that there are ongoing impacts of such neonatal disorders, we have not estimated these costs as there is currently insufficient information for these impacts in a New Zealand setting.
60
4.3.5 Road crashes Road crash costs were estimated from data recorded in the Crash Analysis System (CAS), as well as the cost numbers from various sources, such as the New Zealand Fire Service, New Zealand Transport Agency, and the Insurance Council of New Zealand.
58
The set of conditions is identified by the ICD-10 diagnosis code and, in the case of some injuries, by external cause codes.
59
We have included fetal alcohol syndrome (FASD), ICD-10 code Q86.0, even though it does not appear on Collins and Lapsleys list of alcohol attributable conditions. It seems reasonable to assume that this condition is fully attributable to alcohol misuse.
60
The cost per child, annual incidence of new cases, and annual prevalence of FASD in New Zealand is uncertain (Curtis et al 1994, Leversha and Marks 1995, Alcohol Healthwatch 2007). To provide a sense of magnitude of the social costs of FASD in New Zealand, we draw on Stade et als (2006) examination of the annual social cost of children with FASD in Canada. First we convert their figure per child of CA$13,109.57 to 2005/06 New Zealand dollar terms, allowing for purchasing power differences. While there were four cases of FASD recorded in the NZHIS data in 2005/06, we use a 6-year annual average rate of 2.17 to estimate the number of new cases per year. As State et al (2006) focused on people up to the age of 21, we use this in calculating the annual prevalence of FASD-affected people. That is, we assume that there are approximately 46 people up to age 21 that are affected by FASD. This is likely to be a conservative estimate. On this basis, however, FASD could have resulted in social costs of just over $885,000 in health, education and lost output costs in 2005/06.
49
The number of drug attributable road crash fatalities identified in CAS was scaled down, by drug type, to match those identified from the NZHIS data (see section 4.3.4). Table A.2 of the Ministry of Transports (2006) social cost analysis has scaling factors to allow for under reporting of non-fatal accidents, but it does not give a breakdown by drug type. Therefore, while the estimates of non-fatal accidents are based on reported CAS figures by drug type, we apply the same scaling factor to all serious and minor crashes regardless of drug type. The scaling factors we use for serious and minor accidents are 1.71 and 3.74. The research calculated a similar range of road crash cost components as in Collins and Lapsley (2008). This section focuses on five components, and excludes those components that were estimated elsewhere (health care, justice sector, and intangible costs) to avoid double counting. These five not elsewhere included components consist of loss of output, property damage, travel delays, insurance administration, and fire/emergency service. 4.3.6 Intangible costs from loss of life and lost quality of life The intangible costs of mortality in the 2005/06 year were derived from estimates of alcohol and drug related mortality as described in section 4.2.3, which includes health and road crash-related deaths plus AOD-related homicides. These estimates gave the additional population that would be alive in 2005/06 in the absence of harmful AOD use. The value of a life year for the additional population was estimated as being $106,600.
61
The additional population number was multiplied by a value per life year for a fatal injury to give an estimated value for loss of life of $1,906.5 million, of which $1,519.9 million was for alcohol, $386.5 million was for other drugs, and $380,000 was due to joint AOD use. Lost quality of life was calculated for AOD road crash-related injuries and for victims of crime. Serious road crash injuries were valued at 10 percent of the value of a life year, and minor injuries at 0.4 percent. These values were applied to the estimated number of AODrelated road crash injuries. Similarly, lost quality of life suffered by victims of crime was calculated using the estimated number of AOD-related offences and the intangible cost per offence based on the Treasurys (2006) report on the costs of crime. 4.3.7 Avoidable cost analysis Avoidable costs are the proportion of total costs that could, in principle, be changed given the implementation of appropriate public policies and behaviour changes.
61
The statistical life year is derived from the Ministry of Transports VOSL. For more information, refer to section 9.4.6 on intangible costs.
50
An initial study (Collins and Lapsley 2002) provided a breakdown of avoidable costs to social costs for alcohol (tangible and intangible). This breakdown did not cover the full range of cost components and has been superseded by further research (Collins and Lapsley 2008). The 2008 report uses more recent data and we believe it is a more accurate measure of avoidable costs. However, only alcohol avoidable costs are covered. The report suggests that, of the four methods of estimating avoidable costs, the most appropriate is the intervention effectiveness approach.
62
However, the studys avoidable cost estimate uses a
mix of two other approaches (the Arcadian Normal and exposure-based comparators). This approach reflects the difficulty in estimating in aggregate using intervention effectiveness, which provides avoidable costs per individual intervention. Collins and Lapsley estimated that up to 50 percent of total social costs of alcohol in Australia are avoidable. This study draws on Collins and Lapsleys most recent estimate. That is, we assume that 50 percent of total costs are avoidable. However, this proportion should be viewed as an upper boundary or feasible maximum of potential avoidable costs. As such, the estimate is illustrative rather than a robust figure for the New Zealand context. Table 6.2 shows the avoidable costs for alcohol and illegal drugs. We recommend further research should be done into providing a robust measure of avoidable costs in a New Zealand context. 4.3.8 Injury cost analysis This sub-analysis partitioned off those social costs due to injury resulting from harmful AOD use. This primarily included: health care for victims of crime and people injured as a result of their AOD use lost output due to injury-related mortality, absenteeism and reduced productivity road crash costs related to injury, such as lost output and medical care for crash victims intangible costs such as homicide, road crash and other fatal injuries and reduced quality of life for road crash survivors. The injury cost estimates above are mainly a sub-set of the main analysis components. That is, we identified the quantum of lost output due to injury-related premature mortality, justice and health sector costs, road crash and intangible cost estimates that were injury-related.
62
These methods are detailed in section 3.3.4.
51
One component prepared specifically for this analysis based on ACC data, and drawn into the main cost estimates, was the value of lost output due to absenteeism and reduced productivity for injured workers.
63
ACC data and attribution to harmful AOD use ACC provided data to BERL on the number and cost of both new claims lodged and ongoing claims in 2005/06, as well as public health acute service (PHAS) funding. The $1,973 million of claims were broken down by the seven New Zealand Injury Prevention Strategy (NZIPS) categories: assault, drowning, falls, motor vehicle, other, suicide and self harm, and workplace. This data covered expenditure on ambulance services, elective hospital and medical treatment, benefit and lump sum payments, as well as support for vocational rehabilitation and weekly workers compensation. The PHAS funding of $315.0 million in 2005/06 covered acute inpatient, emergency department, outpatient, other miscellaneous hospital services, cost adjusters, pharmaceuticals and laboratory services. The analysis of this data concentrated on acute inpatient and emergency department funding as discussed below in the sub-section on ACC costs. Section 6.2.1 also discusses rough estimates for PHAS-funded outpatient, pharmaceuticals or laboratory services to provide a sense of their possible magnitude; section 9.4.8 details our approach. Lost output Output lost as a result of workers being incapacitated or impaired by AOD-related injuries was based on the workers compensation payments category. AOD attributable fractions were applied to the relevant NZIPS categories to determine the portion of total ACC workers compensation payments that were due to AOD-related injuries. The ACC figure was then scaled up, as ACC does not compensate workers for 100 percent of lost income, so the social cost of the lost output is greater than the ACC compensation figure. the output lost in the affected workers first week was also added. In addition to the social costs of AOD-related injury, we estimated the portion of injury and social costs borne by ACC. This followed a similar process to that above of applying
64
An estimate of
63
The main absenteeism estimate reflects lost production due to people taking time off as they are unwell rather than injured.
64
Workers are not entitled to ACC compensation in their first week. However, the output lost in this week represents a social cost, and this burden may be carried by employers as sick leave. The resulting estimate is likely to be conservative, however, as we were not able to estimate the value of output lost by injured workers who do not claim ACC workers compensation, for example, people affected for less than one week who do not report t heir injury.
52
relevant attributable fractions to other NZIPS categories. As the NZIPS categories are not explicitly coded with the ICD system, this necessarily involved qualitative matching. For example, the NZIPS assault category is aligned with the fractions for the group of codes related to assaults in Collins and Lapsley (2008). ACC costs In addition to the sub-analysis of AOD-related injury costs, we examined which of these costs would be likely to be borne by ACC. AOD attributable fractions were applied to the relevant NZIPS categories to determine the portion of total ACC funding due to harmful AOD use. This analysis covered expenditure on ambulance services, elective hospital and medical treatment, benefit and lump sum payments, as well as support for vocational rehabilitation and weekly workers compensation. It was not possible to allocate the PHAS funding categories to the NZIPS categories, nor to use the AOD attributable fractions to determine the portion attributable to harmful AOD use. Instead, we concentrated on acute inpatient and emergency department funding. 4.3.9 Costs to government Harmful drug use impacts on both government expenditure and revenue. An indirect cost is that tax revenue from individuals and business is lower as a result of reduced output.
65
Higher drug-related unemployment may also increase unemployment benefit expenditure. GST revenue is reduced due to lower expenditure due to premature mortality and expenditure diverted to purchase illegal drugs. Eliminating harmful use, however, may reduce excise and customs tax collected on alcohol sales, which represents a tax offset. The sections below detail how we estimated the impacts on government expenditure, the revenue impacts from a reduced population and reduced output, and the impacts on consumption-related tax revenue.
65
Although we do not provide a full analysis of the costs of harmful AOD use to the business sector, we include analysis of some of the costs to business of lost output in Appendix Table 24, as this is a component in estimating the costs to the government.
53
Impacts on government expenditure Government expenditure is estimated for the justice and health sectors from budget Vote
66
information and BERLs main estimates, and for emergency services attending drug -related road crashes. Impacts of lost output Estimating the labour market-related costs from a government perspective involved the following elements. Premature mortality individual and business tax lost less unemployment benefits saved Excess unemployment individual and business tax lost plus unemployment benefits incurred Absenteeism individual and business tax lost Reduced productivity business tax lost Lost output from victims of crime
The main estimates of lost output for each of these components were used to estimate the lost tax revenue. Total lost output was decomposed into wage and residual value added components, and average personal and business tax rates were applied to these figures. The impact of drug use on unemployment benefits was calculated by taking the number of people affected times the average annual unemployment benefit payment based on Census data. In the case of premature deaths, the calculated impact represents a cost saving in that these payments did not have to be made. In the case of excess unemployment, benefit payments to people who would have been in employment if not for their harmful drug use represents a cost to government. Appendix Table 24 and Appendix Table 25 provide more detailed breakdowns of the lost output impacts on business and government.
67
66
The term Vote refers to funding approved by parliament for a specified range of outputs and that is the responsibility of a particular government Minister (the Vote Minister).
67
Lost revenue is likely to be a substantial cost to businesses. It was not possible to determine the share of crime, road crash and other costs borne by businesses, so we do not provide a separate business perspective on costs.
54
Impacts on consumption-related taxes In addition to reducing output, premature mortality prevents people spending their income, which would have generated GST revenue. This reduction in GST revenue was based on the main estimate of the reduction in consumption expenditure due to premature mortality. Estimated expenditure on illegal drugs was used to calculate GST revenue lost by expenditure being diverted from taxable, market activity to illegal purchases. We assume that under the counterfactual harmful alcohol expenditure would be diverted to other market expenditure so there is no net impact on GST from such a change. Reducing harmful alcohol use reduces the quantity of alcohol liable for customs and excise duties. This impact is estimated by determining the reduction in alcohol consumed if harmful drinkers were to reduce consumption to the low risk level (as a percentage of total consumption). This percentage is applied to the amount of alcohol excise tax and customs duty collected to calculate the reduction in excise and customs revenue.
55
Results: the Costs of Harmful AOD Use
This section summarises the costs of harmful drug use in 2005/06. The first section presents the overall costs broken down by those costs attributed to alcohol, to other drugs and joint costs that could not be separately estimated. Sections 5.2 and 5.3 reallocate the joint AOD costs according to the relative share of alcohol only and other drug only costs to provide an indication of the total costs of alcohol versus other drugs. 5.1 Costs of harmful alcohol and other drug use overall
Overall, harmful drug use in 2005/06 caused an estimated $6,525 million of social costs. This comprises $4,562 million of tangible resource costs (2.9 percent of GDP in 2005/06) and $1,963 million of intangible welfare costs.
68,69
The tangible costs were equivalent to two
thirds of the GDP ($7,209 million) of New Zealands health services industry in 2005/06. Table 5.1 summarises the tangible resource costs and intangible lost welfare caused by harmful drug consumption in New Zealand in 2005/06. Table 5.1 Social costs of harmful drug use ($m), 2005/06
Source: BERL
Alcohol 2,875.1 1,561.9 4,437.1 68.0%
Other drugs Joint AOD 1,034.2 392.4 1,426.7 21.9% 652.1 8.7 660.8 10.13%
Total 4,561.5 1,963.1 6,524.6 100.0%
Harmful alcohol use in 2005/06 cost New Zealand an estimated $4,437 million of harm, or around two thirds of the social costs. This is equal to almost two fifths of Vote Health in 2006, where the tangible costs were equal to over one quarter of Vote Health. Other drug use accounted for $1,427 million (one fifth of social costs). The total is over half of the justice sector Vote funding (Justice, Customs, Police, Courts, Corrections), while the tangible costs are almost two fifths of this Vote funding.
68
The estimates are GST exclusive figures. GDP is measured at 'market prices', which includes indirect taxes such as GST. Therefore the estimates are not directly comparable with GDP. The percentage figure is indicative of the relative magnitude.
69
Excluding the non-market labour cost estimates, the tangible market cost of harmful AOD use was 3.05 percent of GDP. See footnote 61 for discussion of non-market contributions and their relation to a countrys GDP.
56
Splitting up the joint costs of AOD use ($661 million, nine percent of social costs), total alcohol-related costs equate to $4,939 million (76 percent) and other drug-related costs to $1,585 million (24 percent). Table 5.2 breaks down the $4,437 million of tangible resources diverted by harmful drug by cost component; Appendix Table 2 provides further detail. The three largest tangible cost drivers for alcohol were labour costs, drug production and justice sector costs, which accounted for 84 percent of the tangible costs of alcohol. Similarly, drug production, crime and labour costs accounted for 92 percent of the tangible costs of other drug use. Given an estimated 513,000 harmful alcohol users, 27,000 other drug users and 127,000 joint AOD users, harmful drug use cost approximately $9,800 per user, where over 70 percent of these impacts represented tangible resource costs. Table 5.2 Tangible costs of harmful drug use ($m), 2005/06
Tangible costs ($m) Labour costs Crime n.i.e. Drug production Health care Road crashes n.i.e. Total tangible costs % of tangible costs Alcohol 1,478.3 562.2 342.2 290.2 202.1 2,875.1 63.0% Other drugs Joint AOD 157.6 309.9 518.7 43.2 4.9 1,034.2 22.7% 315.7 239.3 0.0 94.8 2.4 652.1 14.3% Total 1,951.6 1,111.4 861.0 428.2 209.4 4,561.5 100.0%
Note: n.i.e. denotes not included elsew here. Source: BERL
Lost output ($1,952 million), crime ($1,111 million) and drug production ($861.0 million) were the largest resource drains overall. Drug-attributable health care and road crashes not included elsewhere (n.i.e.) cost a further $428 million. Table 5.3 shows the estimated intangible costs of harmful drug use; Appendix Table 3 provides a more detailed breakdown. Intangible costs result from the pain, disability and loss of life resulting from the impacts of AOD use. Table 5.3 Intangible costs of harmful drug use ($m), 2005/06
Intangible costs ($m) Loss of life Lost quality of life Total intangible costs % of intangible costs
Source: BERL.
Alcohol 1,519.6 42.3 1,561.9 79.6%
Other drugs 386.5 6.0 392.4 20.0%
Joint AOD 0.4 8.3 8.7 0.4%
Total 1,906.5 56.6 1,963.1 100.0%
57
Drug users and victims of drug-related crime and road crashes suffered a further $1,963 million of intangible costs (28.5 percent of total social costs). Loss of life was the main contributor to overall intangible costs. Premature mortality n.i.e provided the largest amount to this total, with $1,452 million. Loss of life due to road fatalities caused $452 million of harm , and homicides $2.9 million of intangible costs. The pain and disability caused by the use of drugs and alcohol associated with road crashes and crime was estimated as being equivalent to $56.6 million in 2005/06. The main component of this total was crime, which caused $47.5 million worth of harm. Road crashes imposed an intangible cost of $9.1 million. Lost output Lost output is an indirect cost of harmful drug use, as harmful drug use reduces the amount and quality of labour resources. The resulting net output loss cost $1,952 million in 2005/06. Appendix Table 16 details the component costs of lost output. The total includes estimated gross output losses of $1,937 million of market output, $71 million of output lost in the unpaid household sector and $330 million of crime and road crash-related lost output. The cost of lost output was the largest social cost component, and was around one third of all social costs of drug use. The largest cause of lost output was for excess unemployment, at $1,104 million, over half of all labour costs. The reduction in the workforce due to premature mortality cost $600 million in 2005/06, of which almost 80 percent was for males. Premature mortality led to saving consumption resources of $388 million, of which the majority was associated with people engaged in the labour market. Without this saving, total lost output costs for drug use would have reached $2,340 million. Justice sector Tangible costs of crime made a significant proportion of social costs related to drug use. Appendix Table 17 provides further detail on the justice sector figures.
70
70
According to NZHIS data, approximately 25 percent of alcohol-related fatalities, and 17 percent of other drugrelated fatalities, are caused by road accidents. Hence, loss of life due to road fatalities is a significant portion of intangible costs.
58
In 2005/2006, total crime cost was $1,111 million, 16.2 percent of all social costs. Of this number $562 million was for alcohol (just over one half), $310 million was for other drugs (just over one quarter) and $232 million (just over one fifth) was for joint AOD use. The largest crime expenditure related to drug use was for Police expenditure, at $306 million. Police time spent dealing with drug-related driving offences accounted for just over one quarter of Police expenditure on harmful drug use. Around one quarter of drug-related Police expenditure was related to specific drug offences. Prison expenditure related to drug use was $296 million. Around 38 percent of AOD-related prison expenditure was related to alcohol use, 30 percent for other drug use, and 32 percent for joint AOD use. Lost property was another significant component within crime costs. Lost property cost $223 million. Theft due to alcohol and drug use was the largest contributor to this cost. Drug production Diverted inputs into harmful AOD production cost $861.0 million in 2005/06. Of this figure, just under two fifths ($342 million) was for diverted inputs to alcohol, and just over two fifths ($519 million) was for other drugs. Health care We calculate that in 2005/06, net healthcare costs related to AODs were $428 million, of which $290 million was for alcohol, $43 million was for other drugs, and $95 million was for both drugs and alcohol combined. Appendix Table 18 provides further detail on the health sector figures. The largest cost of healthcare was for inpatient treatment of harmful drug users, at $130.3 million. This would have been $137.3 million if it were not for an estimated $7.0 million savings due to premature mortality. Appendix Table 19 and Appendix Table 20 detail the number of cases attributed to harmful alcohol and drug use between 2001 and 2006, and the corresponding annual average costs. Health care for victims of AOD-related crime cost $123.7 million. Specialist drug treatment services in the community were estimated to cost approximately $92.9 million. As substance users cared for by these services are typically poly-drug users, it was not generally possible to separately allocate these costs to alcohol or other drugs.
59
Out-of-pocket expenditure and government subsidisation of AOD-related GP services is estimated to amount to a net cost of $52.0 million. Other primary care services for harmful AOD users cost an estimated $18.7 million, including $17.4 million for drug and alcohol workers. Drug-related A&E/non-hospital admission cost $5.7 million, and drug-related ambulance services cost $4.7 million.
71
The latter estimate is based on the ACC data for people
affected by alcohol and St John Ambulance Service data for people transported to hospital for other drug-related treatment, and is therefore likely to be an underestimate. To provide a sense of the possible magnitude of AOD-related outpatient funding, we made a rough estimate based on the proportions, by drug type, of total inpatient care that were AODrelated and total outpatient funding. Total public outpatient funding in 2005/06 was $1,733.6 million. Our estimates indicate that $76.8 million of outpatient care was alcohol-related and $6.9 million was other drug related, or approximately 4.8 percent of outpatient funding in total was AOD-related. We have not included these estimates in the main social cost estimates as we are concerned about their reliability. Road crashes The social costs of road crashes were $778 million for year 2005/06. A number of road crash costs are reported in other cost components, such as health care or lost output. Those costs that are not included in other cost components amounted to $209 million. The majority of these costs not included in other components were due to harmful alcohol use, accounting for around 97 percent of the total costs not included in other components. Appendix Table 21 details the road crash cost estimates. Approximately 2.3 percent of the total costs n.i.e. were due to other drug use, totalling $4.9 million in 2005/06; joint AOD use cost around $2.4 million (1.1 percent) over the year. The intangible costs of AOD-related road crashes were estimated at $461.0 million, and are recorded as part of total intangible costs discussed below. Intangible costs from loss of life and lost quality of life Loss of life and lost quality of life due to AODs imposed an intangible cost of $1,963 million. Of this total, $1,561 million was due to alcohol, and $393 million due to illegal drugs.
71
For harmful users of alcohol only we estimated use of drug and alcohol workers. As better information is available for other drug users, we include estimates for this group of the use of counsellors, drug and alcohol workers, social workers, psychologists, and psychiatrists.
60
5.2
Costs of harmful alcohol use
This section reallocates the joint AOD costs according to the relative share of alcohol only and other drug only costs by individual component. For example, alcohol accounts for approximately 76 percent of total tangible costs attributed to alcohol only and other drugs only. Therefore, approximately 76 percent of the tangible costs of harmful joint AOD use are added to the alcohol only cost estimates to illustrate the full other drug costs below.
72
Table 5.4 illustrates the full costs caused by harmful alcohol consumption in New Zealand in 2005/06. Table 5.5 details the components that make up the tangible alcohol costs. Table 5.4 Social costs of harmful alcohol use ($m), 2005/06
Source: BERL
Alcohol 3,369.8 1,569.5 4,939.3 75.7%
Total 4,561.5 1,963.1 6,524.6 100.0%
Table 5.5 Tangible costs of harmful alcohol use ($m), 2005/06

Tangible costs ($m) Labour costs Crime n.i.e. Drug production Health care Road crashes n.i.e. Total tangible costs % of tangible costs Alcohol 1,763.6 716.5 342.2 343.0 204.5 3,369.8 73.9% Total 1,951.6 1,111.4 861.0 428.2 209.4 4,561.5 100.0%
72
This example is illustrative only. The reallocation is performed at the sub-component level, rather than using the total. For example, AOD-related Police expenditure is estimated to be $306.3 million. Of this total $43.5 million is attributed to joint AOD use, while $172.2 million was attributed to alcohol alone. This latter figure is approximately 65.5 percent of the sum of the alcohol category plus the other drug category. We reallocate $28.5 million (65.5 percent of the joint AOD costs) to give the total Police costs of harmful alcohol use of $200.7 million. This figure is combined with other justice sector sub-components to give the Crime n.i.e. total of $716.5 million in Table 5.5. We make an exception to this method in the case of community-based substance abuse treatment services. People with alcohol or other drug abuse disorders are typically poly-drug users (Ministry of Health 2006). No costs were estimated as solely attributable to alcohol use for this component, so the process above would result in a zero estimate for alcohol-related use of these services. The Ministry of Healths (2006) mental health survey notes Alcohol disorders were more common than drug disorders but were less likely to be classified as serious, and that the rates of service contact for people with alcohol disorders is lower than for drug disorders. So it is unclear whether alcohol users would have higher or lower use of community-based treatment services than other drug users. Therefore, the joint AOD cost was split equally for this sub-component. For total other drug costs, this share was combined with preventative services related specifically to other drug use, such as methadone treatment costs.
61
Based on Table 5.4, total alcohol costs could equate to $4,939 million. This would make total alcohol costs equivalent to two thirds of the output of New Zealands food, beverage and tobacco industry. Lost output Labour costs for lost output due to harmful alcohol use cost $1,764 million in 2005/06. This was slightly less than one half of all tangible costs for alcohol and one third of all social costs. The largest cause of alcohol lost output was for excess unemployment, at $1,105 million, around three fifths of alcohol labour costs. The reduction in the workforce for alcohol-related premature mortality was the second largest component of alcohol labour costs, at $459.2 million. Over 80 percent of alcohol mortality costs were for males. Premature mortality saved consumption resources of $308.4 million, of which the majority were for people who would have been in the labour market. Without this saving, total lost output costs for alcohol use would have reached $2,072 million. Justice sector We calculate that in 2005/2006, tangible crime costs n.i.e. related to alcohol use only was $716.5 million. This is just under two thirds of all crime costs related to harmful drug use. The largest part of this total was $200.7 million of Police costs, or just over quarter of alcohol crime costs n.i.e. The cost of property losses due to alcohol-related crime was $170.1 million. Just under one quarter of Police costs related to alcohol crime n.i.e. was for activities related to property loss or damage. The third largest crime cost n.i.e. was $166.1 million on prisons in 2005/2006 for alcohol-related offences. Drug production Resources diverted to manufacture alcohol that is used harmfully cost $342.2 million in 2005/06. Health care We calculate that in 2005/2006, net healthcare costs related to alcohol use only was $343 million.
62
The largest component of this number was inpatient hospital costs of $121.0 million, or around a third. The greatest hospital-related costs for alcohol came from occupational and machine injuries, fall injuries, road injuries, and self harm. Another significant contributor to healthcare costs was the treatment of victims of alcoholrelated crime, at $113.2 million. Most victims suffered from violent offences. As noted above, it was not possible to separately identify community care costs related to harmful drug use, as most people receiving care are poly-drug users. We apportion $40.6 million of the total AOD service treatment costs to total alcohol costs. Health care costs related to harmful alcohol use would have been $358.7 million, were it not for $13.5 million savings caused by premature mortality. Road crashes The total social costs of road crashes related to alcohol were estimated at $700.4 million in 2005/06. The total tangible costs of alcohol-related road crashes were estimated as $307.2 million. A significant amount of these costs are captured in other cost components. For example, the total tangible cost figure includes health care costs of $14.6 million, which is recorded in the health care component in the previous section. Those costs not included elsewhere (n.i.e.) in other cost components amounted to $204.5 million (two thirds of total alcohol-related road crash costs). The three largest cost components n.i.e. were insurance administration, travel delays and property damage, and cost $133.7 million, $27.1 million and $22.0 million respectively. A large proportion of the total alcohol-related road crash costs (54.9 percent, $384 million) were due to loss of life. Lost quality of life attributed to alcohol misuse was estimated to cost a total of $8.9 million over 2005/06 (1.3 percent of the total alcohol-related road crash costs). The intangible road crash costs are recorded in the intangible costs section below. Intangible costs from loss of life and lost quality of life In the 2005/2006 year, 14,250 additional people would have been alive had there been no harmful use of alcohol. Assuming a value of a life year of $106,600, the value of these deaths is estimated at $1,520 million. Costs were broken down into three categories, of which premature mortalities (excluding road fatalities and homicides), cost $1,133 million, road fatalities cost $384 million, and homicide cost $2.3 million.
63
5.3
Costs of harmful other drug use
This section reallocates the joint AOD costs according to the relative share that other drug costs represent of the alcohol only and other drug only costs. For example, other drug costs account for approximately 24 percent of total tangible costs attributed to alcohol only and other drugs only. Therefore, approximately 24 percent of the tangible costs of harmful joint AOD use are added to the other drugs only cost estimates to illustrate the full other drug costs below.
73
Table 5.6 shows the full costs that result from the harmful consumption of other drugs in New Zealand during 2005/06. Table 5.7 details the corresponding distribution of tangible costs by component. Table 5.6 Social costs of harmful other drug use ($m), 2005/06
Source: BERL
Other drugs 1,191.7 393.6 1,585.3 24.3%
Total 4,561.5 1,963.1 6,524.6 100.0%
Table 5.7 Tangible costs of harmful other drug use ($m), 2005/06
Tangible costs ($m) Labour costs Crime n.i.e. Drug production Health care Road crashes n.i.e. Total tangible costs % of tangible costs Other drugs 188.0 394.9 518.7 85.2 4.9 1,191.7 26.1% Total 1,951.6 1,111.4 861.0 428.2 209.4 4,561.5 100.0%
Based on Table 5.6, total other drug costs could amount to $1,585 million. This would make total other drug costs equivalent to one percent of New Zealands total market output (GDP).
73
This example is illustrative only as the reallocation is performed by component, rather than using the total.
64
Lost output Labour costs for lost output due to other drug use cost $188 million in 2005/06. This was around 16 percent of all tangible costs for other drugs, and 4 percent of all tangible social costs. The largest cause of other drug lost output was for the reduction in the workforce due to premature mortality, at $141 million. Around three quarters of other drug mortality costs were for males. Premature mortality led to consumption resources of $80 million being saved; the majority of this was related to market labour. Without this saving, total lost output costs for other drug use would have been $268 million, or 42 percent higher. Lost output of prisoners incarcerated for other-drug-attributable offences cost $87.9 million, or over one third of the equivalent for alcohol ($242 million), despite total labour costs due to other drug use being one-eighth of alcohol labour costs. Because of the illegal nature of most other drugs examined, there is a greater incarceration rate, and hence, greater lost output due to imprisonment. Drug production Resources diverted to other drug production cost $518.7 million in 2005/06. This was around 44 percent of all tangible costs for other drugs, and 10 percent of all tangible social costs. The largest costs were for cannabis production ($243.2 million), of which over 99 percent is produced locally, and amphetamines ($230.9 million), of which around 80 percent is imported. Justice sector We calculate that in 2005/2006, tangible crime costs n.i.e related to other drug use only was $395 million. The largest component of this number was the prison cost for other drug related offences, at $129.7 million, or around one third of tangible other drug crime costs n.i.e. Theft and property damage were also significant factors, costing $52.6 million. Another substantial contributor to the cost of drug crime was Police costs, at $105.7 million. This mainly driven by Police resource diverted by specific drug offences and other offences, such as burglary or theft. Health care We calculate that in 2005/2006, net healthcare costs related to other drug use only was $85.2 million.
65
Contributing to this total was pharmaceutical treatment of $11.8 million. This included Pharmac subsidised over-the-counter treatments, including restricted medicines such as methadone ($1.5 million). Other pharmaceutical treatments included Hepatitis C treatments ($7.9 million). Community-based AOD treatment (methadone programmes) cost $52.3 million, of which $11.7 million was for methadone treatment. With the exception of methadone treatment, the reallocated total figure should be interpreted as primarily illustrative, as these services tend to serve poly-drug users. Hospital costs were $11.3 million. The main drug-related conditions treated were road injuries, drug psychoses, self harm, and accidental poisoning by psychostimulants. Health care costs attributed to both alcohol and other drugs are covered in section 5.1. Road crashes As road crashes attributed to other drugs accounted for a small portion of the total road crashes in 2005/06, the costs associated with it were relatively small. Over the period 2005/06, other drug-related road crashes caused $86.2 million of tangible social costs, of which $4.9 million were not included in other cost component estimates. Due to the smaller number of fatal road crashes attributed to other drugs, we estimate that the total cost of lost life was approximately $67.5 million for 2005/06, about one seventh the level for alcohol-related road fatalities. Intangible costs from loss of life and lost quality of life In the 2005/2006 year, 3,620 additional people would have been alive had there been no use of illegal drugs causing premature death. Assuming a value of a life year of $106,600, the value of these deaths is estimated at $386.5 million. The three sub-categories of intangible costs are estimated at $318.8 million of premature mortalities (excluding road fatalities and homicides), $67.5 million of road fatalities and $0.2 million of homicide-related impacts.
66
6
6.1
Additional Analytical Focuses

Avoidable costs
Estimates of the total costs of drug misuse comprise avoidable and unavoidable costs. Avoidable costs are the proportion of total costs, or the burden of misuse, that could in principle be changed given the implementation of appropriate public policies and behaviour changes. Collins et al (2006) argue in the International Guidelines for the Estimation of the Avoidable Costs of Substance Abuse that identifying the social costs of substance misuse involves estimating the relevant avoidable proportion of each cost category and applying these proportions of avoidable cost to the relevant aggregate cost estimates. The International Guidelines argue that avoidable costs are the potential economic benefits (i.e. costs avoided) from substance misuse harm minimisation strategies. These impacts can be used in policy to determine the appropriate level of resources that should be devoted to these strategies (Collins et al 2006). An initial step in estimating avoidable cost is to determine the proportion of a disease in a given population that is due to specific past and present exposure to one or more risk factors (Rehm et al 2006). This proportion of the disease, it is argued, could be reduced or eliminated if the exposure to the risk was reduced or eliminated and is conceptualised even if the reduction is not achievable in practice. An attributable fraction is used to estimate this proportion. However, attributable fractions do not include a time dimension and do not adequately incorporate changes in underlying exposures and disease factors over time. As such, a feasible minimum is used to estimate the avoidable burden of a disease in a society. However, an Arcadian normal is an approximation for a given place and time period only, and is based on disease. It is not exposure specific. Therefore it might be better not to formally define a feasible minimum burden of disease but model shifts in the risk factors in the direction of the theoretical minimal burden. This would involve looking at the distributional changes resulting from intervention packages and modelling avoidable burden accordingly. Feasible minimums can also be estimated using WHO data on The feasible minimum level of disease may also be termed the Arcadian normal, and relates to t he minimum level of mortality for a range of conditions. drug-attributable fractions, morbidity and mortality. This data can be used to identify best performance among countries that WHO has grouped according to common characteristics. However, the International Guidelines for the Estimation of the Avoidable Costs of Substance Abuse caution that this method is highly
67
imprecise given the assumptions that are made when comparing countries. Countries may be so different that they may not be comparable as regards achievable disease reduction. The WHO study does not aim to compare the prevalence rate of alcohol misuse between countries, but the International Guidelines argue that this method of comparison allows estimation of the percentage of avoidable burden and therefore avoidable costs in a country. This approach is still consistent with an approach that focuses on exposure to a risk rather than outcomes. To calculate avoidable costs a counterfactual situation needs to be estimated where there is little or no alcohol misuse. This way the extra costs this misuse imposes on a community can be calculated. Hence, avoidable costs are those which could not have been incurred if there had been no, or minimal, problems associated with alcohol misuse. Some avoidable costs of alcohol misuse are associated with acute harm, such as injuries from road crashes. Avoidance of these harms would result in immediate and longer-term savings. Treatment, law enforcement and some productivity costs are also avoidable costs. Avoidable costs are therefore those public resources that are directed into the prevention or reduction of substance misuse and are consequently unavailable to the community. We used Collins and Lapsley (2008) to estimate avoidable costs.
74
The 2008 report was the
only information source found that measured avoidable costs at an aggregate level (instead of individual interventions), while providing an adequate methodology. However, only alcohol avoidable costs are covered. That report provides the basis for our estimate of avoidable costs for both alcohol and other drugs. The Collins and Lapsley report suggests that, of the four methods of estimating avoidable costs, the most appropriate is the intervention effectiveness approach, where there are a growing number of studies, such as Rehm (2006). However, their total avoidable costs estimates are based on the Arcadian Normal and exposure-based comparators. Table 6.1 Potential avoidable alcohol consumption and mortality in Australia, 2004/05
Measure Alcohol consumption per capita Total alcohol-attributable deaths
Source: Collins and Lapsley (2008)
Approach taken Exposure-based Arcadian Normal
Potential reduction 40% 48%
74
Collins and Lapsley (2002) provided estimates of the avoidable portion of social costs for alcohol (tangible and intangible). That study estimated avoidable costs for a range of components and in greater detail than the level used in this report. Components were estimated to have avoidable costs ranging from 12 to 90 percent. However, the study excluded illegal drugs, had some gaps, and a detailed methodology was not shown. Therefore, the newer 2008 Collins and Lapsley study was used in favour of the 2002 study.
68
Collins and Lapsley (2008) does not provide breakdowns of avoidable costs at lower levels, for example, measuring the amount of alcohol crime that is avoidable. This means that we can only present avoidable costs in an aggregate form. Also, it does not provide solid evidence of the level of avoidable costs; it merely uses estimated falls in mortality and consumption rates to suggest that around half of costs could be avoided. No linkages between these falling rates and how they affect social costs are made. Therefore, the estimate provided should be viewed as an upper boundary or feasible maximum of potential avoidable costs. Collins and Lapsley (2008) does not try to estimate the total value of avoidable costs, which suggests they were not confident in providing a total figure. Given the lack of robust information available to carry out a thorough avoidable cost analysis, we provide a rough estimate of costs that can be avoided. We use a simple approach: we assume that 50 percent of costs are avoidable and apply this to our estimate of total costs. Table 6.2 shows the avoidable costs for alcohol and illegal drugs based on this approach. This should be treated as an upper limit of potential avoidable costs, and not an exact figure. Table 6.2 Potential avoidable costs of harmful drug use in New Zealand, 2005/06
Avoidable costs ($m) Tangible costs Intangible costs Total avoidable costs % of avoidable costs % of social costs
Source: BERL
Alcohol 1,440 780 2,220 70% 34%
Other drugs Joint AOD 520 200 710 20% 11% 330 0 330 10% 5%
Total 2,280 980 3,260 100% 50%
The potential avoidable costs amount to approximately $3,260 million, of which $2,280 million are tangible costs. This is equivalent to just under one quarter (21.4 percent) of Vote Health in 2006. This is an area that we recommend should be a focus for further research. Specific focus should be made in estimating Arcadian Normal and prevalence approaches for a New Zealand context, and further deriving avoidable costs from the results. The preferred approach advocated by Collins and Lapsley (2008) was avoidable costs estimates for individual interventions, based on an intervention effectiveness approach. The fact that it focuses on individual interventions and not on avoidable costs as a whole means it cannot be applied to this study. Firstly, Collins and Lapsley (2008) states that the avoidable costs for individual interventions are not additive, as many interventions are related and not necessarily mutually exclusive, meaning a total estimate could not be
69
derived. Secondly, many potential interventions with benefits that could not be well quantified were excluded from the study, and the list was not exhaustive. Thirdly, the intervention policies are specifically tailored to an Australian environment, which has a number of regulatory differences to New Zealand, meaning many of the potential interventions would not apply or would have a different effect. According to Collins et al (2006) avoidable cost estimates on their own do not indicate the rate of return a community gets from investing in alcohol misuse prevention and intervention over a period of time. A cost benefit analysis is required to produce this information. But to complete a cost benefit analysis, researchers need to know the avoidable cost component of alcohol misuse prevention and intervention. Avoidable cost estimation will therefore facilitate program analysis through a cost benefit analysis and the design of appropriate programs that achieve the best rates of return. In this context, Single et al (2003) notes that estimates of avoidable costs do not indicate how these cost reductions might be achieved or whether the social benefits that result from these programs exceed their social costs. Single et al (2003) states that project appraisals must be undertaken that evaluate the efficiency of alternative policies or interventions and/or treatments. These appraisals could use cost-effectiveness analysis to compare the cost of alternative policies or interventions and/or treatments, but they should be undertaken from the viewpoint of the community as a whole, not just the viewpoint of the government budget. 6.2 Injury costs
The injury cost sub-analysis aimed to estimate what portion of the estimated social costs of harmful AOD use resulted from injury. The main components examined were : health care for victims of crime and people injured as a result of their AOD use lost output due to injury-related mortality, absenteeism and reduced productivity road crash costs related to injury, such as lost output and medical care for crash victims intangible costs such as homicide, road crash and other fatal injures and reduced quality of life for road crash survivors. For both conceptual and data limitation reasons, a number of tangible costs included in the main estimates are not included in the injury sub-analysis. For example, conceptually, we do not expect justice sector costs such as lost property or preventative measures to be related to AOD-caused injuries. Due to an absence of data, we were unable to estimate what proportion of output lost because of excess unemployment was due to AOD-related injuries. As such, intangible costs are likely to make up a greater proportion of total injuryrelated costs than in the total social costs. This is because intangible costs pick up injury-
70
related loss of life and lost quality of life, where the former is relatively well recorded and has a high social cost. The calculation method is detailed in section 4.3.8 and is expanded on in section 9.4.8. Table 6.3 details the portion of total social costs of AOD use that were due to injury. Table 6.3 Injury costs of harmful drug use ($m), 2005/06
Injury costs ($m) Alcohol Other drugs Joint AOD Total
Tangible costs Labour costs Health care Road crash n.i.e. Total tangible costs 401.1 221.3 2.0 624.4 Intangible costs Premature mortality n.i.e. Road fatalities Homicide Road crashes, quality of life Total intangible costs Total injury costs % of injury costs % of social costs
Source: BERL
67.4 15.3 0.0 82.7
0.0 16.9 0.0 17.0
468.5 258.6 2.1 729.1
572.0 384.4 2.3 8.8 967.5 1,591.9 81.6% 24.4%
186.5 67.5 0.2 0.2 254.3 337.1 17.3% 5.2%
0.0 0.0 0.4 0.2 0.5 17.5 0.9% 0.3%
758.5 451.8 2.9 9.1 1,222.4 1,951.4 100.0% 29.9%
Total injury costs due to harmful drug use were estimated to be $1,951 million (29.9 percent of the total social costs of harmful drug use). This is equal to almost two thirds (65 percent) of the GDP of the insurance and services to finance and insurance industries in 2006. According to Table 6.3, tangible costs associated with alcohol misuse totalled $624.4 million, and intangible costs were estimated to be around $967.5 million. Injuries related to alcohol misuse were estimated to cost the community approximately $1,591.9 million in total for the year 2005/06. Over one third of the total injury costs ($729.1 million) were estimated to be tangible costs , with the largest part of the tangible costs attributable to labour costs (almost two thirds, 64,3 percent). Health care costs, however, accounted for over one third of tangible injury related costs.
75
75
Intangible costs make up a relatively larger proportion of injury costs than other costs. This is because many of the tangible costs were excluded as they were not injury related, or in some cases, for example the lost output for victims, it was impossible to separate out injury and non-injury related costs.
71
The majority of road crash impacts in Table 6.3 are included under the health care and lost output components, with the residual listed as not included elsewhere. Focusing on all road-crash components together, we estimate that over one seventh of tangible injury costs are related to road crashes. Table 6.4 decomposes the road crash-related injury costs embedded across the range of cost components in Table 6.3 to focus on the total tangible and intangible road crash injury costs.
76
Table 6.4 Total tangible and intangible costs of road crash injuries ($m), 2005/06
Tangible road crash injury costs ($m) Hospital/medical Emergency/pre-hospital Follow-on Loss of output (fatality) Loss of output (temporary disability) Total tangible costs % of tangible costs
Source: BERL.
Alcohol 6.9 3.9 3.5 83.7 2.0 100.0 87.9%
Other drugs Joint AOD 0.3 0.1 0.1 12.8 0.0 13.4 11.8% 0.2 0.1 0.1 0.0 0.0 0.3 0.3%
Total 7.4 4.1 3.7 96.5 2.1 113.7 100.0%
Intangible road crash injury costs ($m) Loss of life Pain and suffering Total intangible costs % of intangible costs
Source: BERL.
Alcohol 384.4 8.8 393.1 85.3%
Other drugs Joint AOD 67.5 0.2 67.7 14.7% 0.0 0.2 0.2 0.0%
Total 451.8 9.1 461.0 100.0%
Table 6.4 suggests that the bulk (87.9 percent) of the tangible road crash injury costs are attributed to alcohol, with the remainder due to other drug-related crashes. AOD-related intangible costs due to road crash injuries are over one third (37.7 percent) of the total intangible costs due to injury. Below we estimate the injury cost per person due to harmful drug use. The per user figures are likely to underestimate the full social cost per injured person, as the total injury costs are divided across all harmful users rather than the number of injured people, due to the difficulty in robustly estimating the number of people injured due to their alcohol or drug use. We estimated that approximately 513,000 people (aged 13 years old plus) drink in a harmful manner but did not use other drugs.
77
Taken alongside the costs of harmful alcohol use only,
76
We have not attempted to determine the proportion of the following tangible cost components associated with road crash injuries: legal and court, property damage, travel delays, insurance administration, fire/emergency service. As such, the road crash-related injury cost estimates are likely to be conservative.
77
A further 127,000 people drink and use other drugs, that is, use AOD harmfully.
72
this equates to approximately $3,100 per harmful drinker, of which $1,200 is tangible and $1,900 is intangible. Note, the per user figures here concentrate on the alcohol only costs, consistent with the denominator, and do not apportion joint AOD costs. Other drug use accounted for a relatively smaller proportion of total injury costs (17.5 percent), but it had a substantial estimated impact of $337.1 million. The majority of this total was due to intangible costs ($254.4 million in 2005/06), reflecting the large number of deaths attributed to other drug use. Approximately 27,000 people (aged 13 years old plus) were estimated to use illegal drugs and not alcohol, implying costs of $12,300 per drug user, of which $3,000 is tangible and $9,300 is intangible. These figures indicate a higher burden per user for other drug users compared to harmful alcohol users. This difference is mainly driven by the justice sector costs associated with illegal drug use, and intangible costs related to opiate attributable deaths. The research also investigated the impacts of joint AOD injuries. The tangible costs of injury due to harmful joint AOD use were estimated to be around $17.5 million. The bulk of these costs (96.9 percent) were tangible costs associated with health care. 6.2.1 Implications for ACC Below we examine the impacts on ACC, which represent a proportion of total social costs due to AOD-related injury. Table 6.5 summarises the estimated number of claims and the associated costs to ACC in the 2005/06 year, according to the six NZIPS categories, plus estimated AOD-related PHAS funding for inpatient and A&E services.
78 ,79
78
This is likely to underestimate the total number of claims, as these categories account for about four fifths of injury deaths and serious injuries.
79
We do not have figures for the total number of claims related to the public health acute services funding.
73
Table 6.5 ACC claim numbers and costs due to harmful drug use ($m), 2005/06
ACC claims Assault Drowning Falls Motor Vehicle Other Suicide and self harm Workplace PHAS - inpatient PHAS - A&E Total
Source: BERL, based on ACC data.
Alcohol # 10,674 30 76,787 12,875 118,208 128 61,728 280,429 $m 10.9 0.1 45.9 36.5 72.8 0.3 77.5 24.8 6.1 274.8
Other drugs # $m
4,159 57 0 4,215
11.5 0.12 0.00 0.53 0.13 12.3
The bulk of the estimated injury claims were due to harmful alcohol use. This reflects, in part, limitations in the data, as no assaults, drowning, other or workplace injuries are attributed to other drug use. However, a large proportion of the identified claims for suicide and self harm injuries (30.1 percent) and motor vehicle (24.4 percent) were attributed to other drugs rather than alcohol. Table 6.6 provides a sense of the AOD-related injury cost burden borne by ACC; this is based on Table 6.5 and the main social cost estimates. Table 6.6 Proportion of AOD injury and social costs borne by ACC, 2005/06
ACC claims Assault Drowning Falls Motor Vehicle Other Suicide and self harm Workplace PHAS - inpatient PHAS - A&E Total
Source: BERL, based on ACC data.
Total $m 10.9 0.1 45.9 48.0 72.8 0.4 77.5 25.3 6.2 287.1
% of tangible % of tangible injury costs 1.5 0.0 6.3 6.6 10.0 0.1 10.6 3.5 0.9 39.4 social costs 0.2 0.0 0.9 1.0 1.5 0.0 1.6 0.5 0.1 5.8
We estimate that ACC accepted approximately 284,640 compensation claims due to drugrelated injuries. These claims cost ACC $255.6 million in 2005/06. We estimate that it spent a further $31.6 million of PHAS funding on AOD-related inpatient and A&E services. The
74
total estimate of $287.1 million shows that harmful drug use is a substantial drag on ACCs resources, and one that has substantial scope for amelioration. We also estimated that it could have spent an additional $1.30 million on outpatient services, $1.17 million on pharmaceuticals and $82,000 on laboratory services due to harmful AOD use. These latter estimates, however, have not been included in the main estimates as we are not confident that the approach used for these elements generate robust estimates. However, we have included them as indicative of the possible magnitude of these costs. Expressing the estimated ACC costs in terms of tangible costs (to concentrate on the resource impacts), we estimate that ACC bore almost two fifths (39.4 percent) of drug related injury costs, and 5.7 percent of the tangible social costs.
80
The share of total injury
costs partly reflects the method used in this study. In particular, approximately one third of the tangible costs of AOD-related injury is due to premature mortality. Therefore, aside from benefits paid to eligible people who died in 2005/06, it would not be expected that ACC would cover this third of tangible costs. Excluding the net cost due to injury-related premature mortality of $233.4 million, ACC would bear over one half (51.6 percent) of the remaining AOD-related injury costs. Equivalently, no crime or diverted production costs are estimated to be due to AOD-related injuries, which should be taken into account when comparing ACCs injury-related AOD costs to total tangible social costs. These two costs components represent almost half (48.7 percent) of the tangible social costs of harmful AOD use. By excluding these and focusing on lost output, health and road-related costs only, ACC bears approximately one tenth (9.9 percent) of these tangible social costs. Table 6.6 indicates that the three highest cost categories are workplace, other, and motor vehicle injuries. These are closely followed by alcohol-related falls. Workplace injuries account for a disproportionately large share of compensation costs: they result in 22.0 percent of the alcohol-related injury claims (Table 6.5) but 31.8 percent of the non-PHAS claim costs (Table 6.6).
81
Similarly, although at a lower level, alcohol related assaults
represent 3.8 percent of claims but 4.5 percent of non-PHAS claim costs.
80
As the six NZPIS categories account for approximately 80 percent of injuries, the full cost could be up to one quarter higher than the estimate. If this were case, the number of AOD-related injuries would rise to 355,800 claims at a cost of $339.8 million. This would be just under half the estimated cost of injuries (47.2 percent), and over one sixteenth (6.6 percent) of the total social costs of harmful AOD use.
81
The percentages of claims and claims costs are based on the NZIPS-related figures, as we do not have figures for the total number claims related to the public health acute services funding.
75
6.3
Costs to the government
This sub-analysis reports on the estimated costs of AOD misuse borne by the government. Harmful drug use impacts on the government accounts by creating direct costs of higher expenditure as well as indirect costs resulting from lower tax revenue and higher benefit payments. The estimates based on the main social cost estimates related to government expenditure plus estimates of major transfer payments, such as labour taxes and unemployment benefits and consumption-related taxes. The calculation method is detailed in section 9.4.9. The estimates of the costs of harmful drug use from a government perspective cover increased (net) justice and health expenditure and reduced (net) tax revenue related to output and consumption. Table 6.7 summarises the estimated costs to the government. Appendix Table 26 provides further details. Table 6.7 Costs of harmful drug use government perspective ($m), 2005/06
Tangible costs to government ($m) Labour costs Consumption-related taxes Crime n.i.e. Health care Road crashes n.i.e. Total tangible costs to government % of tangible social costs
Source: BERL
Alcohol 464.9 -226.4 367.8 286.0 19.9 912.3 20.0%
Other drugs Joint AOD 46.4 72.6 254.8 31.0 0.5 405.3 8.9% 101.3 0.0 169.3 13.6 0.2 284.4 6.2%
Total 612.6 -153.8 791.9 330.6 20.6 1,601.9 35.1%
The costs of harmful drug use from a government perspective amount to an estimated $1,602 million, or over one third (35.1 percent) of the total tangible costs to society. Reallocating the joint costs, almost 70 percent of the costs are due to alcohol and 30 percent are attributable to other drugs. Justice sector costs related to harmful drug use impose the largest burden on the government, accounting for just under half (49.4 percent) of the estimated cost to the government, and is equivalent to 17.4 percent of the total tangible costs to society. Reallocating the joint costs, other drugs account for around two fifths of the estimated cost to the justice sector. The second largest cost is lost tax revenue from reduced output, and equated to 38.2 percent of the governments estimated costs. Premature death (including from road crashes), excess unemployment and losses suffered by victims of crime are the main
76
sources of these costs, and 90.9 percent of the labour costs are attributed to harmful alcohol use.
82
Health sector costs amounted to 20.6 percent of the governments cost burden. Again, alcohol-related harm drove the health sector costs from the governments perspective, accounting for over 90 percent of the health sector costs. While a substantial portion of this cost of alcohol is for hospital-based treatment of users and people injured in road crashes, over one third of it resulted from health care for victims of crime. The net impact of harmful drug use on consumption-related tax revenue is negative. That is, on balance the government receives more tax revenue from excise taxes and customs duties ($225.7 million) than it would forgo if harmful consumption ceased ($101.9 million). However, the net contribution of $94.1 million in consumption-related taxes pales against the total (gross) costs from the governments perspective of almost $1,755 million of lost revenue from output-related tax and increased expenditure.
82
This partly reflects that it was not possible to estimate excess unemployment and reduced productivity related to other drug use, and therefore the zero estimates for these components are likely to be conservative.
77
References
Abelson P. (2003). The Value of Life and Health for Public Policy. The Economic Record (Australia) Vol. 79 Special Issue, June 2003, S2-S13. Alcohol Healthwatch. (2007). Fetal Alcohol Spectrum Disorder. ANCD and NZDF. (2008). Beyond 2008 Regional Report - Australia and New Zealand. Australian Institute of Health and Welfare. (1999). National Drug Strategy Household Survey 1998. Canberra: Commonwealth of Australia. Australian Institute of Health and Welfare. (2006). Health expenditure Australia 200405. Health and welfare expenditure series, Number 28. Canberra: Commonwealth of Australia. Begg S, Vos T, Barker B et al. (2007). The burden of disease and injury in Australia, 2003. Brisbane: University of Queensland. BERL. (2006). Interventions by specialised AOD clinicians in justice settings. Report to ALAC. BERL. (2007a). Fiscal impacts of immigration 2005-06. Report to the Department of Labour. BERL. (2007b). The value of statistical life for fire regulatory impact statements. Report to New Zealand Fire Service. BERL. (2008). New Zealand Drug Harm Index. Report to the New Zealand Police. Borges G, Cherpitel C, Orozco R, Bond J, Ye Y, Macdonald S, Giesbrecht N, Stockwell T, Cremonte M, Moskalewicz J, Swiatkiewicz G and Poznyak V. (2006). Acute Alcohol Use and the Risk of Non-Fatal Injury in Sixteen Countries. Addiction 101, 993-1002. Brand S and Price R. (2005). The economic and social costs of crime. London: UK Home Office. Brazier J, Beverill M, Green C, et al. (1999). A review of the use of health status measures in economic evaluation. Health Technology Assessment 3(1). British Medical Association Board of Science. (2008). Alcohol misuse: tackling the UK epidemic. London: British Medical Association. Brooks R, with the EuroQol Group. (1996). EuroQol: the current state of play. Health Policy 37(1): 53-71.
78
Buchanan J, Colquhoun A, Friedlander L, et al. (2005). Maxillofacial fractures at Waikato Hospital, New Zealand: 1989 to 2000. New Zealand Medical Journal 118 (1217), 1-9. Bush, R., Wooden, M. (1994). Smoking, Alcohol and Absence in the Workplace: An Analysis of the 1989/90 National Health Survey, National Institute of Labour Studies, The Flinders University of South Australia, March 1994. Catalyst Health Economics. (2001). Alcohol Misuse in Scotland: Trends and Costs - Final Report. Chetwynd J and Rayner T. (1985). The economic costs to New Zealand of lost production due to alcohol abuse. New Zealand Medical Journal, 694-97. Chisholm D, Doran C, Shibuya K, and Rehm J. (2006). Comparative cost-effectiveness of policy instruments for reducing the global burden of alcohol, tobacco and illegal drug use Drug and Alcohol Review 25, 553-565. Choi B, Robson L and Single E. (1997). Estimating the Economic Costs of the Abuse of Tobacco, Alcohol and Illegal Drugs - A Review of Methodologies and Canadian Data Sources. Collins D and Lapsley H. (2002). Counting the cost - estimates of the social costs of drug abuse in Australia in 1998-9. Collins D and Lapsley H. (2008). The Costs of Tobacco, Alcohol and Illegal Drug Abuse to Australian Society in 2004/05. Canberra: Commonwealth of Australia. Collins D and Lapsley H. (2008). The avoidable costs of alcohol abuse in Australia and the potential benefits of effective policies to reduce the social costs of alcohol. Commonwealth of Australia: 2008. Collins D, Lapsley H, Brochu S, Easton B, Perez-Gomez A, Rehm J, and Single E. (2006). International Guidelines for the Estimation of the Avoidable Costs of Substance Abuse Ottawa: Health Canada. Connor J, Broad J and Jackson R. (2005). The Burden of Death, Disease and Disability due to Alcohol in New Zealand. ALAC Occasional Publication No. 23. Curtis J. (1994). Alcohol and Pregnancy: Information for health professionals. Alcohol Advisory Council.
79
Devlin N, Scuffham P and Bunt L. (1997). The social costs of alcohol abuse in New Zealand. Addiction 92(11): 1491-1505. Devlin N, Hansen P, Kind P et al. (2000). The health state preferences and logical inconsistencies of New Zealanders: a tale of two tariffs. Discussion paper, vol. 180. New Zealand: The University of York Center for Health Economics and the University of Otago. Drummond M, Stoddart G and Torrance G. (1994). Methods for the economic evaluation of health care programmes. Oxford Medical Publications. Easton B. (1997). The social costs of tobacco use and alcohol misuse. Public Health Monograph Series No. 2. Wellington School of Medicine, University of Otago. English D, Holman C, Milne E et al. (1995). The quantification of drug caused morbidity and mortality in Australia. 1995 edition. Canberra: Commonwealth Department of Human Services and Health. Ezzati M, Lopez A, Rodgers A, Murray C, editors. (2004). Comparative Quantification of Health Risks: Global and Regional Burden of Disease due to Selected Major Risk Factors. Geneva: WHO. Ezzati M, Vander Hoorn S, Rodgers A, Lopez A, Mathers C and Murray C. (2004). Potential health gains from reducing multiple risk factors In Ezzati et al. Comparative Quantification of Health Risks. Fergusson D, Horwood L and Boden J. (2008). Is driving under the influence of cannabis becoming a greater risk to driver safety then drink driving? Findings from a longitudinal study. Accident Analysis and Prevention 40, 1345-1350. Fielden J, Cumming J, Horne G, Devane P, Slack A and Gallagher L. (2005). Waiting for Hip Arthroplasty - Economic Costs and Health Outcomes. Journal of Arthroplasty 20(8): 990-97. Fillmore K, Kerr W, Stockwell T, Chikritzhs T and Bostrom A. (2006). Moderate alcohol use and reduced mortality risk: Systematic error in prospective studies. Addiction Research and Theory 14(2): 101-132. Fillmore K, Kerr W, Stockwell T, Chikritzhs T and Bostrom A. (2007). Has alcohol been proven to be protective against coronary heart disease? Addiction Research and Theory 15(1): 35-46. Foxcroft D. (2006). Letters to the Editor. Addiction 101, 1057-1059.
80
Gibson A and Shanahan M. (2007). Costs and outcomes of treatments for excessive alcohol consumption: Making policy decisions with the available data. Drugs: education, prevention and policy 14 (1) 1-17. Gilmore I and Sheron N. (2007). Editorial: Reducing the harms of alcohol in the UK. BMJ 335, 1271-1272. Hales J, Bowen J and Manser J. (2007). NZ-ADAM 2006 annual report. Prepared for the New Zealand Police. Hales J and Manser J. (2007). NZ-ADAM 2007 annual report. Prepared for the New Zealand Police. Harwood H; Fountain D and Livermore G. (1998). The Economic Costs of Alcohol and Drug Abuse in the United States 1992. Report prepared for the National Institute on Drug Abuse and the National Institute on Alcohol Abuse and Alcoholism, National Institutes of Health, Department of Health and Human Services. Harwood, H. (2000). Updating Estimates of the Economic Costs of Alcohol Abuse in the United States: Estimates, Update Methods, and Data. Report prepared by The Lewin Group for the National Institute on Alcohol Abuse and Alcoholism. HSRC. (2007). Evaluation of the Primary Health Care Strategy - Practice Data Analysis 2001-2005. Wellington: Health Services Research Centre and CBG Health Research Ltd. Humphrey G, Casswell S and Han D. (2003). Alcohol and injury among attendees at a New Zealand emergency department. New Zealand Medical Journal, 116(1168), 298-307. Johansson P, Jarl J, Eriksson A et al. (2006). The social costs of alcohol in Sweden 2002. Stockholm University. Jones C, Donnelly N, Swift W and Weatherburn D. (2006). Preventing cannabis users from driving under the influence of cannabis. Accident Analysis and Prevention 38, 854-861. Jones S, Casswell S and Zhang J-F. (1995). The economic costs of alcohol-related absenteeism and reduced productivity among the working population of New Zealand. Langley J and Marshall S. (1994). The severity of road traffic crashes resulting in hospitalisation in New Zealand. Accident Analysis and Prevention, 26(4), 549-54. Lee K and Snape L. (2008). Role of alcohol in maxillofacial fractures. New Zealand Medical Journal, 121 (1271) 15-23.
81
Leversha A and Marks R. (1995). Alcohol and pregnancy: Doctors attitudes, knowledge and clinical practice. NZ Medical Journal, 108, 428-300. Lindberg M and Amsterdam E. (2008). Alcohol, wine and cardiovascular health. Clinical Psychology, 31(8), 347-51. MacDonald Z, Tinsley L, Collingwood J, Jamieson P and Pudney S. (2005). Measuring the harm from illegal drugs using the Drug Harm Index. UK Home Office. Mann R. (2008). Reducing Alcohol-Related Deaths on Canadas Roads. http://www.camh.net, accessed on 28 July 2008. Markandya A and Pearce D. (1989). The social costs of tobacco smoking. British Journal of Addiction 84(10), 1139-1150. Mathew S, Kitson K, Watson P. (2001). Assessment of Fetal Alcohol Syndrome and Other Alcohol Related Effects in New Zealand. A Survey of Midwives in New Zealand. Palmerston North: Massey University. Mayhew P and Reilly J. (2007a). The New Zealand Crime and Safety Survey 2006. Wellington: Ministry of Justice. Mayhew P and Reilly J. (2007b). Community Safety - Findings from the New Zealand Crime and Safety Survey 2006. Wellington: Ministry of Justice. McFadden, M. (2006). The Australian Federal Police Drug Harm Index: A New Methodology for Quantifying Success in Combating Drug Use .Australian Journal of Public Administration, 65(4): 68-81. Miller T and Blewden M. (2001). Costs of alcohol-related crashes: New Zealand estimates and suggested measures of use internationally. Accident Analysis and Prevention 33, 783791. Miller J and Pikora T. (2007). Alcohol consumption among recreational boaters: Factors for intervention. Accident Analysis and Prevention 40: 496-501. Ministry of Health. (1999). Our Health, Our Future. Wellington: Ministry of Health. Ministry of Health (2006). Te Rau Hinengaro - The New Zealand Mental Health Survey. Wellington: Ministry of Health. Ministry of Health. (2007). Alcohol Use in New Zealand: Analysis of the 2004 New Zealand Health Behaviours Survey Alcohol Use. Wellington: Ministry of Health.
82
Ministry of Health. (2008). Portrait of Health - New Zealand Health Survey 2006-07. Wellington: Ministry of Health. Ministry of Justice. (1997). Home Detention - The evaluation of the home detention pilot programme 1995-1997. Wellington: Ministry of Justice. Ministry of Justice. (2006). Conviction and sentencing of offenders in New Zealand - 1996 to 2005. Wellington: Ministry of Justice. Ministry of Justice. (2007a). An Overview of Conviction and Sentencing Statistics in New Zealand 1997 - 2006. Statistical Bulletin, number 1. Wellington: Ministry of Justice. Ministry of Justice. (2007b). Benchmarking Study of Home Detention Programs in Australia and New Zealand. Ministry of Justice. (2007c). Conviction and Sentencing of Offenders in New Zealand 19972006. Wellington: Ministry of Justice. Ministry of Transport. (2006). The social cost of road crashes and injuries - June 2006 update. Wellington: Ministry of Transport. Mirrlees-Black C. (1999). Domestic violence: findings from a new British Crime Survey selfcompletion questionnaire. London: UK Home Office. ODea D, Thomson G, Edwards R, and Gifford H. (2007a). Tobacco Taxation in New Zealand. Wellington: Smokefree Coalition and ASH NZ. O'Dea D. (2007b). Economic and Social Costs of Smoking in New Zealand - Some Valuation Issues. December 2007, Auckland: HSPR conference. Office of National Drug Control Policy. (2001). The Economic Costs of Drug Abuse in the United States. Washington DC: Executive Office of the President. National Health and Medical Research Council. (1992). Is There a Safe Level of Daily Consumption of Alcohol for Men and Women? Canberra: NHMRC. Pidd K, Berry G, Roche A and Harrison J. (2006). Estimating the cost of alcohol-related absenteeism in the Australian workforce: the importance of consumption patterns. MJA 185 (11/12), 637-641. Productivity Commission. (1999). Australias Gambling Industries. Inquiry Report: Volume 1.
83
Ratcliff J, Ryan M and Tucker J. (1996). The costs of alternative types of routine antenatal care for low-risk women: shared care vs care by general practitioners and community midwives. Journal of Health Services Research and Policy 1(3):135-40. Rayner T and Chetwynd J. (1987). The economic cost of alcohol-related health care in New Zealand. British Journal of Addiction, 82, 59-66. Rayner T, Chetwynd J and Alexander T (1984) Costs of alcohol abuse in New Zealand - a preliminary study. Report to ALAC. Rehm et al. (2004). Alcohol Use. In Ezzati et al. Comparative Quantification of Health Risks Rehm J, Taylor B, Patra J and Gmel G. (2006). Avoidable burden of disease: conceptual and methodological issues in substance epidemiology. International Journal of Methods in Psychiatric Research 15 (4), 181-191. Rehm J, Patra J and Popova S. (2006). Alcohol-attributable mortality and potential years of life lost in Canada 2001: implications for prevention and policy. Addiction 101, 373-384. Ridolfo B and Stevenson C. (2001). The quantification of drug-caused mortality and morbidity in Australia, 1998. Room R, Babor T, Rehm J. (2005). Alcohol and public health, The Lancet, 365, 9458. Roper T and Thompson A. (2006) Estimating the costs of crime in New Zealand in 2003-04. Treasury Working Paper. Wellington: New Zealand Treasury. Single E. (1995). International Guidelines for Estimating the Economic Costs of Substance Abuse. Paper presented at the 2 Window of Opportunity National Congress. Brisbane. Single E, Collins D, Easton B, Harwood H, Lapsley H, Kopp P, and Wilson E. (2001). International Guidelines for Estimating the Costs of Substance Abuse. First edition. Single E, Collins D, Easton B, Harwood H, Lapsley H, Kopp P, and Wilson E. (2003). International Guidelines for Estimating the Costs of Substance Abuse. Second edition. Geneva: WHO. Sethi D, Racioppi F, Baumgarten I and Betrollini R. (2006). Reducing inequalities from injuries in Europe. The Lancet 368, 2243-2250. Snively S. (1994). The New Zealand economic cost of family violence. Report to the Department of Social Welfare. Wellington: Coopers & Lybrand.
nd
84
Stade B, Ungar W, Stevens B, Beyene J and Koren G. (2006). The burden of prenatal exposure to alcohol: measurement of cost. Journal of Fetal Alcohol Syndrome International, 4(5), 1-14. Ulrich J and Hanke M. (2002). Alcohol-Attributable Mortality in a High Per Capita Consumption Country Germany. Alcohol & Alcoholism. 37 (6), 518-581. UK Cabinet Office. (2003). Alcohol misuse: How much does it cost? London: Strategy Unit. United States General Accounting Office. (1970). Substantial cost savings from establishment of alcoholism program for Federal civilian employees by the Comptroller General of the United States. Report to the Special Subcommittee on Alcoholism and Narcotics. Van der Steeren A, Anderson I and Thorpe L. (2006). Individual harms, Community harms: reconciling Indigenous values with drug harm minimisation policy. Drug and Alcohol Review (25), 219-225. Watson P and McDonald B. (1999). Nutrition during Pregnancy: Report to the Ministry of Health. Palmerston North: Massey University. White I, Altman D and Nanchatal K. (2002). Alcohol consumption and mortality: modelling risks for men and women of different ages. British Medical Journal 325, 191-97. Wilkins C, Girlng M and Sweetsur P. (2006). Recent trends in illegal drug use in New Zealand, 2006. Auckland: SHORE. Wilkins C, Girlng M and Sweetsur P. (2007). Recent trends in illegal drug use in New Zealand, 2005-2007. Auckland: SHORE. Wilkins C and Sweetsur P. (2003). Drug Use in New Zealand. 2003 Health Behaviour Survey. Auckland: SHORE. Wilkins C and Sweetsur P. (2007). Trends in drug use in the population in New Zealand: Findings from national household drug surveying in 1998, 2001, 2003 and 2006. Auckland: SHORE. Wilkins C, Sweetsur P and Girlng M. (2008). Recent trends in illegal drug markets in New Zealand, 2005-2007. Auckland: SHORE. World Health Organisation. (2002). World Health Report 2002: Reducing risks, promoting health life. Geneva: WHO.
85
Glossary
Description A situation in which no harmful drug use has occurred. Two categories of psychoactive substances: alcohol and other drugs AOD use that results in a net social cost A broad range of personal, economic, and social impacts from past and present harmful drug use where the consumer does not bear all the costs or their decisions are not fully informed, rational and consistent. The proportion of total social costs that could in principle be changed by implementing appropriate public policies and behaviour changes. The proportion of total social costs due to injury resulting from harmful AOD use. The proportion of total social costs borne by the government, including transfers such as lost tax and increased benefit payments.
Concept Counterfactual Drugs Harmful use Social costs
Avoidable costs Injury costs Government perspective
Organisations ACC BERL BTE ESR HOI ICNZ LTNZ MoH NDIB NZFS NZHIS NZP SHORE StatsNZ TNZ WHO
Name Accident Compensation Corporation Business and Economic Research Ltd Bureau of Transport Economics Institute of Environmental Science and Research Limited Health Outcomes International Ltd Insurance Council of New Zealand Land Transport New Zealand Ministry of Health National Drug Intelligence Bureau New Zealand Fire Service New Zealand Health Information Service New Zealand Police Centre for Social and Health Outcomes Research and Evaluation Statistics New Zealand Transit New Zealand World Health Organisation
86
Abbreviation A&E AIE AMS AOD BZPs DUCO GDP GP GST HBV HCV HES HIV/AIDS ICD IDMS IDUs LSD MDMA MSTs NMDS NZ-ADAM PHAS SSRI VOSL VOSLY WIES YLL YLD DALY QALY
Description Accident and emergency Annual inmate equivalent New Zealand Polices Activity Monitoring System Alcohol and other drugs Benzylpiperazines Drug Use Careers of Offenders survey, Australian Institute of Criminology Gross Domestic Product General practitioner Goods and Services Tax (a value added tax applying to almost all goods and services, at a uniform rate of 12.5 percent since the late 1980s) Hepatitis B virus Hepatitis C virus Household Economic Survey (StatsNZ) Human immunodeficiency virus/ acquired immunodeficiency syndrome International Classification of Disease Illicit Drug Monitoring System Injecting drug users Lysergic acid diethylamide Methylenedioxy-methamphetamine Morphine sulphate tablets National Minimum Data Set New Zealand Arrestee Drug Abuse Monitoring Public Health Accute Services (ACC funding category) Selective serotonin reuptake inhibitor Value of Statistical Life Value of Statistical Life year Weighted inlier equivalent separations Years of life lost Years of life lost due to disability Disability-adjusted life year Quality-adjusted life year
87
Appendix Method and Calculation Detail
This section details the method used in the study, and is the basis for the summary provided in section 4 of the main report. This section necessarily duplicates some of the information provided in the summary. Both the summary and this detailed appendix are complemented by the methodological information provided in BERLs (2008a) illegal drug harm study. 9.1 Method
This section sets out the general framework that BERL used for the research. The AOD cost study draws on a similar methodology to that used in BERLs (2008) NZDHI project and the Australian estimates of drug harm (Collins and Lapsley 2008). All three studies draw on an internationally recognised standard for measuring the harmful substance use set out in Single et al (2003).
83
These sources, and their background documents, detail
the general methodology used in this study. As such, this report briefly outlines the method but concentrates on departures from that general method and the use of New Zealand data. 9.1.1 Conceptual framework This study uses a prevalence approach to calculate the current impact of harmful drug use. The prevalence approach estimates resource diverted in a given year due to the impacts of past and present harmful drug use. The costs estimated using this approach are compared to a counterfactual situation, in this case where there was no past or present harmful drug use. That is, in order to determine the harm avoided by reducing drug consumption we compare the current situation with drug use to a hypothetical case where there is no harmful drug use. The prevalence approach focuses on the impacts of both past and present, and is likely to give an informative picture of the impacts of harmful drug use in a given year. Collins and Lapsley (2002:13) make the following observation with respect to tobacco use, but it is relevant for a range of harms stemming from drug use.
To examine [the effects of smoking in a given year (incidence)] would not be useful since most abuse-related morbidity or mortality in a given year results from abuse in earlier periods we should be examining the impact of abuse over an extended period of time and
83
Single et al's (2003) guidelines are based on an international collaboration of substance abuse experts, epidemiologists and economists, and have been widely reviewed. As Collins and Lapsley (2008) note, the guidelines facilitate the development of valid and credible estimates of the economic costs of substance abuse. The guidelines treat methodological and data issues in detail.
88
this implies comparison with the counterfactual situation in which there was no abuse over this extended period.
The prevalence approach has the advantage of using currently available health data, such as mortality and morbidity figures related to drug use, to define what a counterfactual population would have looked like today. This is likely to result in more robust estimates than under the major alternative approach based on incidence.
84
The current approach
provides estimates that are comparable with other New Zealand cost-of-harm studies, such as Easton (1997) on tobacco and alcohol, ODea on suicide (2005) and tobacco (2007) and BERL (2008) on illegal drugs. The study aimed to develop estimates from the bottom up based on the behaviour of drug users. This approach focuses on the harm resulting from the consumption of particular types of drug. These estimates can be aggregated to give total drug-related harm. 9.2 Methodological issues
9.2.1 Types of cost The study looked at a range of social costs associated with illegal drugs, and quantified them by drug category and in aggregate. We took a conventional approach of dividing harmful substance use costs into tangible and intangible costs. Social costs The study focuses on social costs, that is, costs borne by the individual or wider society where the consumer does not bear all the costs or the decisions are not fully informed, rational or consistent with their long-term welfare. While there are arguments as to the legitimacy of, and value to be placed on, the private impacts of drug consumption (see Collins and Lapsley 2002: 17-19), this study does not attempt to explicitly value private benefits. Collins and Lapsley (2008, 2002: 20-21) and earlier studies study assume that all illegal drug consumption is harmful and imposes a social cost. While it is possible for illegal drug use to have benign or beneficial impacts, the epidemiological evidence base referenced for this study shows no net beneficial impacts on average for the health conditions considered. On
84
The incidence approach requires projections into the future and the use of discount rates to estimate the potential loss in output incurred due to morbidity and mortality resulting from harmful use. Both techniques, however, involve assumptions that affect the estimated cost.
89
the basis of this evidence, all resources diverted by illegal drug consumption are regarded as social costs.
85
The study aimed to estimate net social costs, rather than gross social costs of drug use. That is, drug use may offset some costs as users reduce the burden on societys scarce resources. For example, premature death reduces the health care that users might otherwise have required if they had lived longer. Tangible costs Tangible costs relate to resources used (or diverted) due to the harmful use of drugs in New Zealand. Tangible costs can be divided into direct costs and indirect costs. These costs may be borne by the individuals or by third parties such as taxpayers in general or employers. Direct costs Direct costs relate to resources directed away from an alternative use as a result of harmful drug use. Direct costs relate to the immediate impacts of illegal drug use borne by the individual, community and government. The most important direct costs in dollar terms, according to Collins and Lapsley, are: crime costs caused by harmful drug use resources diverted to drug production from beneficial consumption or investment road accidents health care costs.
86
At a conceptual level, direct costs also include the unpaid time given up by family and friends to take care of those who are ill as a result of harmful drug use, as well as time spent seeking or participating in treatment by persons affected by harmful drug use. Estimation of these costs would require information on the quantum and value of time involved, for example, whether such care displaces productive activities or leisure. This study does not estimate these impacts.
85
Sensitivity analyses showing the effect of relaxing this assumption for drugs and alcohol are shown in Section 11.1 of this report.
86
These components may also involve indirect and intangible costs. For example, in addition to health care related to road crashes, time off work would be counted as an indirect cost, while lost quality of life or loss of life would be measured as an intangible cost.
90
Indirect costs Indirect costs are borne by the wider society, and refer to potential resources or output that is not generated as a result of misuse. The primary indirect costs of harmful drug use are: production lost to the economy as a result of premature death of drug users reduced production by those who fall ill as a result of drug misuse
Intangible costs Intangible cost can only be borne by individuals and do not have (productive) resource implications for society. That is, reductions in intangible costs only benefit the individual. In the case of harmful drug use, intangible costs include: premature death as a result of drug misuse reductions in the quality of life due to pain and disability caused by harmful drug use.
These costs are borne by individual people who misuse drugs and their families and other associates who experience pain, disability and other reductions in quality of life. A decrease in the quality of life of one individual, however, cannot be transferred to another individual and these impacts have no direct resource implications for society. 9.3 Population patterns and impacts related to drug use
This section outlines general assumptions and inputs related to the population and drug use patterns. These are used in the individual cost component estimates in section 9.4 below. 9.3.1 Prevalence of drug use The study estimated the prevalence of alcohol, alcohol and other drug, and other drug only use over the 2005/06 period. This was done using data from a New Zealand study into the burden of disease due to alcohol (Connor et al 2005), two national illegal drug use surveys by the Centre for Social and Health Outcomes Research and Evaluation (SHORE) at Massey University (Wilkins and Sweetsur 2003, 2007) plus summary data from 2006 national drug use survey provided by SHORE. The 2006 national drug use survey data allowed BERL to separately identify the prevalence of alcohol use and other drug use.
87
frequent drug user survey (Wilkins et al 2006) was also used to estimate illegal drug use.
87
The 2006 national drug use survey (Wilkins and Sweetsur 2007) interviewed more than 2,000 individuals in the 13 to 45 age group about their drug-use history. It reports the percentages of adults who were current users of
91
Table 9.1 gives the population prevalence of alcohol consumption by drinking category from Connor et als (2005) study. These categories are based on average alcohol consumption per day, and were used in the WHO comparative risk assessment for alcohol and epidemiological studies by English et al (1995).
88
Table 9.2 shows the prevalence of total drug use by age and gender for the studys three drug categories: other drugs only, both alcohol and other drugs and alcohol only. This table is based on summary data provided by SHORE to BERL.
89
It allows for poly-drug use; see
discussion below. The categories are mutually exclusive, so a person is allocated to one category only. Table 9.1 and Table 9.2 were used to construct prevalence profiles by age, gender, drug use pattern.
90
These profiles were applied to demographic profiles to estimate the number of
people by age and gender who use specific types of drugs (alcohol only, alcohol and other drugs, or other drugs only), and by drinking pattern for alcohol.
alcohol or illegal drugs at the time of survey. The survey breaks down drug use in the last 12 months by age group and sex, for 13 to 45 year-olds.
88
In most studies examined (see the footnote below) the abstinent drinker category is set at 0 grams per day. English et al (2005) note that "quantities of alcohol consumed up to one quarter of a standard drink [2.5 grams] per day were included with the exposure category of 'abstinence' to provide sufficient tolerance for the inclusion of studies with small amounts of baseline contamination (commonly <1 drink per week or <1 drink per month).
89
The SHORE data have a fine degree of information by age, gender and drug use patterns for people aged 13-45 years old. We assume that alcohol and other drug use prevalence for people aged over 45 years old is the same as for people aged 40-45. Connor et al (2005) provide prevalence information for all people aged 15+ on alcohol use (but not other drug use). We note that the average population prevalence rate for alcohol use for people aged 45+ from Connor et al (78.4 percent) is close to the estimate from the SHORE data (77.7 percent). Therefore, the assumption we use for alcohol prevalence in older age groups is likely to be conservative.
90
The average daily consumption ranges are consistent with the WHO categories (Rehm et al 2004), the National Health and Medical Research Council (NHMRC) Australian alcohol guidelines (Pidd et al 2006), and recent New Zealand study on the burden of death, disease and disability due to alcohol (Connor et al 2005).
92
Table 9.1 Alcohol consumption by drinking pattern, sex and age group, 2005/06
Risk category Males

Age group 15-29 11.9% 58.3% 9.7% 20.1% 15-29 16.2% 57.2% 12.6% 14.0% 30-44 12.3% 66.4% 9.1% 12.3% 30-44 17.5% 65.9% 11.0% 5.6% Women 0.0 0-19.99 20-39.99 40+ 45-59 11.9% 71.3% 8.7% 8.1% 45-59 17.6% 66.4% 11.0% 5.0% Men 0.0 0-39.99 40-59.99 60+ 60-69 17.8% 70.0% 8.2% 4.0% 60-69 25.8% 66.3% 6.5% 1.4% 70-79 26.6% 65.3% 5.5% 2.7% 70-79 41.9% 53.6% 3.7% 0.8% 80+ 42.2% 53.9% 2.6% 1.2% 80+ 48.4% 48.1% 2.9% 0.6%
Females
Grams of alcohol per day

Source: Connor et al (2005)
Table 9.2 Total alcohol and other drug use by sex and age group, 2005/06
Age group Males
OD only A+OD Alcohol only
13-14 0.0% 5.0% 53.9% 13-14 0.0% 2.4% 30.9%
15-17 0.0% 21.0% 52.3% 15-17 0.0% 15.2% 56.4%
18-19 1.3% 40.4% 51.0% 18-19 1.7% 28.0% 59.8%
20-24 3.4% 39.4% 45.5% 20-24 0.0% 34.9% 54.1%
25-29 0.9% 23.5% 69.6% 25-29 0.0% 17.3% 67.5%
30-34 0.0% 23.3% 72.5% 30-34 0.0% 12.2% 70.1%
35-39 2.3% 13.4% 71.3% 35-39 0.5% 2.3% 81.3%
40-45 1.2% 12.5% 76.5% 40-45 0.4% 6.0% 80.3%
46-64 1.2% 12.5% 76.5% 46-64 0.4% 6.0% 80.3%
65+ 1.2% 12.5% 76.5% 65+ 0.4% 6.0% 80.3%
Females
OD only A+OD Alcohol only
Source: BERL, based on SHORE data.
The hazardous and high risk drinking patterns represent harmful alcohol use for the purposes of this study. Any illegal drug use is assumed to be harmful, reflecting the absence of evidence for the non-medical health benefits from the consumption of illegal drugs (Ridolfo and Stevenson 2001). Table 9.3 reports the estimated number of harmful drug users in 2005/06 by drug type.
93
Table 9.3 Harmful drug users by sex and age group, 2005/06
Total Both 19,100 85,800 88,000 77,200 56,600 62,400 65,400 53,900 112,000 46,500 666,900
Source: BERL, StatsNZ.
The estimates indicate that approximately 666,900 people aged 13 plus engaged in harmful alcohol, other drug or joint AOD use in 2005/06. This includes 513,000 people who use alcohol harmfully, 127,000 people who use both AOD and 27,000 people who only use other drugs. The most common drug used harmfully was alcohol, involving about 95 percent of harmful users. A further 2.27 million people used alcohol in a non-harmful way. 9.3.2 Drug consumption in New Zealand Poly-drug use Approximately one fifth of drug users in New Zealand are estimated to use more than one type of drug. Poly-drug use poses an issue for some calculations in this study where they are based on the number of drug users by drug type. For example, drug-related absenteeism should focus on an individual rather their drug use, which may see them recorded twice: once as a harmful alcohol user and once as a cannabis user. For this person, a day off work should be counted once. The resulting costs are included in the joint AOD cost category. To minimise this risk of double counting, the number of people using mutually exclusive drug categories was calculated. For example, the total number of harmful alcohol users comprises those people estimated to drink alcohol in a harmful manner and not use other drugs at all. Similarly, the other drug category covers a relatively small group of people estimated to use other drugs but not alcohol. The other drug group includes people who use multiple substances, but only where this does not include alcohol. BERL (2008a) describes how other drug users were allocated to mutually exclusive categories so a poly-drug using
94
person was counted only once. The final group of joint AOD users use both alcohol and other drugs. The prevalence of harmful drug use calculated above uses these mutually exclusive user groups. These figures were calculated using information on use of a range of drugs, and drinking patterns.
91
Comparison of drug use in New Zealand and Australia Table 9.4 and Table 9.5 compare New Zealand and Australian drug use prevalence. Table 9.4 New Zealand (2005/06) and Australian (2004/05) alcohol use prevalence by 92 drinking risk
Age group Australia Low risk Hazardous High risk Any use New Zealand Low risk Hazardous High risk Any use
Source: BERL, ABS.
1824 47.4% 7.9% 6.3% 61.5% 1824 57.3% 11.1% 16.8% 85.2%
2534 51.5% 7.6% 5.8% 64.8% 2534 64.5% 11.0% 13.3% 88.8%
3544 52.0% 8.5% 6.3% 66.8% 3544 66.8% 10.2% 8.8% 85.8%
4554 50.3% 9.6% 6.2% 66.1% 4554 70.7% 10.2% 6.7% 87.6%
5564 50.0% 9.1% 6.6% 65.7% 5564 73.2% 9.3% 5.1% 87.6%
6574 47.4% 6.7% 3.8% 57.9% 6574 77.7% 7.3% 2.6% 87.6%
75+ 41.1% 3.7% 1.7% 46.4% 75+ 80.5% 5.2% 1.7% 87.4%
Total 49.5% 8.0% 5.6% 63.2% Total 68.7% 9.7% 8.7% 87.1%
Table 9.4 suggests that New Zealand has a substantially higher prevalence of alcohol consumption than Australia. This is consistent across all age groups and drinking patterns, bar high risk drinking in later middle age groups, which is lower for New Zealanders.
91
In particular, the estimates drew on Connor et al (2005), Wilkins and Sweetsur (2003, 2007) and Wilkins et al (2006).
92
This table is based on the New Zealand figures estimated for this study and the ABS (2006) National Health Survey 2004-05. For comparability with the Australian figures, the table is based on people age 18+ years only.
95
Table 9.5 New Zealand (2006) and Australian (1998) illegal drug use prevalence by age 93 group
New Zealand Illegal drug Cannabis Opioids Cocaine Amphetamines (excl CM and MDMA) Crystal methamphetamine (CM) Ecstasy (MDMA) LSD
Source: BERL.
Australia 15-45 y.o. 24.2% 1.0% 1.8% 5.0% n.a. 3.2% 4.0% 14 y.o.+ 17.9% 0.8% 1.4% 3.7% n.a. 2.4% 3.0%
15-45 y.o. 17.9% 0.2%-0.4% 1.1% 3.4% 0.8% 3.9% 1.8%
New Zealand has lower drug use prevalence than Australia for all drug types except ecstasy. 9.3.3 Drug-attributable mortality Estimated mortality Data were supplied by the New Zealand Health Information Service (NZHIS) on all deaths in New Zealand over the five-year period from 2001/2002 to 2005/2006. Out of this file we extracted those that were wholly attributable (e.g. G62.1 Alcoholic polyneuropathy) or partly attributable to either the harmful use of alcohol, or other drug use, by ICD-10 code. These were identified from the table supplied by Collins and Lapsley (2008). Attributable fractions were taken from the Australian work by Collins and Lapsley (2008).
94
These fractions estimate the proportion of deaths that are attributable to harmful alcohol use, and separately to other drugs use, for each ICD-10 code by age and gender. The fractions are, however, Australian estimates. Applying these on New Zealand data may overestimate the deaths attributable in some areas, such as illegal drug-related conditions where New
93
This table is based on the 2006 figures for New Zealand (Wilkins et al 2006) and 1998 Australian figures reported in McFadden 2006 (based on AIHW 2000). For comparability we have estimated the Australian 15-45 year old prevalence rates, as well as the rates for all people aged 14 years old and over as reported in AIHW (1999).
94
The previous drug-harm study used a single fraction across all age groups and both genders. This study, however, uses an updated set of fractions with age and gender information. Therefore, there are some differences in the numbers of deaths attributable to drugs compared with BERLs prior work. For example, the range of conditions and some corresponding attributable fractions differ from the illegal drug harm study. For the drug study, schizophrenic disorders were not included (i.e. they implicitly had a zero fraction), while in the current study they have positive fractions for 15-64 year olds. This means more deaths and disability have been attributed to drug use in the current study. The classification for Hepatitis no -A, and B changed between Collins and Lapsleys 2002 and 2008 studies, with the removal of ICD categories B17.2-B17.8 and the inclusion of B17.0.
96
Zealands prevalence is lower, while underestimating others, such as alcohol -related conditions where New Zealands prevalence is higher.
95
Given the studys focus on harmful use, only positive attributable fractions were used. This affected alcohol-related conditions where there are net protective effects (represented by negative attributable fractions). A zero fraction was applied to conditions for which Collins and Lapsley report a negative attributable fraction. This may lead to an underestimate of the number of alcohol-attributable deaths due to harmful use, as we implicitly assume that harmful AOD use has a zero impact on these conditions. Applying the attributable fractions by condition to our dataset resulted in an estimate of the total number of deaths attributable to the harmful use of alcohol or other drugs, over a 5 year period, by age and gender. From these we derived mortality rates (number of deaths per 1000 population by age and gender). Estimates of the Additional Population The additional or missing population are those people who would be alive today if it were not for their deaths in the past by AODs. Essentially, this involves estimating the population that would have existed in 2005/06 based on modified mortality rates - assuming no harmful alcohol use or drug use occurred in the past. This estimation process involves the following steps. Apply current mortality rates (as derived above) to past populations to estimate past deaths. Calculate the proportion of those past deaths that would subsequently have died anyway from normal causes during the intervening period. Sum the remainder to get the estimated additional population who would currently be alive. Following standard practice, no attempt is made to quantify the births which did not take place because of lives lost through harmful alcohol or drug usage (Collins and Lapsley, 2002: 84). Assumptions underlying the additional population estimates Some simplifying assumptions were required in deriving these estimates. These include:
95
It was not feasible to re-calibrate attributable fractions using New Zealand prevalence statistics due to data constraints. In particular, it was not possible to get recent drug use prevalence statistics in a form that would suitably match up with the relevant epidemiological literature on the relative risks associated with harmful drug use.
97
That the mortality experience of those dying as a result of harmful drug use would otherwise have been identical to that of the rest of the population.
That mortality rates summarised in the New Zealand Life tables for 2000-2002 are applicable to earlier periods. In fact mortality rates were higher in earlier years, and the numbers otherwise surviving to the current period will be somewhat overstated.
That the mortality experience, and usage of alcohol and drugs, of emigrants and immigrants was identical to that of the rest of the population.
That the age-structure, by gender, of the population prior to 1991, back to 1951, was identical to that observed in 1991. This was required by the lack of readily available information on the age-structure in those earlier years. (Total populations prior to 1991 were scaled up by a factor of 2 percent, to make the earlier totals correspond to the Estimated Resident population definition used by StatsNZ from 1991 onwards.)
That AOD usage rates, and the consequences for premature mortality, were the same in earlier years, prior to 2001-2005.
Of these assumptions, that most likely to lead to significant error is the last. The latest New Zealand Health Behaviours Survey (NZHBS) indicates that the prevalence of hazardous alcohol consumption over the last decade has remained relatively constant (Ministry of Health 2008) at around 26 or 27 percent. However, an earlier survey referred to by Easton (1997) from the Alcohol Research Unit in 1993 on drinking in the Auckland region stated that 21 percent of drinkers had suffered some harmful effects from their own drinking over the previous 12 months. Therefore, it is possible that the prevalence of harmful alcohol use has increased since the 1950s and 1960s. Similarly, other drug use patterns and trends have changed over time (Wilkins and Sweetsur 2007). The mortality rates are based on data from the 2001-2005 period. But these rates are likely to overstate mortality from heroin and cocaine use in the 1950s or 1960s when use of these drugs in New Zealand was likely to be lower. Similarly, the data is not sufficient to include changes in recent trends, such as an increase in amphetamine use (and the related specific health conditions) over the last decade and a fall in cocaine use. Therefore our estimates for the additional population in the absence of harmful drug use are likely to be over-stated. 9.4 Cost calculations
The sections below details the key assumptions and inputs to the studys cost calculations by the major cost components. The components are: justice sector
98
lost output drug production health care road crashes intangible impacts on lost life and quality of life
The 2005/06 financial year is our base year, and where possible, data from this period was used. All figures are reported in 2005/06 dollar terms, and are exclusive of GST. An exception to the base year rule is for the additional population calculations. Mortality rates were based on a longer term average to capture the effects of past use on current mortality. The additional population estimates are, however, generated for the 2005/06 year. 9.4.1 Lost output A societys capacity to produce output depends on its labour force as a factor of production and the non-market resources in households.
96
This study considers four forms of

97
production loss due to the effects of harmful drug use on the labour force and households: Premature mortality Excess unemployment Absenteeism due to sickness or injury Reduced productivity
Table 9.6 and Table 9.7 show the distribution of additional population in 2005/06 by workforce status. These are based on the additional population and harmful user calculations, and population statistics on workforce participation and employment characteristics, for example, whether the person was employed full time or part time.
98,99
The participation rate and workforce structure by age group are based on StatsNZ figures in
96
Premature death and illness of those not in paid employment reduce a countrys capacity to support itself. That is, although unpaid, the non-market activities people do for their own household and those around them are valuable. We note, however, that there is no agreed substitute non-market valuation for the contribution of these people. Furthermore, non-market contributions are not counted as part of a countrys Gross Domestic Product (GDP), and it would be inconsistent to include this when expressing aggregate harm as a proportion of GDP.
97 98
BERLs drug harm project considered illegal-drug related premature mortality and absenteeism impacts only.
The workforce structure of people who use drugs harmfully is based on the working age population, that is, people aged 15-64 years old. The working-age population total differs from the estimated total number of harmful users in Table 9.3, as that table include people aged under 15 and over 65 years old.
99
We prepare separate estimates for the lost output of people in prison for drug-related offences. This number of people (by drug type, age and gender) is deducted from the workforce figures in Table 9.7 before calculating the main lost output estimate, in order to avoid double counting.
99
2005/06. These figures are used to calculate the labour costs due to premature mortality and illness described below. Table 9.6 Workforce status of the additional population with no harmful AOD use, 100 2005/06
Additional people Full time Part time UE/NiLF Total
Source: BERL.
Alcohol 5,476 1,461 7,240 14,177
Other drugs 1,682 449 1,488 3,619
Joint AOD 0 0 0 0
Total 7,158 1,910 8,728 17,796
Table 9.7 Workforce status of working-age harmful AOD users by drug type, 2005/06
Work force status Full time Part time UE/NiLF Total
Sources: BERL, StatsNZ.
Alcohol 277,800 74,100 105,400 457,300
Other drugs 14,300 3,800 5,300 23,400
Joint AOD 67,800 18,100 34,700 120,600
Total 359,900 96,000 145,400 601,300
The impact of drug use on mortality and labour market outcomes differs depending on the type(s) of drug used. The hypothetical situation where a person never engaged in harmful drug use, assumes that all people would have an equal likelihood of mortality or illness and a similar pattern of labour market outcomes. That is, we use a population average earnings and output figures.
101
Earnings were calculated using data on median earnings per hour and the average number of hours worked per week by age, sex and employment status (part time and full time). Earnings statistics were sourced from the StatsNZ New Zealand Income Survey for the June 2006 quarter. The Household Labour Force Survey (2004) provided hours worked per week. The value to society of lost output is considerably larger than lost earnings alone, for example, in addition to lost wages there is also lost profit. As such, the earnings profiles were scaled up to reflect the difference between wages and residual value added. The resulting output profiles were based on the assumption that the average GDP per FTE
100 101
The label UE/NiLF in the table refers to people who are unemployed or not in the labour force.
While the social cost estimates are based on a counterfactual of no harmful use in the past or present, not all of the estimated labour market costs in the future would be avoidable by reducing harmful drug use. Previous drug use may have persistent negative effects on a persons employability.
100
(BERL Forecast Database) is 1.87 times the average wage income (StatsNZ). These output profiles by age, gender and workforce status are used in the calculations below. Workforce losses due to premature mortality Drug-attributable mortality causes a reduction in societys productive capacity, which society could have benefited from in the counterfactual case (a world without illegal drug use). This cost is a function of how many people die prematurely due to drug use and what those people could have earned. The first element is the profile of working-age groups for the estimated additional population by drug type. The earning capacity of these groups was determined by matching them by age and gender to give labour force characteristics. This included estimating the proportion of people engaged in the labour market and of those people how many were unemployed, in part time employment and in full time employment. The output profiles were then applied to the lost workforce profiles to give drug-attributable production losses due to premature mortality. The losses by drug type were aggregated to give gross mortality related production losses. Premature mortality has an offsetting effect by reducing the demand on societys scarce resources through reduced consumption. Consumption resources released by the lost population were based on average private and public consumption expenditure per person. Private expenditure was derived from Household Economic Survey figures on average household expenditure and the average number of persons per household. Public expenditure per capita (excluding transfers such as social welfare benefits) was based on a study of population-related expenditure completed by BERL for the Department of Labour (BERL project reference #4497). Workforce losses due to excess unemployment An estimate of excess unemployment due to harmful drug use of 10 percent was taken from Rayner, Chetwynd and Alexander (1984, p47). Although this study was based in New Zealand, it was a preliminary estimate and only for alcohol-related costs. Therefore, no estimate is made for other drug attributable excess unemployment, and we take a conservative approach and apply the proportion only to identified high risk drinkers, rather than including the wider category of hazardous drinkers. We estimated this involved 14,600 full time and 3,900 part time people. These people are removed from the estimated labour force below, to avoid double counting.
101
As above, we apply age and gender output profiles to the estimated number of full time, part time and unemployed/NiLF people affected to estimate the lost output. Workplace losses due to absenteeism To estimate the number of people absent from work due to alcohol-attributable illness we used recent Australian research (Pidd et al 2008). This research indicated that 4.2 percent of male drinkers and 2.5 percent of female drinkers have time off work due to alcohol-related illness. This study also reported on the average number of days off (by drinking pattern) due to alcohol-related illness. For our study these figures were applied to the estimated number of harmful drinkers to estimate the number, and value, of days lost to alcohol-related illness. Other drug attributable absenteeism was calculated in a similar fashion to Collins and Lapsley. It begins by drawing on probability estimates from Bush and Wooden (1994) on the impact of substance smoking and alcohol on absences from the workplace. As in Collins and Lapsley, we assume that the probability of absenteeism is the same for tobacco use and illegal drug use. Overall, male drug users are assumed to be absent 70 percent more days and females 20 percent more days a year than abstainers. The probability of excess absenteeism by gender was distributed by drug type according to health service use. This had the effect of making absenteeism differ by the type of drug used. For example, male cannabis users were estimated to take 8 percent more sick days than the average (abstinent) male worker, while opioid users took off 40 percent more days. Australian Bureau of Statistics (2003) survey data on employee absences was used in the absence of robust New Zealand data. The survey provided information on how many hours a person taking sick leave took per week and the proportion of workers taking sick leave. These weekly figures used to estimate the number of hours of sick leave per worker per annum. Estimates of the other drug-using population by age and gender were used to develop a workforce profile as for harmful alcohol users To ensure poly-drug users were not counted twice, separate estimates were made for people who use alcohol only, other drugs only and both alcohol and other drugs. Combining the estimates of hours of sick leave, hourly output and the number of workers affected generated the value of drug-related absenteeism by drug type and the total cost of absenteeism.
102
ACC data on workers compensation for injury was used to calculate injury-related absenteeism due to harmful AOD use. This process is described in section 9.4.8. Workforce losses due to reduced productivity We estimate output lost as a result of harmful alcohol use impairing worker productivity using information on the number of days affected by drinking pattern and an estimate of reduced work efficiency on such days. As for the excess unemployment estimates, we do not estimate reduced productivity losses due to other drugs, and we take a conservative approach and estimate this loss for identified high risk drinkers only. That is, we focus on the estimated high risk group of 146,100 full time, 39,000 part time and 103,400 UE/NiLF high risk drinkers and other drug users.
102
Jones et al (1995) quantified absenteeism and reduced productivity costs associated with alcohol consumption using New Zealand survey data. This study showed that the top 10 percent (heavy) of drinkers
103
experienced 3.93 days per person on average when work
performance was reduced. We take a conservative approach and focus on the high risk drinking group, by workforce status, and calculate the total number days affected by using Jones et als figures and applying them to the age and gender output profiles discussed above. We assume that there is a 25 percent reduction in workplace productivity on days affected by harmful alcohol use. This is the same estimate used by Jones et al (1995) and Rayner et al (1984). However, this estimate is based on dated American research that uses expert opinion on the reduction in efficiency by alcoholics (United States General Accounting Office 1970). As such, this estimate illustrates the potential magnitude of lost output due to reduced productivity. Contemporary New Zealand research would be desirable to validate this estimate. We note that Jones et al (1995) also estimate the number of days affected for other levels of drinking. The second highest drinking group (the top 10-25 percent of drinkers) were affected for an average of 1.84 days per year. Our study estimates 11.21 percent of drinkers were in the hazardous drinking risk group, which is a slightly lower proportion than the 15 percent focused on by Jones et al (1995). As their wider band would include more moderate drinkers than our narrower hazardous drinking group, an estimate of days affected for the hazardous drinking group would be likely to be conservative and underestimate the number
102 103
UE/NiLF: Unemployed and Not in the Labour Force.
This corresponds very closely to the estimated proportion of drinkers in the high risk drinking group (9.99 percent) used in this study.
103
of days affected. Using the same approach as described above, but applied to the hazardous drinking group, we estimate that reduced productivity in the workplace would add a further $24.4 million to tangible labour costs, and reduced productivity in the household sector would add a further $1.0 million. Non-market output losses Losses in the non-market sector of the economy were calculated along broadly similar lines to the market sector calculations outlined above. Two elements that differed for these estimates were the amount and value of time spent on unpaid activity. The number of hours per person spent on unpaid activities in the home was based on StatsNZs Time Use Survey (2001). This survey found that women spent an average of 4.8 hours per day on unpaid work and men 2.8 hours per day. Grossed up to an annual figure, this equates to a woman spending approximately 1,750 hours on unpaid work and men 1,020 hours. The Time Use Survey is based on a sample of people inside and outside the labour force. If non-employed people do more unpaid work than employed people, then these annual estimates would tend to under-estimate household losses due to premature mortality. Ratcliffe et al (1996) have examined the value placed on activities in the household. Based on this research, we conservatively assume that lost output in households was valued at half the median hourly wage for elementary occupations. The value of an hour of unpaid activity was assumed to be the same for men and women. 9.4.2 Drug production Single et al (2003) discuss the value that should be placed on inputs diverted from legitimate uses to drugs produced for harmful use. This study assumes that resources used in such drug production have alternative uses, that is, their opportunity cost is not zero.
105 104
Therefore, resources used to produce drugs that are used harmfully represent a cost to society.
104
While harmful drug production arguably delivers benefits to some communities in New Zealand, this study assumes that resources freed up by lower drug production would flow to another, equally productive use. It is possible that some resources are better suited, for example, to illegal drug production than to their best alternative. However, this seems unlikely given the wide range of industries these resources could alternatively be employed in. Therefore, resources used to produce drugs that are used harmfully impose an opportunity cost on society.
105
This is similar to how other cost components are treated within this study. For example, it is assumed that AOD treatment professionals could work productively in another profession were there no harmful alcohol or other drug use; hence their use is a cost to society.
104
Alcohol In the case of alcohol, we draw on the consumption estimates in section 4.2.2 as one component in the calculation of resources diverted for harmfully used alcohol. That is, we assume that 50 percent of all alcohol consumed is done so in a harmful manner. We applied this proportion to an estimate of New Zealanders annual aggregate expenditure on alcohol. Annual aggregate expenditure on alcohol was estimated from StatsNZ s two most recent Household Economic Surveys (HES) 2003/04 and 2006/07.
106
The earlier survey
estimated that aggregate annual expenditure by all private households on alcohol was $1,594 million. The later survey reports that an estimated 1,596,200 households spent an average of $19.00 per household per week on alcohol, or $1,550 million. Taking the average of these two figures, we use an estimate of $1,572 million (GST inclusive) in the 2005/06 year. Excluding GST, duties and alcohol excise tax gives an annual aggregate expenditure of $684.5 million on alcohol.
107
Applying the 50 percent proportion to the annual aggregate expenditure, we estimate that $342.2 million of resources were diverted to produce alcohol that was then used in a harmful fashion. Illegal drugs The value of resources diverted by illegal drug production was based on BERLs (2008) drug harm study estimate. The calculations are described more fully in section 5.3 of that report. The estimate covered the value of domestic production plus imported drugs. The volume of drugs, by drug type, were typically valued a fraction of their street price, following Collins and Lapsleys approach. However, the value of domestically sourced pharmaceuticals diverted for opioid production was based on work by Sheerin (2004), and we allowed for differences in cost of imported and domestically sourced resources used in the production of methamphetamine.
106
Alcohol consumption is likely to be substantially under-reported by survey respondents in the HES. Therefore, the estimate of expenditure on alcohol consumption is likely to be conservative.
107
For consistency we exclude GST, and also exclude customs duties and excise tax on alcohol, as these taxes are transfers of resources rather than a social cost.
105
9.4.3 Justice sector costs This analysis aims to capture the costs of alcohol and drugs to the justice sector, including people affected by alcohol and drug crime, and agencies that deal with such crime and its effects.
108
Customs Two aspects of the Customs Service operations are covered in the section: its drug enforcement programme, and alcohol excise and customs duty operations. The Customs Service provided information on its total budget ($107.0 million) and drug enforcement expenditure ($28.7 million) in the 2005/06 financial year.
109
Enforcement
expenditure was based on the Customs Services analysis of the prop ortion of its budget dedicated to drug enforcement activity. It covered a range of cost centres, such as drug detector dog teams and drug investigations units (100 percent drug-attributable), mail centre operations (30 percent) and intelligence targeting (30 percent). The Customs figure was developed using estimates of drug-related activity under its various costs centres. In practice, it is difficult to accurately separate out drug enforcement activities from Customs other border roles. This is particu larly an issue for screening and search activity. For example, screening a passenger at the airport, searching a craft or examining a cargo container all address a range of border related risks or commodities. This caveat also applies to overhead expenditure, which was allocated in proportion to estimated drug enforcement expenditure. For the alcohol excise tax operation, data was obtained from the Customs Service on customs and excise duty expenditure obtained over the 2005/06 financial year. The attributable duty to alcohol was estimated using breakdowns of alcohol and other duties from December 2006 Crown financial statements and Vote funding for revenue collection.
110
108
The analysis does not include the costs of funding regulatory agencies in New Zealand such as the Alcohol Advisory Council of New Zealand (ALAC), the Liquor Licensing Authority or District Licensing Authorities. Some of the functions of these agencies are considered discretionary policy costs in terms of this studys methodology and are therefore outside its scope. While such agencies aim to minimise the harm from alcohol consumption, they are likely to exist as preventative measures even with changes in the actual level of harmful alcohol consumption. Other functions are to do with the regulation of the liquor market, such as licensing premises that serve alcohol, rather than being directly related to harmful alcohol consumption.
109
Total enforcement expenditure is composed of $16.4 million of operational expenditure plus $12.3 million of nonoperational expenditure. This latter figure equals the fraction of total non-operational expenditure that is proportional to drug enforcement operational expenditure.
110
Alcohol excise was 23.4 percent of crown excise tax revenue ($516 million) in 2005/06. Customs duty on alcohol was $197 million.
106
Based on the assumption that 50 percent of alcohol use is harmful, the cost due to harmful alcohol use was $0.13 million. Community Costs Victims of Crime Alcohol-and-drug-related costs borne by victims of crime were estimated using an approach similar to that in Roper and Thompson (2006). This included preventative expenditure, property losses, lost output, health service use and intangible costs. Intangible costs result from the pain, disability, and in the case of homicide loss of life of crime victims in 2005/06.
112 111
The loss of life values in the Treasury report (Roper and
Thompson 2006) were calculated using a different method to the prevalence approach used in this research. Where relevant, intangible costs of crime have been rebased from a human capital cost to be consistent with the life year values used in this study. Section 9.4.6 discusses this process. Community costs were calculated as the number of alcohol-and-drug-attributable crimes multiplied by the average cost per offence. These calculations were based on estimated actual crime rather than recorded crime. This involved applying crime multipliers from the Treasury study (Roper and Thompson 2006) to the offences recorded in 2005/06 to estimate actual crime levels (see Appendix Table 13).
113
The fraction of total criminal offences related to drug use and the distribution by drug type was based on information from the NZ-ADAM programme (discussed below) and New Zealand data on drug crime, described in the Police section below. For example, this data suggests that alcohol has been consumed prior to just over one in three offences (34.9 percent) and for just over one in five offences (21.9 percent) alcohol was a significant contributory factor to the commission of that offence.
114
111
Preventative expenditure comprises all those costs that individuals, households, businesses and institutions incur to prevent crime, e.g. security alarms, fencing and deadlocks [and] insurance administration (Roper and Thompson 2006: 13). It focuses on measures to protect property, and does not aim to include health prevention interventions or treatment, such as methadone programmes for opioid users.
112
Ideally, the study would have calculated the additional population that would have existed in 2005/06 in the absence of drug-attributable homicides. However, there were insufficient data on the pattern of homicides to robustly calculate this. As such, this component is likely to be very conservative. It measures drug-attributable homicides estimated from Police data in 2005/06 year rather than the people that would have survived to 2005/06.
113
Roper and Thompson (2006: 8) note that multipliers derived from the New Zealand National Survey of Crime Victims 2001 are not readily convertible to the particular crime sub-categories covered in this study. A similar caveat holds here with respect to the more recent 2006 New Zealand Crime and Safety Survey (Mayhew and Reilly 2007a). BERLs (2008a) report on illegal drug harm includes an appendix examines the impacts of using a limited set of New Zealand crime multipliers imputed from the Crime and Safety Survey (Mayhew and Reilly 2007a).
114
These figures are comparable to the findings from the New Zealand crime and safety survey 2006, which notes Nearly one-half of victims of crimes in public places who were in contact with the offender said the offender had been drinking. This compares to just under one-third of incidents in private places (Mayhew and Reilly 2007b).
107
Police and related expenditure NZP resources diverted due to alcohol and drug consumption were estimated from total Budget appropriations for the 2005/06 year. Crown revenue and receipts were removed from the total appropriations to get net expenditure of $942.6 million. Police time by offence category First, NZP expenditure was allocated using data from the NZPs Activity Monitoring System (AMS) on Police hours spent dealing with different offences in 2005/06.
115
BERL organised
the AMS data into eight offence categories, in order to align the data with information from other sources, such as the NZ-ADAM data and the crime multipliers. Appendix Table 15 details the offences included in these categories; these broad categories are based on the New Zealand Offence Hierarchy (NZOH) system. The data used to construct Table 9.8, below, relates to just over 7.63 million hours of Police time. This time equates to just over half the NZPs recorded activities that are coded to specific offences. A further 7.15 million hours of time was spent on incident and service activity; this time is not coded to specific offence categories. We assume that total Police time is distributed by offence according to the recorded statistics, as shown in Table 9.8. Table 9.8 Police activity by offence category, 2005/06
Offence category % of hours Violent 15.7% Property 5.8% Illegal drugs 7.2% Traffic 23.5% Drink/ Breaches Disorder Drugged Driving 1.7% 6.1% 7.5% Other 32.5% Total - all offences 100.0%
Source: BERL, based on NZP data
Offences attributable to harmful AOD use The amount of Police expenditure attributable to harmful AOD use is mainly based on the NZ-ADAM research. Information from the NZP is used to estimate the distribution of offending by drug type for drink/drugged driving.
116
Below, we discuss how we used these
This research also notes that three-quarters of offences occurred in private places and one quarter in public places. Applying these weights to the reported figures on crime where the offender had been drinking suggests just under 35.5 percent of offences involve alcohol.
115
Traffic offence data did not separately identify drink driving offences from drugged driving offences. Collectively, drink/drugged driving offences accounted for almost one quarter of road policing hours or 3.8 percent of total policing hours. Collins and Lapsley (2002) reported that this category accounted for no significant percentage of Australian police resource time for illegal drugs. We have used NZP advice to divide drink/drugged driving across the three categories of alcohol only, drugs only and joint AOD.
116
At the end of this subsection, we discuss an alternative New Zealand data source on the proportion of offences where alcohol is involved the NZPs Alco-Link database. While the data are different, our analysis of the alternative data confirms the general pattern shown by the NZ-ADAM data, and reinforces the validity of using this source.
108
data sources to estimate the proportion of all offences attributable to alcohol or other drug misuse.
117
The NZ-ADAM programme measures drug and alcohol use among people who have recently been apprehended and detained in watch houses by police (Hales and Manser, 2008). The NZP fund Health Outcomes International (HOI) Limited to conduct this research. Data are collected from four Police watch houses around New Zealand (Whangarei, Henderson, Hamilton and Dunedin). Detainees at these watch houses who meet specific criteria and who consent are interviewed. Each year since the research started in 2005, more than 2,000 detainees have been available to participate, with a total of 7,087 over three years. Of these, 42.1 percent met the inclusion criteria and agreed to be interviewed and 37.7 percent completed the entire interview process. BERL commissioned HOI to provide summary data using offence categories appropriate to our study. To strengthen the statistical power of the NZ-ADAM sample, data from 20052008 waves of NZ-ADAM were combined. The calculated proportions, therefore, represent averages over this period. HOI provided BERL with summary data for the first three waves of data, from July 2005 to June 2008. The detainee data were cross-tabulated by the number of offenders, offence type and the degree to which alcohol or other drugs contributed to the commission of the alleged offence. BERL used this data to estimate the fraction of offences (by offence category) where participants believed their alcohol or other drug use had contributed significantly to the commission of their alleged offence. The fractions are based on participants who believed that their alcohol use or drug-taking had contributed All, Some or A lot to the commission of their offence. While the attribution is based on self-report, Hales and Manser (2007) note that there is a high degree of correspondence between self-reported drug use and positive urinalysis for a range of the illegal drugs analysed. The correlation ranged from 69 percent to 90 percent for use within the past 30 days and 40 percent to 63 percent for those reporting use in the previous 48 hours. On average, across all the NZ-ADAM offence categories, alcohol or other drug use was a significant factor for more than one quarter (26.6 percent) of detainees. There are two exceptions to this process. In the case of drug offences, 28.8 percent of detainees said alcohol or drug use significantly contributed to their drug offences. This
117
The NZ-ADAM data do not separately identify drink/drugged driving offences from other traffic offences. We assume that all specific drink/drugged driving offences are attributable to alcohol or other drugs, so the total for this category sums to 100 percent. The distribution of this offence category by drug type is based on advice from the NZP that drug driving accounts for only about 0.26 percent of drink/drugged driving offences.
109
includes 25.7 percent of detainees who attributed their offending to other drugs only and 3.1 percent who used both alcohol and drugs. We assume that all drug offending is due to drug or joint AOD use, and scale the reported proportions up so they sum to 100 percent. In the case of drink/drugged driving offences, which are not separately reported in NZADAM, we assume that all such offences are attributable to alcohol or other drugs, so the total for this category sums to 100 percent. We base the distribution across drug types on advice from the NZP on the proportion of apprehensions leading to a charge for drink driving (almost all cases) versus drugged driving (about 2.6 in 1,000 cases). To ensure that drink/drugged driving offences are not double counted in the (other) traffic offence category, we scale down the AOD-attributable (other) traffic offence fractions by drug type (from NZ-ADAM) according to the estimated proportion of traffic offences due to drink/drugged driving by drug type. That is, the traffic category is for all other traffic offences bar drink/driving offences, which are separately estimated. The NZP data suggest that 24.1 percent of recorded road policing time is spent on drink/drugged driving offences. We split the percentage of road policing time by drug type using the NZPs advice on the distribution of drink/drug driving offences by drug type. For example, the bulk of drink/driving offences (99.74 percent) are due to alcohol, therefore, we estimate that 24.04 percent of total traffic offences are due to drink driving, with the remainder due to drugged driving. The NZ-ADAM figures suggest, however, that lower proportion (20.2 percent) of total traffic offences are due to alcohol. Therefore, it is likely that the bulk of alcohol-related traffic offences are due specifically to drink driving offences, and we conservatively assume that no (non-drink/drugged driving) traffic offences are due to alcohol. We estimate that 0.06 percent of total traffic offences are due to drugged driving. The NZ-ADAM figures suggest that 2.64 percent of total traffic offences are due to other drugs. We assume, therefore, that 2.58 percent of (non-drink/drugged driving) traffic offences are due to other drug use. That is, we assume that some traffic offences, which are not specifically drugged driving offences, are related to other drug use. Table 9.9 gives the fraction of reported crimes related to drugs only, alcohol only and both alcohol and drugs combined. Appendix Table 14 and Appendix Table 15 detail the offence categories used to construct Table 9.9 below.
110
Table 9.9 AOD-related apprehensions by offence category, 2005/06

Offence category Violent % offences attributable to 1.5% Drugs only Alcohol only A&OD only Total AOD 20.9% 3.4% 25.8% Property 2.9% 23.9% 8.1% 34.9% Illegal drugs 89.2% 0.0% 10.8% 100.0% Traffic 2.6% 0.0% 4.5% 7.1% Drink/ Breaches Disorder Drugged Driving 4.8% 17.3% 4.8% 26.9% 0.7% 33.1% 2.7% 36.5% 0.26% 99.74% 0.0% 100.0% Other 6.2% 11.8% 4.7% 22.7%
Source: BERL, based on HOI data and NZP advice.
The percentages in Table 9.8 (Police time by offence) and Table 9.9 (drug attributable offences) are jointly applied, by offence category, to net Police expenditure to calculate the cost of NZP resources diverted by AOD-related offending. The results of these calculations are shown in Table 9.10. Table 9.10 AOD-related Police expenditure ($m) by offence category, 2005/06
Offence category Police expenditure ($m) Drugs only Alcohol only A&OD only Total AOD Violent 2.2 31.0 5.0 38.2 Property 1.6 13.2 4.4 19.2 Illegal drugs 60.7 0.0 7.3 68.1 Traffic 5.7 0.0 10.1 15.8 Drink/ Breaches Disorder Drugged Driving 0.8 2.8 0.8 4.3 0.4 18.9 1.6 20.9 0.2 70.2 0.0 70.4 Other 19.1 36.2 14.3 69.6 Total - all offences 90.7 172.2 43.5 306.3
In addition to expenditure by the NZP, we include expenditure by the New Zealand Defence Force for the RNZAFs No 3 Squadron, which supported the NZPs drug eradication programme. This expenditure was included in the other drug category. In the 2008/09 financial year this support amounted to estimated $576,000. This figure was adjusted to 2005/06 dollar terms of $553,000, and included with the estimated Police expenditure. Alco-Link data In addition to the NZ-ADAM figures, we examined data from the NZPs Alco-Link database, in order to assess whether this information could be used to validate the NZ-ADAM figures. That Alco-Link database captures information on people apprehended for an alleged offence and whether alcohol was consumed prior to committing the alleged offence. The NZP provided BERL with exploratory figures from the Alco-Link database that had been linked to offender records in order to output the Alco-Link data by offence category. Strengths of the Alco-Link database are that it: contains population information, rather than being based on a sample of offenders
111
has routinely collected data can be linked with offender data contained in other Police databases.
Weaknesses of this data source for this research, compared with NZ-ADAM, are that it does not explicitly report: to what degree alcohol may have contributed to the commission of an offence the involvement of drugs other than alcohol.
It was unknown whether alcohol was involved in about one seventh of the records, but this proportion ranged between 1.4 percent (for drink/drugged driving traffic apprehensions) to 53.5 percent for the other crime category.
118
The analysis below concentrates on those
records where it was known whether alcohol had been consumed prior to committing an alleged offence. Table 9.11 compares the offence and apprehension rates from NZ-ADAM and Alco-Link.
119
The first column reflects the proportion of an offence category where alcohol or alcohol and drugs jointly were significant in the commission of an offence, i.e. the detainee said their drug use was between some to all of the cause for their offending.
120
The second column
includes all people who had committed an offence and had consumed alcohol, regardless of the reported level of involvement of alcohol in the commission of the crime. The third column reports the proportion of people apprehended for an alleged offence where alcohol was known to have been consumed before an alleged offence. In all cases, the proportion in the second column should be greater than that in the first. However, while the first column provides a sense of how important alcohol was in the commission of a crime, the second column is more directly comparable with the Alco-Link database, which records whether alcohol had been consumed prior to an alleged offence.
118
In the case of drink/drugged driving, the not known figure reflects recording errors, as technically all of these offences are due to alcohol or illegal drugs.
119
Note, the Other crime category includes sexual, dishonesty and a residual set of other offences. The Other crime category in main report does not include this last set, so the figures for this section are not directly comparable with the main report.
120
Note that these proportions differ from those in the main body of the report. The figures in Table 9.11 refer to where any alcohol was involved, which captures where alcohol only had been consumed plus where alcohol and other drugs had been consumed.
112
Table 9.11 Comparison of NZ-ADAM and Alco-Link offence and apprehension rates with alcohol involvement by offence category, 2005/06
% of incidents involving alcohol Violence Disorder Drugs Other Property Traffic Breaches All offences
NZ-ADAM Some-All 24.3% 35.8% 3.1% 16.5% 32.0% 24.7% 22.1% 21.9% Any 43.4% 66.3% 7.1% 26.5% 47.7% 37.0% 33.2% 34.9%
Alco-Link % yes 33.7% 22.8% 48.5% 10.7% 26.3% 45.7% 26.0% 34.2%
Source: BERL, based on NZP data.
The figures indicate that alcohol had been consumed prior to an offence in about one third of cases. This proportion is very close across both sources of data (34.9 percent in NZ-ADAM research and 34.2 percent in the Alco-Link database). The first column shows that in about one fifth of all cases alcohol was a significant influence on the commission of the crime. That is, comparing the first and second columns, alcohol was reported as a causal factor in about two thirds of offences where alcohol had been consumed. While there are differences between the average rates by offence category, these differences are not statistically significant at a 5 percent level (we have not reported these figures here). That is, the difference in rates by offence category between these two sources could be due to chance rather than representing significant differences between the groups examined. Given there are no statistically significant differences in the proportions between these two sources, we have used the NZ-ADAM figures, as they have the advantage of reporting the degree of causality of alcohol in the commission of an offence. Courts Court related expenditure was based on the 2005/06 Budget appropriations for Vote Courts and the output classes of Custody of Remand Prisoners plus Escort and Custodial Supervision from Vote Corrections. After removing Crown revenue and receipts, net court related expenditure was $353 million. Total court related expenditure was allocated to the offence categories using the Police time proportions from NZ-ADAM given in Table 9.9.
113
Prisons Prison expenditure in the 2005/06 financial year was based on the number of inmates incarcerated for crimes where alcohol and or other drugs significantly contributed to their offence times the average cost per prisoner in 2005/06. The Department of Corrections provided information on the prison muster in 2005/06 by offence type, age and sex. Drug-and-alcohol-attributable prisoner fractions were based on the Australian Institute of Criminologys Drug Use Careers of Offenders (DUCO) survey as New Zealand data were unavailable. These fractions were adjusted by New Zealand to Australia drug and alcohol prevalence rates to suit a New Zealand context.
121
Based on these figures, approximately 3,553 people (37.7 percent of the prison population) were incarcerated as a result of alcohol-and-drug-related crime.
122
Using an average cost
per prisoner of $68,879, this generated a total of $244.7 million. As this estimate is based on Australian attributable fractions, and in light of the high proportion of the New Zealand prison population that have drug and alcohol problems, the actual cost of drug-related imprisonment in New Zealand could be considerably higher than this conservative estimate. Incarceration poses a further cost due to the lost output of inmates. Lost output estimates were calculated on the basis of the age and gender profile of inmates jailed due to alcohol use, drug use, and combined alcohol and drug use. This totalled $76.8 million. Section 9.4.1 describes the general process to calculate lost output. The estimates of lost output are an exception to the general approach of estimating net costs. That is, the estimates do not take into account prisoners earning while in prison. The Department of Corrections, however, advises that working prisoners are paid a token amount per day (approximately $1). Therefore, the gross and net figures are unlikely to differ substantially. For example, if all 3,553 alcohol-and-drug-attributable prisoners were paid for every standard working day (5 days a week, 46 weeks a year), this would amount to just $817,000 or less than one percent of the estimated loss of prisoner output. Community sentences Alcohol-and-drug-related community sentence expenditure was based on the number of people serving community-based sentences and orders in 2005/06. Net appropriations for
121 122
New Zealand has 66.9 percent the rate of Australian drug use, but 137.9 percent the rate of alcohol use.
A Ministry of Health report (2007:16) suggests that a substantially larger proportion (almost 90 percent) of the prison population has an alcohol and/or drug problem. This suggests the estimated proportion of alcohol-and-drugattributable incarceration is likely to be conservative.
114
this output class were $69.8m (Department of Corrections 2007). This equals total appropriations less a small surplus. The distribution of community sentences by offence type was available from Ministry of Justice sentencing statistics. The alcohol-and-drug-attributable conviction fractions used for the prison estimates were applied to the community sentence figures. These figures were used to estimate the part of net expenditure that related to drug use. This amounted to $46.3 million. Home detention involves Corrections Services managing adult offenders within the precincts of a specified residence during specified hours.
123
Alcohol-and-drug-related home detention
numbers were based on the proportion of the corrective services population on home detention and the overall proportion of crime leading to a criminal sentence (16.7 percent). This process gave a figure of 132 drug-related home detainees. The cost of a home detainee per year of $14,800 was based on the average length of time on home detention multiplied by the capital and operational cost per person.
124
In total,
alcohol-and-drug-related home detentions amounted to approximately $1.94 million. To avoid double counting, the home detention estimate was deducted from the estimate for community sentences. 9.4.4 Health care Pharmaceuticals Data was sourced from Pharmac on pharmaceuticals involved in AOD-related treatment. As such the estimates cover government subsidised pharmaceuticals only, and do not include over-the-counter pharmaceuticals. Particular pharmaceuticals included were:
Alcohol Naltrexone hydrochloride Buprenorphine hydrochloride Disulfiram
Other drugs Methadone hydrochloride Combivir plus efavirenz (HIV) Pegatron (Hepatitus C)
Pharmaceuticals for co-morbid depression in illegal drug users were estimated as per the BERL (2008) drug harm study.
123 124
Home detention was made a separate sentence type in 2007.
The underlying figures were based on two Ministry of Justice reports on home detention (Ministry of Justice 1997 and 2007a).
115
For some pharmaceuticals used for multiple conditions that were not AOD-related, we did not include estimates. For example, the alcohol calculations do not include anti-nausea drugs. Primary care and accident & emergency care Data outputted from the NZHBS for this project indicated no excess use above the general population average of primary care services due to harmful alcohol consumption. However, this is likely to be an underestimate as some specialist AOD treatments require a referral from a primary care physician. This study estimated GP cost based total GP funding and the ratio of people admitted to hospitals with AOD-related conditions, following Collins and Lapsleys (2008) method and advice. That is, we assumed that 4.4 percent of people consulted a GP in regard to alcoholrelated problems. We applied this proportion to estimated total funding from public and private sources for general practitioner services. We estimate this funding amounted to approximately $1,163 million in 2005/06, based on the Ministry of Healths Health Expenditure Trends series for 2002/03 (Ministry of Health 2005) and adjusted to the 2005/06 year assuming similar funding growth as in previous years. We also estimated the cost of providing the services of drug and alcohol workers for harmful alcohol users. The method used to calculate this component is the same as for illegal drug users, which is discussed below. We assume the proportion of harmful alcohol users engaging these services was the same as for illegal drug users. Illegal drug-related primary care use was based on reported health service use by frequent drug users in Wilkins and Sweetsur (2007). The primary health care services examined include general practitioners (GPs), counsellors, drug and alcohol workers, social workers, psychologists, and psychiatrists. Hospital based accident and emergency (A&E) service use was also reported. Wilkins and Sweetsur (2007) interviewed methamphetamine, ecstasy and injecting drug users. For this study, we conservatively assume that other stimulant and LSD users have similar levels to ecstasy users (who have the lowest rates out of the three groups interviewed). The SHORE study did not report on primary care use by cannabis users;
116
therefore we assume that this group does not have excess primary care use. No information was available on health service use by occasional drug users.
125
The proportion of people using health care services as a result of their drug use was applied to the number of frequent drug users by age and drug type. The number of people in mutually exclusive drug use sub-sets was used to avoid double counting. This gave the number of people using specific health services. This figure was multiplied by the average cost of the specified health services.
126
Savings from premature mortality were calculated where counterfactual data were available. Data from the General Practice Computer Databases were used to develop an agegender profile of GP use.
127
This was applied to the lost population to estimate savings.
Ambulance service data recorded the total number of New Zealanders transported to hospital. We use this to generate a conservative estimate of the number of people that would not usually have used the ambulance service and A&E due to premature mortality.
Savings in other primary care services were not estimated.
The General Practice Computer Databases are an alternative source of data on drug-related GP use compared to the NZHBS. These databases evolved from the Royal New Zealand College of General Practitioners (RNZCGP) Computer Research Network. The Network consists of general practices from throughout urban and rural New Zealand. Data is collected from computerised records of consultation notes, prescriptions, investigations and referral forms. Data from 30 practices between 2005 and 2007 were examined for patients aged 10 years plus treated for harmful alcohol and drug use. These patients were identified from GP Read codes as having an alcohol or drug abuse problem. The search results represent an underestimate of all such patients at these practices having such problems since Read coding only occurs for around 20 percent of consultations. Just over 0.65 percent of 142,353 patients were identified as having AOD problems. Females had an average of 8.4 additional GP visits per annum above the population rate,
125
A sensitivity analysis indicated that if occasional drug users level of use was one fifth that of frequent drug users, then this would add $150,000 to the net health care cost. Therefore, excluding occasional drug users from these calculations is unlikely to substantially alter the estimate.
126
These rates were based on expert advice BERL sought from researchers in the Schools of Medicine in Wellington, Christchurch and Dunedin, except for GP rates. GP rates were based on publicly available rates and a recent Health Research Council report (HSRC 2007).
127
The databases draw on the Royal New Zealand College of General Practitioners Computer Research Network.
117
while males had an average of 5.4 additional GP visits per annum. To get an indication of the cost of AOD-related primary care based on these figures we multiply together: the percentage of people with AOD-related problems visiting a GP age-gender excess GP consultation rates national population age-gender profiles (aged 10 years plus) cost per consult (for those less than 18 years old and for those 18 years or older)
After allowing for savings from premature mortality, this calculation suggests a figure of $145.1 million, or 4.0 percent of the tangible costs estimated in the main report. This dollar figure is approximately one fifth of the net medical (primary care) services figure estimated in Collins and Lapsley (2008). Collins and Lapsleys estimate suggests that just under five percent of national spending on medical services was due to AODs. Our figure above corresponds to approximately 12.5 percent of New Zealands expenditure on general practitioners. Therefore, the figure is broadly in line with New Zealands smaller population and differing drug use prevalence, although it may overestimate the true cost. However, the figure calculated above is based on a sample of less than 500 people with identified AOD problems. Therefore, the estimate should be regarded as indicative, and it is not used in the main report estimates. Other community care Total national expenditure on alcohol and drug programs were obtained from the Addiction Treatment Services division within the Ministry of Health. This was $96.7m for the 2006/2007 year. This figure adjusted into 2005/2006 terms was $92.9m. We could not separate data into alcohol programs and drug programs; this is because many of the programs were interrelated and dealt with both harmful alcohol and drug use. This included community workers, residential beds, and independent medical detox beds. However the value of drug-specific program methadone placement ($11.7m) was noted. Ambulance services Estimates of alcohol-related and illegal drug-related ambulance service use were based on information provided by ACC and the St John Ambulance Service. The data related specifically to harmful alcohol use were estimated using ACC ambulance conveyance numbers. The estimates are likely to under-estimate as they rely on only callouts for injuries.
118
The drug data is related specifically to drug-related overdose, ingestion and poisoning. They do not cover other drug-related callouts such as falls or violent assaults. Therefore the estimates are likely to under-estimate actual ambulance service use. Hospital care The cost of drug attributable hospital care was based on inpatient treatment.
128
Data were
supplied by the New Zealand Health Information Service (NZHIS) on all publicly-funded hospital discharges in New Zealand over the 6 years from 2001 to 2006. Out of this file we extracted each record that was wholly attributable or partly attributable to either the harmful use of alcohol, or other drug use, by ICD-10 code. tables sourced from Collins and Lapsley (2008).
129
The set of conditions was based on
130
Each discharge has a case-weight attributed to it, given in the data-set as weighted inlier equivalent separations (WIES). This broadly shows the relative complexity of the particular case, relative to the average over all cases. For example, a case-weight of 2 represented a case twice as complex as an average case. Attributable fractions were taken from the Australian work by Collins and Lapsley (2008). These fractions estimate the proportion of hospital discharges attributable to harmful alcohol use, and separately to other drugs use, for each ICD-10 code by age and gender. The fractions are, however, Australian estimates, and so applying these on New Zealand data may overestimate the deaths attributable in some areas (such as drug-related conditions), while underestimating others. These fractions were applied to the case weights in our data for each Collins and Lapsley category. We could then estimate the total average cases attributable to the use of either alcohol or other drugs, over the 6 year period. The total average cases can then be applied to the average cost (or case -weight multiplier) for a specified year to give the approximate total treatment cost to the hospital for cases attributable to drug and alcohol abuse in that year. The case-weight method is likely
128
Some health care services for drug-related conditions are provided in an outpatient setting. The NZHIS has some data on outpatient treatment in its National Non-Admitted Patient Collection dataset. However, this data is organised by purchaser code rather than by diagnosis, so it is not possible to directly use this data to estimate drug attributable outpatient costs. As such, the hospital cost estimates are likely to be conservative.
129
External cause codes (starting with V, W, X or Y) were also used to identify drug-caused deaths and hospital discharges. These codes are primarily related to injuries, for example, assaults or road injuries.
130
One exception is the inclusion of fetal alcohol syndrome (FASD), ICD-10 code Q86.0, which does not appear on Collins and Lapsleys list of alcohol attributable conditions. It seems reasonable to assume that this condition is fully attributable to alcohol misuse.
119
to be a conservative estimate of alcohol and drug-related hospital costs as they were not designed to capture the total cost of hospital procedures. Table 9.12 shows case-weight multipliers for medical/surgical inpatient cases, excluding GST, for recent years. A case-weight multiplier shows the average cost of treating a patient with a case-weight of one, or multiple thereof. Table 9.12 Case-weight multipliers, 1998/99 2007/08
Financial year 1998/99 1999/00 2000/01 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08
Source: NZHIS.
Medical/Surgical Neonatal Inpatient Inpatient 2433.62 2399.22 2487.16 2479.01 2617.72 2728.55 2854.88 2949.09 3151.01 3740.38 None 2761.48 2732.47 2677.23 2827.03 2946.72 3124.17 3124.17 No longer used
The case-weight multipliers given are for financial years, ending June. This report uses the case-weight multiplier for the 2005/06 financial year, namely $2,949.09. Total Hospital Costs Total publicly-funded hospital costs caused by alcohol use and other drug use were calculated by multiplying the number of case-weights by the case-weight multiplier. Table 9.13 shows the annual costs for the six years 2001 to 2006, in 2006 dollar terms. Table 9.13 Estimated hospital costs of AOD-caused cases 2001 to 2006
Year 2000/01 2001/02 2002/03 2003/04 2004/05 2005/06
Source: BERL.
Alcohol Case-weights 18,358.9 37,416.9 38,572.6 38,789.6 39,576.7 42,734.4 Cost ($m) 54.1 110.3 113.8 114.4 116.7 126.0
Other drugs Case-weights 2,148.8 3,190.2 3,330.5 3,267.0 3,295.1 3,822.8 Cost ($m) 6.34 9.41 9.82 9.64 9.72 11.27
120
Table 9.13 shows that gross alcohol-attributable hospital costs amounted to $126.0 million in 2005/06, while other drug attributable hospital costs were $11.3 million. Detailed cost breakdowns by category in Appendix Table 19 and Appendix Table 20 show the distribution of hospital costs for AOD-related treatment for the period from 2001 to 2006. Savings in hospital costs from premature mortality were based on the figures in Collins and Lapsley (2008). They estimated savings were approximately 4.6 percent of gross costs of alcohol and 23.2 percent of other drugs. These rates were adjusted for New Zealands lower rate of premature mortality. These calculations suggest savings of $5.0 million for alcohol and $2.0 million for other drugs. Therefore, net hospital costs were $121.0 million for alcohol and $9.2 million for other drugs. 9.4.5 Road crashes The key components of tangible road crash costs attributed to alcohol misuse, other drug use, or joint AOD include: Hospital or medical Emergency or pre-hospital Follow-on costs Loss of output due to premature mortality or temporary disability Legal and court Property damage Travel delays Insurance administration Fire or emergency service.
The intangible costs primarily relate to loss of life and pain and disability. The number of AOD-related road crashes The number of road crashes was estimated from Ministry of Transport data recorded in the Crash Analysis System (CAS) where drugs were a contributing factor. These records have some issues for this analysis: they do not establish causality, nor do they accurately report the actual number of non-fatal crashes (Ministry of Transport 2006). In order to be conservative, the number of road crash fatalities was scaled down in proportion to the difference between Ministry of Transport and NZHIS data on drug-attributable road deaths (see section 9.4.4).
121
The number of non-fatal accidents where drug use was suspected was scaled up to allow for under-reporting. The scale factors were drawn from Table A.2 of the Ministry of Transports (2006) social cost analysis. The report has factors by severity of crash but it does not have any information on the degree of under-reporting by drug type. Therefore, we apply the same scaling factor to all serious and minor crashes regardless of drug type. The scaling factors we use for serious and minor accidents are 1.71 and 3.74. The costs of hospital/medical, emergency/pre-hospital, follow-on health care, legal and court, and intangible costs were included in their respective sections. The remaining cost components (n.i.e. not included elsewhere) are looked at in detail below: Loss of output (due to temporary disability) Property damage Travel delays Insurance administration Fire or emergency service
Appendix Table 21 summarises the road crash cost estimates. Loss of output due to temporary disability Output lost due to temporary disability due to a road crash is based on the number of injuries (estimated as part of this project) and the cost per injury (Ministry of Transport 2006). This component amounted to $2.05 million, and is primarily attributed to alcohol use. Output lost as a result of a premature workforce (net of consumption resources saved) is included in the main lost labour estimates. Property damage Road crashes attributed to either alcohol misuse or other drug use caused property damage of $22.6 million in 2005/06, of which $21.7 million was due to alcohol misuse. The degree of damage is related to the severity of the crash on average, $8,800 for a fatal crash, $5,600 for a serious crash, and $4,500 for a minor crash (Ministry of Transport 2006). Travel delays The calculation of travel delays was based on the assumption that only fatal and/or serious crashes would cause delays in time; hence the impacts of minor road crashes were excluded from our estimation.
122
Out of the total $27.7 million loss because of travel delays, $26.8 million was attributed to alcohol misuse. This is due to a significantly large number of alcohol attributed road crashes. The cost of a fatal or serious crash in urban areas averaged $1,326 in urban areas, and $1,086 in rural areas. Insurance administration Alcohol attributed road crashes cost $45 million in 2005/06, according to our research, whereas other drugs attributed crashes had an impact of $2.4 million. As vehicle insurance accounts for around 50 percent of the total claims in 2005/06
131
, it is
safe to assume a majority of the fixed/overhead costs of the insurance business are due to vehicle insurance. Fire/emergency service In 2005/06, approximately $46 million (18 percent) of the New Zealand Fire Services operating expenditure was due to motor vehicle crashes for 2005/06, based on information from the New Zealand Fire Service.
132
Of this, $19.4 million was estimated to be for emergency service responses to road crashes attributed to alcohol misuse. Another $47,000 and $23,000 were spent on emergency services because of the road crashes attributable to joint AOD and other drug use. 9.4.6 Intangible costs Intangible costs result from the pain, disability and loss of life resulting from the impacts of AOD use. They represent psychological and welfare costs. These costs differ from the tangible costs of lost output as they do not divert productive resources. Loss of life This study uses a value of $106,600 per year of life lost due to the premature mortality of harmful alcohol and drug users, and alcohol-and-drug-related homicides and road accident fatalities. This value is based on LTNZs VOSL of $3.05 million (in 2006 dollars). The VOSL is converted to a single year estimate using the Pharmac discount rate of 3.5 percent.
133
131 132 133
The Annual Insurance Industry Review, Insurance Council NZ, Annual Report, 06-07 Based on operating expenditure, NZ Fire Service Annual Report, 2006, p65.
For more details on the reasoning and kinds of calculations involved, see for example Easton (1997), Abelson (2003), and O'Dea (2007b). The estimates are based ultimately on the 'Value of a Statistical Life' (VOSL) as
123
The number of people who died in 2005/06 as a result of alcohol and drug use draws on the estimates from hospital mortality data, outlined in detail in section 9.3.3 . We separately identified road accident fatalities in this dataset when calculating the drug-related mortality rates. We scaled the Ministry of Transport road fatality data to be consistent these figures. Homicides committed as a result of alcohol and drug use were estimated using the method detailed in section 9.4.1 on justice sector costs. The resulting estimated deaths were multiplied by the value of life lost to result in an estimated cost of $1,906.5 million, of which $1,519.9 million was for alcohol, $386.5 million was for other drugs. The remaining balance was for joint AOD use. The main contributor was premature mortality, which cost $1,133 million for alcohol and $319 million for drugs. Road fatalities cost $384 million for alcohol and $66 million for drugs, while homicide cost $2 million for alcohol and $0.2 million for drugs. Lost quality of life Lost quality of life was calculated for AOD-related road crash injuries and victims of crime. The number of road crash injuries was estimated from Ministry of Transport data on crashes where drugs were a contributing factor. The reported number of non-fatal road crash injuries were scaled using factors from the Ministry of Transports (2006) social cost analysis to allow for under-reporting. The scaling factors for serious and minor crashes were 1.71 and 3.74. The Ministry of Transport (2006) value the pain and suffering resulting from a serious traffic accident at one tenth of the value of a lost life and from a minor accident is valued at 0.4 percent. This study applies these proportions to the value of a statistical life year. Lost quality of life suffered by victims of crime was based on the estimated number of AODrelated offences and the Treasurys (2006) report on the costs of crime. The VOSL-based intangible costs per offence, by offence type, were scaled so that they are consistent with the prevalence approach and life year values used in this research. In particular, the cost per homicide is set at $106,600 and the average intangible cost, for example, of other violent offences is $244.
134
The average intangible cost of property damage and theft, which
reflects the fear and stress induced by such crime, is $22 per offence.
estimated in New Zealand by the transport sector government agencies. Easton estimated for a 10 percent discount rate a value of NZ$200,000 for 1990, which in 2005 values would be about NZ$275,000. O'Dea considered a value of the order of at least NZ$100,000 in 2005, using a 3.5 percent discount rate, to be appropriate. Abelson proposes a value of AU$100,000 for Australia in 2002. Note that results of the calculations are sensitive to the chosen discount rate - the higher the assumed discount rate, the higher the estimated value of a life-year.
134
The average intangible cost of violent offences borne in a given year covers assaults ranging across threatening behaviour to grievous assault.
124
9.4.7 Avoidable cost analysis Estimates of the total costs of drug misuse comprise avoidable and unavoidable costs. Avoidable costs are the proportion of total costs, or the burden of misuse, that could in principle be changed given the implementation of appropriate public policies and behaviour changes. We have taken Collins and Lapsley (2002) and examined proportions of avoidable costs to social costs for alcohol (tangible and intangible). Note that there are some gaps in the available fields, meaning not all avoidable costs may be covered. Individual avoidable cost components ranged from 12 percent to 90 percent. The Collins and Lapsley report also has avoidable costs for tobacco but not illegal drugs. Tobacco avoidable cost components ranged from 41 percent to 61 percent. We have reason to believe that this research is now outdated, however. According to the most recent reference material on avoidable costs (Collins and Lapsley 2008), which used Arcadian Normal and prevalence-based comparators approaches, around half of costs could be avoided in an Australian context, though this figure should be viewed merely as a feasible maximum of potential avoidable costs. No more detail is available at this time. Given the lack of robust information available to carry out a thorough avoidable cost analysis, we provide a rough estimate of costs that can be avoided. We use a simple approach: we assume that 50 percent of costs are avoidable and apply this to our estimate of total costs. This approach suggests that an estimated $4,340 million of social costs are avoidable. This is an area that we recommend should be a focus for further research. Specific focus should be made in estimating Arcadian Normal and prevalence approaches for a New Zealand context, and further deriving detailed breakdowns of avoidable costs from the results for areas such as crime and health. For further detail, refer to section 6.1. 9.4.8 Injury cost analysis This sub-analysis partitioned off those social costs due to injury resulting from harmful AOD use. This primarily included: health care for victims of crime (from the justice sector cost estimates) and people injured as a result of their AOD use (based on the health sector cost estimates using NZHIS data)
125
lost output due to injury-related mortality (based on the additional population there would have been in the absence of AOD-related fatal injuries), and absenteeism and reduced productivity (based on ACC data, discussed below)
road crash costs related to injury, such as lost output and medical care for crash victims (based on the injury-related components of the main road crash cost estimates)
intangible costs such as homicide, road crash and other fatal injures and reduced quality of life for road crash survivors (based on the injury-related premature mortality, road crashes and homicides from the main intangible cost estimates).
The injury cost estimates are mainly a sub-set of the main analysis components. One component prepared specifically for this analysis based on ACC data, and drawn into the main cost estimates, was the value of lost output for injured workers. ACC data and attribution to harmful AOD use ACC provided data to BERL on the number and cost of new and on-going claims in 2005/06 and public health acute service (PHAS) funding. Claims worth $1,973 million were broken down by the six New Zealand Injury Prevention Strategy (NZIPS) categories: assault, drowning, falls, motor vehicle, other, suicide and self harm and workplace. The PHAS funding of $315.0 million in 2005/06 covered inpatient, emergency department, outpatient, other miscellaneous hospital services, cost adjusters, pharmaceuticals and laboratory services. The analysis of this data concentrated on inpatient and emergency department funding as discussed below in the sub-section on ACC costs. Lost output Output lost as result of workers suffering AOD-related injuries was based on ACC data on workers compensation payments for 2005/06. Total workers compensation payments in 2005/06 totalled $722.0 million. AOD attributable fractions were applied to the relevant NZIPS categories to determine the portion of total ACC workers compensation payments that were due to AOD-related injuries. The corresponding figures were $95.2 million and $4.3 million for alcohol and other drugs, respectively. The total of $99.5 million represents almost one seventh of ACC workers compensation payments.
135
135
The main absenteeism estimate reflects lost production due to people taking time off as they are unwell rather than injured.
126
ACC pays 80 percent of a workers taxable income as compensation. As such, the figures were scaled up by one quarter to reflect that the social cost of the lost output is greater than 80 percent compensation figure. This resulted in a total of $124.3 million corresponding to ACC workers compensation. Furthermore, workers are not entitled to compensation from ACC in their first week off work (or on reduced pay). As such, an estimate of the output lost in the affected workers first week was added.
136
This was based on the estimated number of people affected (almost
13,000 alcohol drinkers and 330 other drug users) times the average weekly wage of $610 in 2005/06. We added the $8.1 million of output lost in a workers first week affected to the $124.3 million related to subsequent weeks affected. Output lost due to AOD injury-related time off work and reduced productivity equated to an estimated $132.5 million. Of this, 95.8 percent was attributable to harmful alcohol use and 4.2 percent to other drug-related injuries. ACC costs In addition to the sub-analysis of AOD-related injury costs, we examined which of these costs would likely be borne by ACC. AOD attributable fractions were applied to the relevant NZIPS categories to determine the portion of total ACC funding due to harmful AOD use. This analysis covered expenditure on ambulance services, elective hospital and medical treatment, benefit and lump sum payments, as well as support for vocational rehabilitation and weekly workers compensation. It was not possible to allocate the PHAS funding categories to the NZIPS categories, nor to use the AOD attributable fractions to determine the portion attributable to harmful AOD use. Instead, we concentrated on acute inpatient and emergency department funding. The calculations based on the NZIPS data indicated that approximately 12.5 percent of elective hospital treatment expenditure was AOD-related. We applied this percentage to the total PHAS acute inpatient and emergency department funding lines to estimate the portion attributable to harmful AOD use. We did not include estimates for PHAS funded outpatient, pharmaceuticals or laboratory services in the studys main social cost estimates as applying the inpatient proportion was not likely to yield a reliable estimate. We made rough estimates for these components to provide a sense of their possible magnitude. We applied the proportion used to estimate
136
This estimate is likely to be conservative, as we were not able to robustly estimate the value of output lost by injured workers who do not claim ACC workers compensation, f or example, people affected for less than one week.
127
inpatient and emergency department funding to total PHAS funding for outpatient, pharmaceuticals or laboratory services. These calculations suggest ACC could have spent an additional $1.30 million on outpatient services, $1.17 million on pharmaceuticals and $82,000 on laboratory services due to harmful AOD use. These latter estimates are indicative figures only as we are not confident that the assumption used generate robust estimates, and they have not been included in the main estimates. 9.4.9 Costs to government The sections below detail how we estimated the impacts on government expenditure, the revenue impacts from a reduced population and reduced output, and the impacts on consumption-related tax revenue. Impacts on government expenditure Government expenditure is estimated for the justice and health sectors from budget Vote information and BERLs main estimates, and for emergency services attending drug -related road crashes. The justice sector estimates include expenditure by: the Ministry of Justice the New Zealand Customs Service the New Zealand Police (and Defence Force support) the Department of Courts the Department of Corrections.
Health sector estimates include expenditure by: the Ministry of Health ACC Pharmac.
Government funded areas within the health sector include: community and preventative care government subsidised primary care government subsidised pharmaceuticals A&E and public hospitals
128
ambulances.
Impacts of lost output Estimating the labour market-related costs from a government perspective involved the following elements. Premature mortality individual and business tax lost less unemployment benefits saved Excess unemployment individual and business tax lost plus unemployment benefits incurred Absenteeism individual and business tax lost Reduced productivity business tax lost
138 137
The main estimates of lost output for each of these components were used to estimate the lost tax revenue. Total lost output was decomposed into wage and residual value added components, using the assumption that the average Gross Domestic Product (GDP) per Full Time Equivalent (FTE) is 1.87 times the average wage income. A similar process was used to determine the impacts from a government perspective of output losses suffered by victims of crime. Average tax rates were applied to the wage and value added components of lost output to identify the indirect cost to government of reduced output. In the case of individual tax, data from the 2006 Census was used to estimate a (weighted) average tax rate of 19.1 percent paid by wage and salary earners. Business tax was based on the prevailing 2005/06 rate of 33 percent. The impact of drug use on unemployment benefits was calculated by taking the number of people affected times the average annual unemployment benefit payment. The main labour market estimates were used for the number of people affected. In the case of those who died prematurely, but who would have been unemployed, the calculated impacts represent a cost saving in that these payments did not have to be made. In the case of excess unemployment, benefit payments to people who would have been in employment if not for their harmful drug use represents a cost to government.
137
We assume that leave taken as a result of harmful drug use is unpaid leave, resulting in a loss of wages, and correspondingly a loss in income tax revenue.
138
We assume that workers suffering reduced productivity as a result of harmful drug use receive their usual wage, so there is no impact on income tax revenue, but as businesses suffer reduced output this results in lower business tax payments.
129
Data from the 2006 Census indicate that, on average, those people eligible for a Job Seekers Community Wage (unemployment benefit) at some point during the year received $6,270. This figure is lower than the annualised unemployment benefit as a large proportion of people do not receive this benefit for an entire year. Appendix Table 24 and Appendix Table 25 provide more detailed breakdowns of the lost output impacts on business and government. Impacts on consumption-related taxes In addition to reducing output, premature mortality prevents people spending their income, which would have generated GST revenue. This reduction in GST revenue is estimated by applying the GST rate of 12.5 percent to the main estimate of the reduction in consumption expenditure due to premature mortality. Illegal drug use diverts expenditure from taxable, market activity reducing the governments revenue from GST. To estimate this impact, we assume that the estimated expenditure on illegal drugs of $518.7 million would otherwise have been spent on legal purchases, and would be liable for GST of 12.5 percent. We assume that under the counterfactual harmful alcohol expenditure would be diverted to other market expenditure so there is no net impact on GST from this change. Reducing harmful alcohol use reduces the quantity of alcohol liable for customs and excise duties. This impact is estimated by determining the share of total alcohol consumed that is consumed in a harmful manner, and applying this to the amount of alcohol excise tax and customs duty collected. Although approximately 50 percent of alcohol is estimated to be used in a harmful manner, we assume that harmful drinkers would reduce their consumption to the low risk drinking level rather than stopping altogether. In this case, alcohol consumption is estimated to fall by 36 percent, which is the proportion used in these calculations.
130
10
Appendix Additional Tables

Appendix Table 1 Harmful drug use cost inclusions and exclusions
Range of costs due to harmful drug use
C&L
DHI
AOD
(2008) (BERLa) (BERLb) Tangible costs - direct Costs of abused substances/ diverted inputs Road crashes (emergency services, property damage, travel delays, insurance) Customs and immigration Police Defence force support for Police drug eradication programmes Courts Prisons Community sentences for offenders Pharmaceuticals (for medical treatment) Primary care Ambulance services Hospitalisation Health care for road crash victims Health care for crime victims Nursing home stays Community care Professional services (other than physicians) Specialty institutions (including substance abuse treament facilities) Special equipment for rehabilitation (e.g. wheelchairs) Prevention programs Training costs for physicians and nurses On-going impacts of neonatal disorders and complications related to drug abuse Home care (care of drug user and their house) Retraining and re-education Employee assistance programs provided by employers for drug-using employees Drug testing in workplace Avoidance behaviour costs Administrative costs of private insurance to treat drug disorders Disability pension due to retirement for health reasons Payroll taxes on earnings that finance insurance or other benefits Tangible costs - indirect Labour costs - reduced workforce Labour costs - excess unemployment Labour costs - absenteeism Labour costs - reduced productivity Foregone consumption Reduced property values in drug-ridden communities Intangible costs Pain and suffering - road crash victims Pain and suffering - victims of crime Pain and suffering - related to drug users' health Value of lost life to the deceased Homelessness associated with illicit drug abuse Family disruption Community disruption Other costs related to drug harm Policy costs Research and education General equilibrium (economy-wide) impacts of harmful drug use Adapted from Choi, Robson and Single (1997) C&L = Collins and Lapsley (2008) The costs of tobacco, alcohol and illicit drug abuse to Australian society BERLa = BERL (2008a) New Zealand Drug Harm Index BERLb = BERL (2008b) Costs of Harmful Alcohol and Other Drug Use
131
Appendix Table 2 Tangible costs of harmful drug use ($m), 2005/06 detail
Tangible costs ($m) Alcohol Lost output Market labour Reduction in workforce Excess unemployment (net) Absenteeism Reduced productivity Losses due to injury (n.i.e.) Household labour Premature death Sickness Reduced productivity Road crash related lost output (temporary) Crime related lost output Victims of crime Prisoners Home detainees Total paid and unpaid labour costs Less consumption resources saved Total net labour costs 156.9 36.0 0.7 1,786.7 -308.4 1,478.3 Diverted inputs Total diverted inputs 342.2 Crime n.i.e. Customs Police Defence force Criminal courts Prisons Community sentences Preventative expenditure Property lost Total crime n.i.e. 0.1 172.2 0.0 64.5 112.3 18.7 61.3 133.1 562.2 Health care (net) Pharmaceuticals Hospital Medical Ambulances Treatment for victims of crime Total net health care 0.3 121.0 66.8 4.2 97.8 290.2 Road crashes n.i.e. Property damage Travel delays Insurance administration Fire/emergency service Total road crashes n.i.e. Total tangible
Other drugs Joint AOD
Total
459.2 877.5 36.0 33.9 126.9 53.7 1.7 2.3 2.0
141.3 0.0 8.5 0.0 5.5 11.3 0.3 0.0 0.0 21.9 47.9 0.6 237.3 -79.7 157.6
0.0 227.5 11.9 8.7 0.0 0.0 0.7 0.8 0.0 31.5 34.0 0.6 315.7 0.0 315.7
600.4 1,105.0 56.5 42.6 132.4 65.1 2.6 3.1 2.1 210.3 117.8 1.9 2,339.7 -388.1 1,951.6
518.7
0.0
861.0
23.8 90.7 0.6 34.0 87.7 18.2 13.9 41.2 309.9
0.0 43.5 0.0 16.3 95.9 13.6 21.6 48.5 239.3
23.9 306.3 0.6 114.8 295.8 50.5 96.8 222.8 1,111.4
11.8 9.2 12.7 0.5 9.0 43.2
0.0 0.0 77.9 0.0 16.8 94.8
12.1 130.3 157.4 4.7 123.7 428.2
21.7 26.8 132.2 19.4 202.1 2,875.1
0.6 0.6 3.2 0.5 4.9 1,034.2
0.3 0.3 1.6 0.2 2.4 652.1
22.6 27.7 137.0 20.1 209.4 4,561.5
132
Appendix Table 3 Intangible costs of harmful drug use ($m), 2005/06 detail
Intangible costs ($m)
Alcohol
Other drugs
Joint AOD
Total
Loss of life Premature mortality n.i.e. Road fatalities Homicide Total loss of life Road crashes Crime Total pain and disability Total intangible 1,132.9 384.4 2.34 1,519.6 8.78 33.6 42.3 1,561.9 318.8 67.5 0.17 386.5 0.21 5.8 6.0 392.4 0.0 0.0 0.38 0.4 0.15 8.2 8.3 8.7 1,451.8 451.8 2.89 1,906.5 9.14 47.5 56.6 1,963.1
Lost quality of life
Note: n.i.e. denotes not included elsewhere. Source: BERL
Appendix Table 4 Drug use prevalence 13+ year olds, 2005/06
Total Both 19,100 85,800 88,000 77,200 56,600 62,400 65,400 53,900 112,000 46,500 666,900
Source: BERL, StatsNZ.
133
Appendix Table 5 Total alcohol caused deaths by nature of cause, 2001-2005

Cause of deaths from harmful alcohol use Alcohol dependence/abuse Alcoholic beverage & other EtOH poisoning Alcoholic cardiomyopathy Alcoholic gastritis Alcoholic liver cirrhosis Alcoholic poly-neuropathy Alcoholic psychosis Aspiration Child abuse & Assault Drowning Fall injuries Female breast cancer Fire injuries Gastro-oesophageal haemorrhage Heart failure Hypertension Laryngeal cancer Liver cancer Occupational and machine injuries Oesophageal cancer Oesophageal varices Oropharyngeal cancer Pancreatitis - acute/chronic Road injuries Stroke - haemorrhagic/ischaemic Suicide and Self Injury All Causes
Source: NZHIS, BERL.
Females 29.0 8.0 10.0 1.0 93.0 0.0 10.0 2.6 35.0 5.7 107.2 182.4 17.6 1.4 0.0 0.3 4.8 44.1 13.0 73.7 0.7 30.2 21.7 61.0 68.8 86.1 907.3
Males 58.0 19.0 105.0 6.0 286.0 1.0 35.0 5.3 59.3 32.9 101.3 0.0 22.3 1.4 8.5 9.7 42.8 123.1 50.8 192.7 0.0 111.5 19.9 416.2 156.7 420.8 2285.1
Total 87.0 27.0 115.0 7.0 379.0 1.0 45.0 7.9 94.3 38.6 208.5 182.4 40.0 2.8 8.5 10.0 47.5 167.2 63.8 266.4 0.7 141.8 41.6 477.2 225.5 506.8 3192.4
134
Appendix Table 6 Alcohol caused deaths and age-gender mortality rates, 2001-2005
Deaths Age 0-4 5-9 10-14 15-19 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85+ All Ages
Sources:
Mortality Rates per 1000 Females 0.039 0.038 0.015 0.138 0.216 0.165 0.223 0.252 0.286 0.454 0.497 0.498 0.732 0.670 1.012 1.012 1.898 5.835 0.443 Males 0.085 0.060 0.074 0.725 1.047 1.039 0.867 0.953 1.020 1.196 1.342 1.695 1.989 2.858 3.690 3.638 4.097 6.464 1.160 12.3 9.0
Females 5.4 5.4 2.2 20.0 29.5 21.6 34.4 40.1 46.1 64.3 62.6 53.6 64.0 47.3 63.8 55.7 78.8 212.4 907.3
Males
11.8 108.2 144.7 129.1 122.2 141.0 155.3 164.0 166.2 179.7 168.9 191.0 211.8 161.2 107.0 101.7 2285.1
NZHIS mortality data.Selected ICD-10 AM codes. 2001-05 Population data. StatsNZ. Estimated resident population, mean calendar years.
Appendix Table 7 Total other drug caused deaths by nature of cause, 2001-2005
Cause of deaths from other drug use Accidental opiate poisoning Amphetamine dependence Hepatitis B Hepatitis non A, and B HIV/AIDS Infective endocarditis Low birthweight Opiate abuse Opiate dependence Opiate poisoning Road injuries Suicide and Self Injury All Causes
Females 24.0 0.0 10.5 0.4 4.6 0.4 1.1 0.0 21.0 30.0 31.5 43.8 167.3
Males 48.0 1.0 19.6 0.0 3.4 0.7 1.1 4.0 36.0 57.0 66.4 147.7 384.9
Total 72.0 1.0 30.0 0.4 8.0 1.1 2.2 4.0 57.0 87.0 97.9 191.5 552.2
135
Appendix Table 8 Other drug deaths and age-gender mortality rates, 2001-2005
Deaths Age 0-4 5-9 10-14 15-19 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85+ All Ages
Sources:
Mortality Rates per 1000 Females 0.008 0.000 0.000 0.115 0.093 0.122 0.113 0.157 0.130 0.080 0.121 0.066 0.046 0.063 0.097 0.050 0.112 0.052 0.082 Males 0.008 0.000 0.000 0.190 0.328 0.356 0.367 0.383 0.265 0.208 0.182 0.198 0.175 0.131 0.125 0.157 0.155 0.210 0.195 1.1 0.0 0.0 28.4 45.3 44.2 51.7 56.7 40.4 28.5 22.5 21.0 14.9 8.8 7.2 6.9 4.1 3.3
Females 1.1 0.0 0.0 16.6 12.7 16.0 17.5 24.9 20.9 11.4 15.2 7.2 4.0 4.4 6.1 2.8 4.7 1.9 167.3
Males
384.9
NZHIS mortality data.Selected ICD-10 AM codes. 2001-05 Population data. StatsNZ. Estimated resident population, mean calendar years.
136
Appendix Table 9 Counterfactual population estimates males, 2005/06

Males Age Group 0-4 5-9 10-14 15-19 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90+ All males
Estimated resident population
146,420 149,230 159,500 159,470 146,030 127,100 136,420 149,190 156,030 149,130 129,950 119,820 92,300 75,650 57,470 47,590 30,070 4,990 13,180 2,062,720
Estimated resident Additional population population - no harmful alcohol - no harmful alcohol or drug use use 146,427 6 149,248 159,529 159,577 146,290 127,519 137,027 149,997 157,027 150,276 131,184 121,164 93,626 76,965 58,694 48,674 30,855 18,631 13,357 2,076,067 17 28 91 209 327 464 606 748 875 960 1,067 1,068 1,081 1,032 943 700 420 165 10,806
Additional population - no other drug use 1 1 1 16 52 91 143 201 249 272 274 276 259 234 192 141 86 40 13 2,541
137
Appendix Table 10 Counterfactual population estimates females, 2005/06

Females Age Group 0-4 5-9 10-14 15-19 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90+ All females
139,830 142,610 150,990 154,280 145,540 133,940 149,830 163,430 166,650 156,070 133,750 122,710 95,110 80,000 62,800 56,060 43,900 25,940 13,950 2,137,390
Estimated resident Additional population population - no harmful alcohol - no harmful alcohol or drug use use 139,833 3 142,619 151,003 154,314 145,610 134,044 149,985 163,646 166,927 156,403 134,126 123,126 95,524 80,418 63,201 56,446 44,248 26,290 14,155 2,141,918 8 12 24 46 68 100 139 178 224 264 302 309 321 317 318 299 321 192 3,444
138
Appendix Table 11 Counterfactual population estimates total, 2005/06

Total Age Group 0-4 5-9 10-14 15-19 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90+ Total
286,250 291,840 310,490 313,750 291,570 261,040 286,250 312,620 322,680 305,200 263,700 242,530 187,410 155,650 120,270 103,650 73,970 30,930 27,130 4,200,110
Estimated resident Additional population population - no harmful alcohol - no harmful alcohol or drug use use 286,260 9 291,867 310,532 313,891 291,900 261,563 287,011 313,644 323,954 306,680 265,309 244,289 189,150 157,383 121,896 105,120 75,104 44,921 27,512 4,217,985 25 40 114 255 395 564 745 927 1,099 1,224 1,369 1,376 1,402 1,349 1,261 999 741 357 14,251
Appendix Table 12 NZ-ADAM distribution of crime by offence category and drug type, 2005/06
Offence category Violent % offences attributable to 1.5% Drugs only Alcohol only A&OD only Total AOD 20.9% 3.4% 25.8% Property 2.9% 23.9% 8.1% 34.9% Illegal drugs 89.2% 0.0% 10.8% 100.0% Traffic 2.6% 0.0% 4.5% 7.1% Drink/ Breaches Disorder Drugged Driving 4.8% 17.3% 4.8% 26.9% 0.7% 33.1% 2.7% 36.5% 0.26% 99.74% 0.0% 100.0% Other 6.2% 11.8% 4.7% 22.7%
139
Appendix Table 13 Crime multipliers to estimate actual crime from recorded crime, 2003/04
Offence category Homicide Grievous assaults Intimidation & threats Kidnapping & abduction Other assaults Sexual violation Other sexual offences
Source: Treasury.
Multiplier 1.00 1.80 7.70 1.00 7.70 1.50 6.50
Offence category Robbery Burglary Theft of vehicles Theft from vehicles Other theft & receiving Property damage Fraud
Multiplier 3.70 2.20 1.05 2.80 8.28 4.30 4.00
Appendix Table 14 NZP offence codes and offence categories
Police offences Violent Sexual Drugs (not cannabis) Drugs (cannabis) Anti social Dishonesty Property Damage New Drugs Property Abuse Admin Road policing - other
NZOH code series 1000 2000 3100 3200 3400-3900 4000 5000 5900 6000 7000
Offence category Violence Other Drugs Drugs Disorder Other Property Drugs Property Breaches Traffic Drink/drugged driving
Road policing - drink/drugged driving

Source: NZP.
140
Appendix Table 15 NZ-ADAM offence categories (HOI)
Category
Offence
Category Violence
Offence Homicide Kidnapping and abduction Robbery Assaults Intimidation/ threats
Breaches (Administrative) Against justice Immigration Disorder Drugs Drugs (cannabis only) Drugs (not cannabis) Family offences Property Destruction of property Trespass Littering Animals Traffic
Source: NZ-ADAM.
Other
Burglary Car conversion etc Theft Receiving Fraud Sexual affronts/ attacks Immoral behaviour
141
Appendix Table 16 Lost output due to harmful drug use ($m), 2005/06 - detail
Labour costs ($m) Market output Male Female Total reduction in workforce Male Female Total excess unemployment (net) Male Female Total absenteeism Male Female Total reduced productivity Total paid labour costs Household output Premature death Male Female Total premature death Male Female Total sickness Male Female Total reduced productivity Total unpaid labour costs Other labour costs Losses due to injury (n.i.e.)
Road accidents lost output
Alcohol
Other drugs
Joint AOD
Total
Reduction in workforce 371.7 87.4 104.2 37.0 0.0 0.0 0.0 164.9 62.6 227.5 6.5 5.5 11.9 6.3 2.4 8.7 248.2 475.9 124.5 600.4 721.1 383.9 1,105.0 38.2 18.3 56.5 27.8 14.8 42.6 1,936.9
459.2 141.3 Excess unemployment (net) 556.2 321.3 877.5 Absenteeism 23.9 12.1 36.0 Reduced productivity 21.4 12.4 33.9 1,533.4 0.0 0.0 0.0 7.8 0.7 8.5 0.0 0.0 0.0 155.3
33.0 20.8 53.7 Sickness 0.9 0.8 1.7 Reduced productivity 1.1 1.3 2.3 57.8
6.4 5.0 11.3 0.2 0.0 0.3 0.0 0.0 0.0 11.6
0.0 0.0 0.0 0.3 0.4 0.7 0.4 0.4 0.8 1.5
39.4 25.7 65.1 1.4 1.2 2.6 1.5 1.6 3.1 70.8
126.9 2.0 193.6 195.6 1,786.7
5.5 0.0 70.3 70.4 237.3
0.0 0.0 66.0 66.0 315.7
132.4 2.1 329.9 332.0 2,339.7
Crime related lost output Total other labour costs Total paid and unpaid labour costs
Consumption resources saved Market Household Total consumption resources saved Total net labour costs % Total net labour costs
Source: BERL
-288.1 -20.3 -308.4 1,478.3 75.8%
-75.4 -4.3 -79.7 157.6 8.1%
0.0 0.0 0.0 315.7 16.2%
-363.5 -24.6 -388.1 1,951.6 100.0%
142
Appendix Table 17 Justice sector costs of harmful drug use ($m), 2005/06 - detail
Crime costs ($m) Alcohol Other drugs Joint AOD Total
Tangible costs Customs Police Defence force Criminal courts Prisons Community sentences Preventative expenditure Property lost Health care of victims Lost output of victims Lost output of prisoners Lost output of home detainees Total tangible Homicide Pain and suffering Total intangible costs Total costs Tangible n.i.e. Intangible n.i.e. Total crime n.i.e. % of Total crime n.i.e.
Note: n.i.e. denotes not included elsewhere. Source: BERL
0.1 172.2 0.0 64.5 112.3 18.7 61.3 133.1 97.8 156.9 36.0 0.7 853.6 2.3 33.6 35.9 Intangible costs
23.8 90.7 0.6 34.0 87.7 18.2 13.9 41.2 9.0 21.9 47.9 0.6 389.2 0.2 5.8 5.9 395.2 309.9 309.9 27.9%
0.0 43.5 0.0 16.3 95.9 13.6 21.6 48.5 16.8 31.5 34.0 0.6 322.2 0.4 8.2 8.5 330.7 239.3 239.3 21.5%
23.9 306.3 0.6 114.8 295.8 50.5 96.8 222.8 123.7 210.3 117.8 1.9 1,565.0 2.9 47.5 50.4 1,615.4 1,111.4 1,111.4 100.0%
Total costs of crime 889.5 562.2 562.2 50.6%
143
Appendix Table 18 Health sector costs of harmful drug use ($m), 2005/06 - detail
Health care ($m) Hospitals Pharmaceuticals - treatment Primary (non-A&E) Other community care A&E/non-hospital admission Ambulances Treatment for victims of crime Total health Total health n.i.e.
Alcohol 121.0 0.3 61.4 0.0 5.4 4.2 97.8 290.2 192.4
Other drugs 9.2 11.8 0.8 11.7 0.2 0.5 9.0 43.2 34.2
Joint AOD 0.0 0.0 -3.3 81.2 0.1 0.0 16.8 94.8 77.9
Total 130.3 12.1 58.8 92.9 5.7 4.7 123.7 428.2 304.5
Note: n.i.e denotes not included elsewhere. Source: BERL.
144
Appendix Table 19 Hospital costs due to alcohol use by category, 2001-2006

Cost Categories due to Alcohol Use Directly attributable to alcohol Alcohol dependence/abuse F10.0-10.2 5,505 385 161 386 2,767 40 3,730 1 12,976 6.38 ICD-10 code Case-weight sums six-year total Annual average cost ($m 2005/06)
Alcoholic beverage & other EtOH poisoning X45, Y15, T151.0-151.1, T51.9 Alcoholic cardiomyopathy Alcoholic gastritis Alcoholic liver cirrhosis Alcoholic poly-neuropathy Alcoholic psychosis Fetal alcohol syndrome (dysmorphic) Sub-total Other related causes Child abuse and assault Cholelithiasis Fall injuries Female breast cancer Fire injuries Gastro-oesophageal haemorrhage Heart failure Hypertension Ischaemic heart disease Laryngeal cancer Liver cancer Occupational and machine injuries Oesophageal cancer Oesophageal varicies Oropharyngeal cancer Pancreatitis - acute/chronic Road injuries Self-harm or accidental harm Stroke - haemorrhagic/ ischaemic Supraventricular cardiac dysrhythmias Unspecified liver cirrhosis Sub-total Total
I42.6 K29.2 K70 G62.1 F10.3-10.9 Q86.0
X85-Y09, Y87.1 K80 W00-19, M80-82 C50 X00-19 K22.6 I50-51, I97.1 I11, I13.0, I15 I20-25 C32 C22 V codes, refer App Table 16 C15 I85, I98.2 C00-C14 K85, K86.0-86.1 Several V codes, refer App Table 16 G45, I60-69 I47.1, I47.8-48.9 K74.3-74.6, K76.0, K76.9
9,200 0 40,645 1,438 826 508 379 25 0 1,054 819 115,964 1,825 193 2,469 3,332 15,620 5,711 1,904 0 232 202,144 215,120 99.36 105.73
145
Appendix Table 20 Hospital costs due to other drug use by category, 2001-2006
Cost Categories due to Other Drug Use Directly attributable to opiates Opiate dependence Opiate abuse Opiate poisoning Antepartum haemorrhage due to opiates Low birthweight due to opiates Sub-total Directly attributable to other drugs Cannabis dependence Cannabis abuse Amphetamine dependence Amphetamine abuse Cocaine dependence Cocaine abuse Psychostimulant poisoning Hallucinogen dependence Hallucinogen abuse Hallucinogen poisoning Other psychotropic drug poisoning Anabolic steroid poisoning Antepartum haemorrhage due to cocaine Low birthweight due to cocaine Sub-total Attributable to unclassifiable injecting drug use Hepatitis B Hepatitis C HIV/AIDS Infective endocarditis Sub-total Other related causes Drug psychoses Maternal drug dependence Newborn drug toxicity Road injuries Self-harm or accidental harm Schizophrenia Sub-total Total Notes: Antepartum haemorrhage split between opiates and cocaine in ratio 0.01225 / 0.03725 from Ridolfo & Stevenson (2001) Low birthweight split between opiates and cocaine in ratio 0.02125 / 0.02575 from Ridolfo & Stevenson (2001)
ICD-10 code
Case-weight sums Annual average six-year total cost ($m 2005/06) 551 72 1,620 109 358 2,708 1.33
F11.2-F11.4 F11.0-F11.1 T40.0-T40.3 O20, O44.1, O45-O46, O67, P02.0-P02.1 P05, P07, P22
F12.2-F12.4 F12.0-F12.1 F15.2-F15.4 F15.0-F15.1 F14.2-F14.4 F14.0-F14.1 T43.6 F16.2-F16.4 F16.0-F16.1 T40.7-T40.9 T43.8-T43.9 T38.7 O20, O44.1, O45-O46, O67, P02.0-P02.1 P05, P07, P22
323 397 37 52 0 0 416 2 11 91 42 4 331 433 2,140 6.31
B16, B17.0, B18.0-B18.1 B17.1 B20-B24, R75, Z20.6,Z21 I33
545 221 189 304 1,259 3.71
F11-F19 O35.5 P04.4, P96.1 Several V codes, refer App Table 16 F20-F29
3,049 15 938 4,566 2,441 10 11,019 17,126 32.49 43.85
146
Appendix Table 21 Road crash costs due to harmful drug use ($m), 2005/06 - detail
Tangible costs of road crashes ($m) Hospital/medical Emergency/pre-hospital Follow-on Loss of output (fatality) Loss of output (temporary disability) Legal and court Property damage Travel delays Insurance administration Fire/emergency service Total tangible road crash costs % of tangible costs Total costs n.i.e. Alcohol 6.9 3.9 3.5 83.7 2.0 4.4 21.7 26.8 132.2 19.4 304.5 93.5% 202.1 Other drugs Joint AOD 0.3 0.1 0.1 12.8 0.0 0.3 0.6 0.6 3.2 0.5 18.5 5.7% 4.9 0.2 0.1 0.1 0.0 0.0 0.1 0.3 0.3 1.6 0.2 2.8 0.9% 2.4 Total 7.4 4.1 3.7 96.5 2.1 4.8 22.6 27.7 137.0 20.1 325.8 100.0% 209.4
Intangible costs of road crashes ($m) Loss of life Pain and suffering Total intangible road crash costs % of intangible costs
Source: BERL
Alcohol 384.4 8.8 393.1 85.3%
Other drugs Joint AOD 67.5 0.2 67.7 14.7% 0.0 0.2 0.2 0.0%
Total 451.8 9.1 461.0 100.0%
147
Appendix Table 22 Drug-attributable morbidity and mortality health conditions The following list of ICD-10-AM 1 edition codes was used to assess drug-attributable morbidity and mortality. Data were supplied by the NZHIS on Deaths Registration data (Tables A and E) and Publicly Funded Hospital Discharges data (Tables A, B or V).
Table A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A Code B16 B17.0 B18.0 B18.1 B20 B21 B22 B23 B24 C07 C08.0 C08.1 C11.0 C11.1 C11.2 C11.3 C11.8 C11.9 F11.0 F11.1 F11.2 F11.3 F11.4 F11.5 F11.6 F11.7 F11.8 F11.9 F12.0 F12.1 F12.2 Description Acute hepatitis B Acute delta-(super)infection of hepatitis B carrier Chronic viral hepatitis B with delta-agent Chronic viral hepatitis B without delta-agent Human immunodeficiency virus [HIV] disease resulting in infectious and parasitic diseases Human immunodeficiency virus [HIV] disease resulting in malignant neoplasms Human immunodeficiency virus [HIV] disease resulting in other specified diseases Human immunodeficiency virus [HIV] disease resulting in other conditions Unspecified human immunodeficiency virus [HIV] disease Malignant neoplasm of parotid gland Malignant neoplasm of submandibular gland Malignant neoplasm of sublingual gland Malignant neoplasm of superior wall of nasopharynx Malignant neoplasm of posterior wall of nasopharynx Malignant neoplasm of lateral wall of nasopharynx Malignant neoplasm of anterior wall of nasopharynx Overlapping malignant lesion of nasopharynx Malignant neoplasm of nasopharynx, unspecified Mental and behavioural disorders due to use of opioids, acute intoxication Mental and behavioural disorders due to use of opioids, harmful use Mental and behavioural disorders due to use of opioids, dependence syndrome Mental and behavioural disorders due to use of opioids, withdrawal state Mental and behavioural disorders due to use of opioids, withdrawal state with delirium Mental and behavioural disorders due to use of opioids, psychotic disorder Mental and behavioural disorders due to use of opioids, amnesic syndrome Mental and behavioural disorders due to use of opioids, residual and late-onset psychotic disorder Mental and behavioural disorders due to use of opioids, other mental and behavioural disorders Mental and behavioural disorders due to use of opioids, unspecified mental and behavioural disorder Mental and behavioural disorders due to use of cannabinoids, acute intoxication Mental and behavioural disorders due to use of cannabinoids, harmful use Mental and behavioural disorders due to use of cannabinoids, dependence syndrome
st
148
Table A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A
Code F23.8 F23.9 F24 F25.0 F25.1 F25.2 F25.8 F25.9 F28.8 F29.9 I33 I51.0 I51.1 I51.2 I51.3 I51.4 I51.5 I51.6 I51.7 I51.8 I51.9 I97.1 K76.0 K76.9 M80 M81 M82 O20 O35.5 O44.1 O45 O46 O67 P02.0 P02.1 P04.4 P22.0
Description Other acute and transient psychotic disorders Acute and transient psychotic disorder, unspecified Induced delusional disorder Schizoaffective disorder, manic type Schizoaffective disorder, depressive type Schizoaffective disorder, mixed type Other schizoaffective disorders Schizoaffective disorder, unspecified Other nonorganic psychotic disorders Unspecified nonorganic psychosis Acute and subacute endocarditis Cardiac septal defect, acquired Rupture of chordae tendineae, not elsewhere classified Rupture of papillary muscle, not elsewhere classified Intracardiac thrombosis, not elsewhere classified Myocarditis, unspecified Myocardial degeneration Cardiovascular disease, unspecified Cardiomegaly Other ill-defined heart diseases Heart disease, unspecified Other functional disturbances following cardiac surgery Fatty (change of) liver, not elsewhere classified Liver disease, unspecified Osteoporosis with pathological fracture Osteoporosis without fracture Osteoporosis in diseases Haemorrhage in early pregnancy Maternal care for (suspected) damage to fetus by drugs Placenta praevia with haemorrhage Premature separation of placenta [abruptio placentae] Antepartum haemorrhage, not elsewhere classified Labour and delivery complicated by intrapartum haemorrhage, not elsewhere classified Fetus and newborn affected by placenta praevia Fetus and newborn affected by other forms of placental separation and haemorrhage Fetus and newborn affected by maternal use of drugs of addiction Respiratory distress syndrome of newborn
149
Table A A A A A B B B B B B B B B B B E E E E E E E E E E E E E E E E E E E E E
Code P22.1 P22.8 P22.9 P96.1 R75 T38.7 T40.0 T40.1 T40.2 T40.3 T40.7 T40.8 T40.9 T43.6 T43.8 T43.9 V01.1 V01.9 V02.1 V02.9 V03.1 V03.9 V04.1 V04.9 V09.2 V09.9 V12 V13 V14 V19.4 V19.5 V19.6 V20 V21 V22 V23 V24
Description Transient tachypnoea of newborn Other respiratory distress of newborn Respiratory distress of newborn, unspecified Neonatal withdrawal symptoms from maternal use of drugs of addiction Laboratory evidence of human immunodeficiency virus [HIV] Poisoning by Androgens and anabolic congeners Poisoning by Opium Poisoning by Heroin Poisoning by Other opioids Poisoning by Methadone Poisoning by Cannabis (derivatives) Poisoning by Lysergide [LSD] Poisoning by Other and unspecified psychodysleptics [hallucinogens] Poisoning by Psychostimulants with potential for use disorder Poisoning by Other psychotropic drugs, not elsewhere classified Poisoning by Psychotropic drug, unspecified Pedestrian injured in collision with pedal cycle, traffic accident Pedestrian injured in collision with pedal cycle, unspecified whether traffic or nontraffic accident Pedestrian injured in collision with two- or three-wheeled motor vehicle, traffic accident Pedestrian injured in collision with two- or three-wheeled motor vehicle, unspecified whether traffic or nontraffic accident Pedestrian injured in collision with car, pick-up truck or van, traffic accident Pedestrian injured in collision with car, pick-up truck or van, unspecified whether traffic or nontraffic accident Pedestrian injured in collision with heavy transport vehicle or bus, traffic accident Pedestrian injured in collision with heavy transport vehicle or bus, unspecified whether traffic or nontraffic accident Pedestrian injured in traffic accident involving other and unspecified motor vehicles Pedestrian injured in unspecified transport accident Pedal cyclist injured in collision with two- or three-wheeled motor vehicle Pedal cyclist injured in collision with car, pick-up truck or van Pedal cyclist injured in collision with heavy transport vehicle or bus Driver injured in collision with other and unspecified motor vehicles in traffic accident Passenger injured in collision with other and unspecified motor vehicles in traffic accident Unspecified pedal cyclist injured in collision with other and unspecified motor vehicles in traffic accident Motorcycle rider injured in collision with pedestrian or animal Motorcycle rider injured in collision with pedal cycle Motorcycle rider injured in collision with two- or three-wheeled motor vehicle Motorcycle rider injured in collision with car, pick-up truck or van Motorcycle rider injured in collision with heavy transport vehicle or bus
150
Table E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E
Code V25 V26 V27 V28 V29.4 V29.6 V30 V31 V32 V33 V34 V35 V36 V37 V38 V39.4 V39.5 V39.6 V40 V41 V42 V43 V44 V45 V46 V47 V48 V49.4 V49.5 V49.6 V50 V51 V52 V53 V54 V55 V56
Description Motorcycle rider injured in collision with railway train or railway vehicle Motorcycle rider injured in collision with other nonmotor vehicle Motorcycle rider injured in collision with fixed or stationary object Motorcycle rider injured in noncollision transport accident Driver injured in collision with other and unspecified motor vehicles in traffic accident Unspecified motorcycle rider injured in collision with other and unspecified motor vehicles in traffic accident Occupant of three-wheeled motor vehicle injured in collision with pedestrian or animal Occupant of three-wheeled motor vehicle injured in collision with pedal cycle Occupant of three-wheeled motor vehicle injured in collision with two- or three-wheeled motor vehicle Occupant of three-wheeled motor vehicle injured in collision with car, pick-up truck or van Occupant of three-wheeled motor vehicle injured in collision with heavy transport vehicle or bus Occupant of three-wheeled motor vehicle injured in collision with railway train or railway vehicle Occupant of three-wheeled motor vehicle injured in collision with other nonmotor vehicle Occupant of three-wheeled motor vehicle injured in collision with fixed or stationary object Occupant of three-wheeled motor vehicle injured in noncollision transport accident Driver injured in collision with other and unspecified motor vehicles in traffic accident Passenger injured in collision with other and unspecified motor vehicles in traffic accident Unspecified occupant of three-wheeled motor vehicle injured in collision with other and unspecified motor vehicles in traffic accident Car occupant injured in collision with pedestrian or animal Car occupant injured in collision with pedal cycle Car occupant injured in collision with two- or three-wheeled motor vehicle Car occupant injured in collision with car, pick-up truck or van Car occupant injured in collision with heavy transport vehicle or bus Car occupant injured in collision with railway train or railway vehicle Car occupant injured in collision with other nonmotor vehicle Car occupant injured in collision with fixed or stationary object Car occupant injured in noncollision transport accident Driver injured in collision with other and unspecified motor vehicles in traffic accident Passenger injured in collision with other and unspecified motor vehicles in traffic accident Unspecified car occupant injured in collision with other and unspecified motor vehicles in traffic accident Occupant of pick-up truck or van injured in collision with pedestrian or animal Occupant of pick-up truck or van injured in collision with pedal cycle Occupant of pick-up truck or van injured in collision with two- or three-wheeled motor vehicle Occupant of pick-up truck or van injured in collision with car, pick-up truck or van Occupant of pick-up truck or van injured in collision with heavy transport vehicle or bus Occupant of pick-up truck or van injured in collision with railway train or railway vehicle Occupant of pick-up truck or van injured in collision with other nonmotor vehicle
151
Table E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E
Code V57 V58 V59.4 V59.5 V59.6 V60 V61 V62 V63 V64 V65 V66 V67 V68 V69.4 V69.5 V69.6 V70 V71 V72 V73 V74 V75 V76 V77 V78 V79.4 V79.5 V79.6 V80.3 V80.4 V80.5 V81.1 V82.1 V83.0 V83.1 V84.0
Description Occupant of pick-up truck or van injured in collision with fixed or stationary object Occupant of pick-up truck or van injured in noncollision transport accident Driver injured in collision with other and unspecified motor vehicles in traffic accident Passenger injured in collision with other and unspecified motor vehicles in traffic accident Unspecified occupant of pick-up truck or van injured in collision with other and unspecified motor vehicles in traffic accident Occupant of heavy transport vehicle injured in collision with pedestrian or animal Occupant of heavy transport vehicle injured in collision with pedal cycle Occupant of heavy transport vehicle injured in collision with two- or three-wheeled motor vehicle Occupant of heavy transport vehicle injured in collision with car, pick-up truck or van Occupant of heavy transport vehicle injured in collision with heavy transport vehicle or bus Occupant of heavy transport vehicle injured in collision with railway train or railway vehicle Occupant of heavy transport vehicle injured in collision with other nonmotor vehicle Occupant of heavy transport vehicle injured in collision with fixed or stationary object Occupant of heavy transport vehicle injured in noncollision transport accident Driver injured in collision with other and unspecified motor vehicles in traffic accident Passenger injured in collision with other and unspecified motor vehicles in traffic accident Unspecified occupant of heavy transport vehicle injured in collision with other and unspecified motor vehicles in traffic accident Bus occupant injured in collision with pedestrian or animal Bus occupant injured in collision with pedal cycle Bus occupant injured in collision with two- or three-wheeled motor vehicle Bus occupant injured in collision with car, pick-up truck or van Bus occupant injured in collision with heavy transport vehicle or bus Bus occupant injured in collision with railway train or railway vehicle Bus occupant injured in collision with other nonmotor vehicle Bus occupant injured in collision with fixed or stationary object Bus occupant injured in noncollision transport accident Driver injured in collision with other and unspecified motor vehicles in traffic accident Passenger injured in collision with other and unspecified motor vehicles in traffic accident Unspecified bus occupant injured in collision with other and unspecified motor vehicles in traffic accident Rider or occupant injured in collision with two- or three-wheeled motor vehicle Rider or occupant injured in collision with car, pick-up truck, van, heavy transport vehicle or bus Rider or occupant injured in collision with other specified motor vehicle Occupant of railway train or railway vehicle injured in collision with motor vehicle in traffic accident Occupant of streetcar injured in collision with motor vehicle in traffic accident Driver of special industrial vehicle injured in traffic accident Passenger of special industrial vehicle injured in traffic accident Driver of special agricultural vehicle injured in traffic accident
152
Table E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E V V
Code Description V84.0 Driver of special agricultural vehicle injured in traffic accident V84.1 Passenger of special agricultural vehicle injured in traffic accident V84.2 Person on outside of special agricultural vehicle injured in traffic accident V84.3 Unspecified occupant of special agricultural vehicle injured in traffic accident V85.0 Driver of special construction vehicle injured in traffic accident V85.1 Passenger of special construction vehicle injured in traffic accident V85.2 Person on outside of special construction vehicle injured in traffic accident V85.3 Unspecified occupant of special construction vehicle injured in traffic accident V86.0 Driver of all-terrain or other off-road motor vehicle injured in traffic accident V86.1 Passenger of all-terrain or other off-road motor vehicle injured in traffic accident V86.2 Person on outside of all-terrain or other off-road motor vehicle injured in traffic accident V86.3 Unspecified occupant of all-terrain or other off-road motor vehicle injured in traffic accident V87.0 Person injured in collision between car and two- or three-wheeled motor vehicle (traffic) V87.1 Person injured in collision between other motor vehicle and two- or three-wheeled motor vehicle (traffic) V87.2 Person injured in collision between car and pick-up truck or van (traffic) V87.3 Person injured in collision between car and bus (traffic) V87.4 Person injured in collision between car and heavy transport vehicle (traffic) V87.5 Person injured in collision between heavy transport vehicle and bus (traffic) V87.6 Person injured in collision between railway train or railway vehicle and car (traffic) V87.7 Person injured in collision between other specified motor vehicles (traffic) V87.8 Person injured in other specified noncollision transport accidents involving motor vehicle (traffic) V89.2 Person injured in unspecified motor-vehicle accident, traffic V89.9 Person injured in unspecified vehicle accident X41 Accidental poisoning by and exposure to antiepileptic, sedative-hypnotic, antiparkinsonism and psychotropic drugs, not elsewhere classified X42 Accidental poisoning by and exposure to narcotics and psychodysleptics [hallucinogens], not elsewhere classified X42 Accidental poisoning by and exposure to narcotics and psychodysleptics [hallucinogens], not elsewhere classified X60-X84 Intentional self-harm Y11 Poisoning by and exposure to antiepileptic, sedative-hypnotic, antiparkinsonism and psychotropic drugs, not elsewhere classified, undetermined intent Y12 Poisoning by and exposure to narcotics and psychodysleptics [hallucinogens], not elsewhere classified, undetermined intent Y12 Poisoning by and exposure to narcotics and psychodysleptics [hallucinogens], not elsewhere classified, undetermined intent Y85.0 Sequelae of motor-vehicle accident Y85.9 Sequelae of other and unspecified transport accidents Y87.0 Sequelae of intentional self-harm Z20.6 Contact with and exposure to human immunodeficiency virus [HIV] Z21 Asymptomatic human immunodeficiency virus [HIV] infection status
Source: NZHIS, Collins and Lapsley (2008).
153
Appendix Table 23 Alcohol-attributable morbidity and mortality conditions The following list of ICD-10-AM 1 edition codes was used to assess alcohol-attributable morbidity and mortality. Data were supplied by the NZHIS on Deaths Registration data (Tables A and E) and Publicly Funded Hospital Discharges data (Tables A, B or V).
Table A A A A A A A A A A A A A A A A A A A A A A B, E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E Code C00-C14 C15 C22 C32 C50 F10.3-10.9 F10.0-10.2 G62.1 K70 I11, I13.0, I15 I20-25 I42.6 I47.1, I47.8-48.9 I50-51, I97.1 G45, I60-69 I85, I98.2 K22.6 K29.2 K74.3-74.6, K76.0, K76.9 K80 Q86.0 K85, K86.0-86.1 X45, Y15, T151.0-151.1, T51.9 W00-19, M80-82 X00-19 W65-74 W78-79 V01.0, 02.0, V03.0, V04.0, V05 V06.0, V09.0-09.1, V09.9, V10.0-10.3 V11.0-11.3, V12.0-12.3, V13.0-13.3 V14.0-14.3, V15.0-15.3, V16.0-16.3 V17.0-17.3, V18.0-18.3, V19.0-19.3 V20.0-20.3, V21.0-21.3, V22.0-22.3 V23.0-23.3, V24.0-24.3, V25.0-25.3 V26.0-26.3, V27.0-27.3, V28.0-28.3 V29.0-29.3, V30.0-30.4, V31.0-31.4 V32.0-32.4, V33.0-33.4, V34.0-34.4 V35.0-35.4, V36.0-36.4, V37.0-37.4 V38.0-38.4, V39.0-39.3, V40.0-40.4 V41.0-41.4, V42.0-42.4, V43.0-43.4 V44.0-44.4, V45.0-45.4, V46.0-46.4 V47.0-47.4, V48.0-48.4, V49.0-49.3 V50.0-50.4, V51.0-51.4, V52.0-52.4 V53.0-53.4, V54.0-54.4, V55.0-55.4 V56.0-56.4, V57.0-57.4, V58.0-58.4 V59.0-59.3, V60.0-60.4, V61.0-61.4 V62.0-62.4, V63.0-63.4, V64.0-64.4 V65.0-65.4, V66.0-66.4, V67.0-67.4 V68.0-68.4, V69.0-69.3, V70.0-70.4 V71.0-71.4, V72.0-72.4, V73.0-73.4 V74.0-74.4, V75.0-75.4, V76.0-76.4 V77.0-77.4, V78.0-78.4, V79.0-79.3 Description Oropharyngeal cancer Oesophageal cancer Liver cancer Laryngeal cancer Female breast cancer Alcoholic psychosis Alcohol dependence/abuse Alcoholic poly-neuropathy Alcoholic liver cirrhosis Hypertension Ischaemic heart disease Alcoholic cardiomyopathy Supraventricular cardiac dysrhythmias Heart failure Stroke - haemorrhagic/ ischaemic Oesophageal varicies Gastro-oesophageal haemorrhage Alcoholic gastritis Unspecified liver cirrhosis Cholelithiasis Fetal alcohol syndrome (dysmorphic) Pancreatitis - acute/chronic Alcoholic beverage & other EtOH poisoning Fall injuries Fire injuries Drowning Aspiration Occupational and machine injuries Occupational and machine injuries Occupational and machine injuries Occupational and machine injuries Occupational and machine injuries Occupational and machine injuries Occupational and machine injuries Occupational and machine injuries Occupational and machine injuries Occupational and machine injuries Occupational and machine injuries Occupational and machine injuries Occupational and machine injuries Occupational and machine injuries Occupational and machine injuries Occupational and machine injuries Occupational and machine injuries Occupational and machine injuries Occupational and machine injuries Occupational and machine injuries Occupational and machine injuries Occupational and machine injuries Occupational and machine injuries Occupational and machine injuries Occupational and machine injuries
st
154
Table E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E
Code V80.0-80.2, V80.6-80.8, V81.0, V81.2-81.9 V82.0, V83.4-83.9, V84.4-84.9, V85.4-85.9 V86.5-86.9, V87.9, V88, V89.0-89.1 X60-84, Y87.0 X85-Y09, Y87.1 V01.1-01.9, V02.1-02.9, V03.1-03.9 V04.1-04.9, V06.1-06.9 V09.2-09.3, V10.4-10.9, V11.4-11.9 V12.4-12.9, V13.4-13.9 V14.4-14.9, V15.4-15.9, V16.4-16.9 V17.4-17.9, V18.4-18.9 V19.4-19.9, V20.4-20.9, V21.4-21.9 V22.4-22.9, V23.4-23.9 V24.4-24.9, V25.4-25.9, V26.4-26.9 V27.4-27.9, V28.4-28.9 V29.4-29.9, V30.5-30.9, V31.5-31.9 V32.5-32.9, V33.5-33.9 V34.5-34.9, V35.5-35.9, V36.5-36.9 V37.5-37.9, V38.5-38.9 V39.4-39.9, V40.5-40.9, V41.5-41.9 V42.5-42.9, V43.5-43.9 V44.5-44.9, V45.5-45.9, V46.5-46.9 V47.5-47.9, V48.5-48.9 V49.4-49.9, V50.5-50.9, V51.5-51.9 V52.5-52.9, V53.5-53.9 V54.5-54.9, V55.5-55.9, V56.5-56.9 V57.5-57.9, V58.5-58.9 V59.4-59.9, V60.5-60.9, V61.5-61.9 V62.5-62.9, V63.5-63.9 V64.5-64.9, V65.5-65.9, V66.5-66.9 V67.5-67.9, V68.5-68.9 V69.4-69.9, V70.5-70.9, V71.5-71.9 V72.5-72.9, V73.5-73.9 V74.5-74.9, V75.5-75.9, V76.5-76.9 V77.5-77.9, V78.5-78.9 V79.4-79.9, V80.3-80.5, V80.9 V81.1, V82.1-82.9, V83.0-83.3 V84.0-84.3, V85.0-85.3, V86.0-86.4 V87.0-87.8, V89.2, V89.9, Y85
Description Occupational and machine injuries Occupational and machine injuries Occupational and machine injuries Suicide and self-abuse Child abuse and assault Road injuries Road injuries Road injuries Road injuries Road injuries Road injuries Road injuries Road injuries Road injuries Road injuries Road injuries Road injuries Road injuries Road injuries Road injuries Road injuries Road injuries Road injuries Road injuries Road injuries Road injuries Road injuries Road injuries Road injuries Road injuries Road injuries Road injuries Road injuries Road injuries Road injuries Road injuries Road injuries Road injuries Road injuries
Source: NZHIS, Collins and Lapsley (2008).
155
Appendix Table 24 Cost to business of lost output ($m), 2005/06

Cost to business of lost output ($m) Reduced value added Reduction in workforce Excess unemployment Absenteeism Reduced productivity Sub-total Business tax savings Reduction in workforce Excess unemployment Absenteeism Reduced productivity Sub-total Crime related lost output Reduced value added Business tax savings Sub-total Net cost to business of lost output
Source: BERL.
Alcohol 194.8 409.0 16.8 25.3 645.9 -64.3 -135.0 -5.5 -8.4 -213.2 90.2 -17.2 73.0 505.8
Other drugs 62.6 0.0 4.0 0.0 66.5 -20.6 0.0 -1.3 0.0 -22.0 32.8 -6.3 26.5 71.1
Joint AOD 0.0 106.0 5.6 5.9 117.5 0.0 -35.0 -1.8 -1.9 -38.8 30.8 -5.9 24.9 103.6
Total 257.4 515.0 26.3 31.3 829.9 -84.9 -169.9 -8.7 -10.3 -273.9 153.8 -29.4 124.4 680.5
Appendix Table 25 Cost to government of lost output ($m), 2005/06

Costs to government of reduced output ($m) Reduction in workforce (net) Individual tax Business tax Unemployment benefit savings Excess unemployment Individual tax Business tax Unemployment benefits incurred Absenteeism Individual tax Business tax Reduced productivity Business tax Crime related lost output Individual tax Business tax Net cost to government of lost output
Source: BERL.
Alcohol 42.6 64.3 -11.5 89.4 135.0 90.6 3.7 5.5 8.4 19.7 17.2 464.9
Other drugs 13.7 20.6 -3.5 0.0 0.0 0.0 0.9 1.3 0.0 7.2 6.3 46.4
Joint AOD 0.0 0.0 0.0 23.2 35.0 25.6 1.2 1.8 1.9 6.7 5.9 101.3
Total 56.3 84.9 -15.1 112.6 169.9 116.2 5.8 8.7 10.3 33.6 29.4 612.6
156
Appendix Table 26 Costs to government of harmful drug use ($m) detail, 2005/06
Tangible costs to government ($m) Alcohol Lost output Reduction in workforce (net) Excess unemployment Absenteeism Reduced productivity Crime related lost output Total net labour costs 95.4 315.0 9.2 8.4 37.0 464.9 Consumption-related taxes GST on reduced expenditure GST on untaxed expenditure Excise taxes and customs duties Total net diverted expenditure 29.4 0.0 -255.7 -226.4 Crime n.i.e. Customs Police Defence force Criminal courts Prisons Community sentences Total crime n.i.e. 0.1 172.2 0.0 64.5 112.3 18.7 367.8 Health care (net) Pharmaceuticals Hospital Medical Ambulances Treatment for victims of crime Total net health care 0.3 121.0 66.8 0.0 97.8 286.0 Road crashes n.i.e. Fire/emergency service Loss of output (disability) Total road crashes n.i.e. Total tangible costs to government Proportion of government costs Proportion of total tangible costs
Note: n.i.e. denotes not included elsew here. Source: BERL.
Other drugs Joint AOD
Total
30.8 0.0 2.2 0.0 13.4 46.4
0.0 83.7 3.1 1.9 12.6 101.3
126.1 398.7 14.4 10.3 63.0 612.6
7.7 64.8 0.0 72.6
0.0 0.0 0.0 0.0
37.1 64.8 -255.7 -153.8
23.8 90.7 0.6 34.0 87.7 18.2 254.8
0.0 43.5 0.0 16.3 95.9 13.6 169.3
23.9 306.3 0.6 114.8 295.8 50.5 791.9
11.8 9.2 1.0 0.0 9.0 31.0
0.0 0.0 -3.3 0.0 16.8 13.6
12.1 130.3 64.5 0.0 123.7 330.6
19.4 0.5 19.9 912.3 56.9% 20.0%
0.5 0.0 0.5 405.3 25.3% 8.9%
0.2 0.0 0.2 284.4 17.8% 6.2%
20.1 0.5 20.6 1,601.9 100.0% 35.1%
157
11
Appendix Sensitivity Analysis
The New Zealand AOD harm estimates in 2005/06 drew on a range of data and assumptions where necessary. Sensitivity analysis investigates how the drug harm results in 2005/06 are affected by critical parameters and assumptions. It gives a measure of confidence in how robust the results are to the changes inputs used to calculate them. The sensitivity analyses suggest that the estimates of harmful AOD use in 2005/06 are robust. A change of plus or minus one percent in a given factor alters the estimated total by less than half of a percent. On average, a one percent increase in the factors analysed leads to a 0.1 percent increase in estimated costs for positive changes and a -0.09 percent reduction in estimated costs for negative changes. The results are most sensitive to the assumptions about mortality rates, which affect costs in labour costs, health care, road crashes, lost of life and injury costs. A 10 percent increase in mortality rates leads to a 3.3 percent increase in total social costs. A 10 percent increase in the proportion of the supply of illegal drugs that is imported reduces the total social costs by 1.0 percent. Other sensitive assumptions include AOD use prevalence and the VOSLY. The sensitivity analyses examined the following key factors one at a time.
139
General: drug use prevalence, the quantity used and drug-attributable mortality rates Lost output: the value of time and the level of consumption savings Drug production: the cost of resources diverted for drug production Crime: offences and imprisonments attributed to drugs, and the cost per offence Health service: The amount of hospital care attributable to drug use Harmful drug use and harmful consumption decisions.
Both increases and decreases to the factors are analysed, typically with sensitivity factors of -10 percent, 10 percent and 25 percent applied to each factor.
140
The tables below show the
change from the main estimates (in percentage points). For example, if harmful alcohol and drug use prevalence were 10 percent lower than the rates used in this study, estimated total social costs would be 2.2 percent lower than the main study estimate of $6,525 million.
139
One-way sensitivity analysis examines the impact of varying one assumption at a time, holding all other factors constant.
140
We believe this range of sensitivity factors provide a reasonable indication of how the estimates will vary with modest, but sensible, changes to the assumptions. For example, this study used national drug prevalence rates based on survey data from SHORE. Dr Chris Wilkins from SHORE advises that the drug use prevalence rate for cannabis has been found to be between 10 to 30 percent higher in the Dunedin Longitudinal Study.
158
Appendix Table 27 Sensitivity analysis of general assumptions, 2005/06

Total social costs ($m) Sensitivity factor Tangible costs Intangible costs Total social costs Tangible costs ($m) Sensitivity factor Labour costs Drug production Crime n.i.e. Health care Road crashes n.i.e. Total tangible costs
Note: n.i.e. denotes not included elsewhere.
Harmful alcohol and drug use prevalence -10.0% -3.1% 0.0% -2.2% 10.0% 3.1% 0.0% 2.2% 25.0% 7.7% 0.0% 5.6%
Harmful alcohol and drug use prevalence -10.0% -6.8% 0.0% -0.6% -0.5% 0.0% -3.1% 10.0% 6.8% 0.0% 0.6% 0.5% 0.0% 3.1% 25.0% 17.1% 0.0% 1.4% 1.3% 0.0% 7.7%
Intangible costs ($m) Sensitivity factor Loss of life Lost quality of life Total intangible costs Avoidable costs ($m) Sensitivity factor Tangible costs Intangible costs Total avoidable costs Injury costs ($m) Sensitivity factor Tangible costs Intangible costs Total injury costs
Source: BERL.
Harmful alcohol and drug use prevalence -10.0% 0.0% 0.0% 0.0% 10.0% 0.0% 0.0% 0.0% 25.0% 0.0% 0.0% 0.0%
Harmful alcohol and drug use prevalence -10.0% -3.1% 0.0% -2.2% 10.0% 3.1% 0.0% 2.2% 25.0% 7.7% 0.0% 5.6%
Harmful alcohol and drug use prevalence -10.0% 0.0% 0.0% 0.0% 10.0% 0.0% 0.0% 0.0% 25.0% 0.0% 0.0% 0.0%
159
Appendix Table 28 Sensitivity analysis of lost output assumptions, 2005/06

Hourly wage by age, gender -10.0% -0.1% 0.0% -0.1% 10.0% 0.1% 0.0% 0.1% 25.0% 0.3% 0.0% 0.2%
Hourly wage by age, gender -10.0% -0.3% 0.0% 0.0% 0.0% 0.0% -0.1% 10.0% 0.3% 0.0% 0.0% 0.0% 0.0% 0.1% 25.0% 0.7% 0.0% 0.0% 0.0% 0.0% 0.3%
Source: BERL.
Hourly wage by age, gender -10.0% 0.0% 0.0% 0.0% 10.0% 0.0% 0.0% 0.0% 25.0% 0.0% 0.0% 0.0%
Hourly wage by age, gender -10.0% -0.1% 0.0% -0.1% 10.0% 0.1% 0.0% 0.1% 25.0% 0.3% 0.0% 0.2%
Hourly wage by age, gender -10.0% 0.2% 0.0% 0.1% 10.0% -0.1% 0.0% -0.1% 25.0% -0.3% 0.0% -0.1%
160
Consumption savings -10.0% 1.0% 0.0% 0.7% 10.0% -1.0% 0.0% -0.7% 25.0% -2.6% 0.0% -1.9%
Consumption savings -10.0% 2.0% 0.0% 0.0% 0.0% 5.2% 1.0% 10.0% -2.0% 0.0% 0.0% 0.0% -5.2% -1.0% 25.0% -5.0% 0.0% 0.0% 0.0% -13.0% -2.6%
Source: BERL.
Consumption savings -10.0% 0.0% 0.0% 0.0% 10.0% 0.0% 0.0% 0.0% 25.0% 0.0% 0.0% 0.0%
161
Appendix Table 29 Sensitivity analysis of drug production assumptions, 2005/06

Weekly expenditure on alcohol -10.0% -0.7% 0.0% -0.5% 10.0% 0.7% 0.0% 0.5% 25.0% 1.8% 0.0% 1.3%
Weekly expenditure on alcohol -10.0% 0.0% -3.0% 0.0% 0.0% 0.0% -0.7% 10.0% 0.0% 3.0% 0.0% 0.0% 0.0% 0.7% 25.0% 0.0% 7.4% 0.0% 0.0% 0.0% 1.8%
Source: BERL.
Weekly expenditure on alcohol -10.0% 0.0% 0.0% 0.0% 10.0% 0.0% 0.0% 0.0% 25.0% 0.0% 0.0% 0.0%
Weekly expenditure on alcohol -10.0% -0.7% 0.0% -0.5% 10.0% 0.7% 0.0% 0.5% 25.0% 1.8% 0.0% 1.3%
Weekly expenditure on alcohol -10.0% 0.0% 0.0% 0.0% 10.0% 0.0% 0.0% 0.0% 25.0% 0.0% 0.0% 0.0%
162
Street price of illegal drugs -10.0% -0.6% 0.0% -0.4% 10.0% 0.6% 0.0% 0.4% 25.0% 1.5% 0.0% 1.1%
Street price of illegal drugs -10.0% 0.0% -2.4% 0.0% 0.0% 0.0% -0.6% 10.0% 0.0% 2.4% 0.0% 0.0% 0.0% 0.6% 25.0% 0.0% 6.0% 0.0% 0.0% 0.0% 1.5%
Source: BERL.
Street price of illegal drugs -10.0% 0.0% 0.0% 0.0% 10.0% 0.0% 0.0% 0.0% 25.0% 0.0% 0.0% 0.0%
Street price of illegal drugs -10.0% -0.6% 0.0% -0.4% 10.0% 0.6% 0.0% 0.4% 25.0% 1.5% 0.0% 1.1%
Street price of illegal drugs -10.0% -10.6% 0.5% -4.0% 10.0% -10.6% 0.5% -4.0% 25.0% -10.6% 0.5% -4.0%
163
Resource cost of illegal drugs (% of street value) -10.0% -1.0% 0.0% -0.7% 10.0% 1.0% 0.0% 0.7% 25.0% 2.5% 0.0% 1.8%
Resource cost of illegal drugs (% of street value) -10.0% 0.0% -4.0% 0.0% 0.0% 0.0% -1.0% 10.0% 0.0% 4.0% 0.0% 0.0% 0.0% 1.0% 25.0% 0.0% 9.9% 0.0% 0.0% 0.0% 2.5%
Source: BERL.
Resource cost of illegal drugs (% of street value) -10.0% 0.0% 0.0% 0.0% 10.0% 0.0% 0.0% 0.0% 25.0% 0.0% 0.0% 0.0%
Resource cost of illegal drugs (% of street value) -10.0% -1.0% 0.0% -0.7% 10.0% 1.0% 0.0% 0.7% 25.0% 2.5% 0.0% 1.8%
Resource cost of illegal drugs (% of street value) -10.0% 0.0% 0.0% 0.0% 10.0% 0.0% 0.0% 0.0% 25.0% 0.0% 0.0% 0.0%
164
Imported illegal drugs (% of domestic supply) -10.0% 1.4% 0.0% 1.0% 10.0% -1.4% 0.0% -1.0% 25.0% -3.6% 0.0% -2.6%
Imported illegal drugs (% of domestic supply) -10.0% 0.0% 5.7% 0.0% 0.0% 0.0% 1.4% 10.0% 0.0% -5.7% 0.0% 0.0% 0.0% -1.4% 25.0% 0.0% -10.7% -4.1% 0.0% 0.0% -3.6%
Source: BERL.
Imported illegal drugs (% of domestic supply) -10.0% 0.0% 0.0% 0.0% 10.0% 0.0% 0.0% 0.0% 25.0% 0.0% 0.0% 0.0%
Imported illegal drugs (% of domestic supply) -10.0% 1.4% 0.0% 1.0% 10.0% -1.4% 0.0% -1.0% 25.0% -3.6% 0.0% -2.6%
Imported illegal drugs (% of domestic supply) -10.0% 0.0% 0.0% 0.0% 10.0% 0.0% 0.0% 0.0% 25.0% 0.0% 0.0% 0.0%
165
Dose per occasion of illegal drugs -10.0% -0.7% 0.0% -0.5% 10.0% 0.7% 0.0% 0.5% 25.0% 1.8% 0.0% 1.3%
Dose per occasion of illegal drugs -10.0% 0.0% -2.9% 0.0% 0.0% 0.0% -0.7% 10.0% 0.0% 2.9% 0.0% 0.0% 0.0% 0.7% 25.0% 0.0% 7.3% 0.0% 0.0% 0.0% 1.8%
Source: BERL.
Dose per occasion of illegal drugs -10.0% 0.0% 0.0% 0.0% 10.0% 0.0% 0.0% 0.0% 25.0% 0.0% 0.0% 0.0%
Dose per occasion of illegal drugs -10.0% -0.7% 0.0% -0.5% 10.0% 0.7% 0.0% 0.5% 25.0% 1.8% 0.0% 1.3%
Dose per occasion of illegal drugs -10.0% 0.0% 0.0% 0.0% 10.0% 0.0% 0.0% 0.0% 25.0% 0.0% 0.0% 0.0%
166
Appendix Table 30 Sensitivity analysis of crime assumptions, 2005/06

Alcohol-and-drug-attributable offences -10.0% -0.2% 0.0% -0.1% 10.0% 0.2% 0.0% 0.1% 25.0% 0.4% 0.0% 0.3%
Alcohol-and-drug-attributable offences -10.0% 0.0% 0.0% -0.8% 0.0% 0.0% -0.2% 10.0% 0.0% 0.0% 0.8% 0.0% 0.0% 0.2% 25.0% 0.0% 0.0% 1.9% 0.0% 0.0% 0.4%
Source: BERL.
Alcohol-and-drug-attributable offences -10.0% 0.0% 0.0% 0.0% 10.0% 0.0% 0.0% 0.0% 25.0% 0.0% 0.0% 0.0%
Alcohol-and-drug-attributable offences -10.0% -0.2% 0.0% -0.1% 10.0% 0.2% 0.0% 0.1% 25.0% 0.4% 0.0% 0.3%
Alcohol-and-drug-attributable offences -10.0% 0.0% 0.0% 0.0% 10.0% 0.0% 0.0% 0.0% 25.0% 0.0% 0.0% 0.0%
167
Alcohol-and-drug attributable imprisonments -10.0% -0.5% 0.0% -0.3% 10.0% 0.5% 0.0% 0.3% 25.0% 1.2% 0.0% 0.8%
Alcohol-and-drug attributable imprisonments -10.0% -0.4% 0.0% -1.3% 0.0% 0.0% -0.5% 10.0% 0.4% 0.0% 1.3% 0.0% 0.0% 0.5% 25.0% 1.0% 0.0% 3.3% 0.0% 0.0% 1.2%
Source: BERL.
Alcohol-and-drug attributable imprisonments -10.0% 0.0% 0.0% 0.0% 10.0% 0.0% 0.0% 0.0% 25.0% 0.0% 0.0% 0.0%
Alcohol-and-drug attributable imprisonments -10.0% -0.5% 0.0% -0.3% 10.0% 0.5% 0.0% 0.3% 25.0% 1.2% 0.0% 0.8%
Alcohol-and-drug attributable imprisonments -10.0% 0.0% 0.0% 0.0% 10.0% 0.0% 0.0% 0.0% 25.0% 0.0% 0.0% 0.0%
168
Harm per offence -10.0% -4.9% 0.2% -3.5% 10.0% 1.3% 0.2% 1.0% Harm per offence -10.0% -0.9% 0.0% -2.7% -3.3% 0.0% -1.3% 10.0% 0.9% 0.0% 2.7% 3.3% 0.0% 1.3% 25.0% 2.3% 0.0% 6.8% 8.1% 0.0% 3.2% 25.0% 3.2% 0.6% 2.5%
Source: BERL.
Harm per offence -10.0% 0.0% -9.0% -0.2% 10.0% 0.0% 9.0% 0.2% Harm per offence -10.0% -1.3% -0.2% -1.0% 10.0% 1.3% 0.2% 1.0% Harm per offence -10.0% 1.5% 0.3% 0.8% 10.0% 1.7% 0.0% 0.6% 25.0% 4.3% 0.0% 1.6% 25.0% 3.2% 0.6% 2.5% 25.0% 0.0% 22.6% 0.6%
169
Appendix Table 31 Sensitivity analysis of health assumptions, 2005/06

Hospital cost weights -10.0% -0.3% 0.0% -0.2% 10.0% 0.3% 0.0% 0.2% Hospital cost weights -10.0% 0.0% 0.0% 0.0% -3.6% 0.0% -0.3% 10.0% 0.0% 0.0% 0.0% 3.6% 0.0% 0.3% 25.0% 0.0% 0.0% 0.0% 9.0% 0.0% 0.7% 25.0% 0.7% 0.0% 0.5%
Source: BERL.
Hospital cost weights -10.0% 0.0% 0.0% 0.0% 10.0% 0.0% 0.0% 0.0% Hospital cost weights -10.0% -0.3% 0.0% -0.2% 10.0% 0.3% 0.0% 0.2% Hospital cost weights -10.0% -1.4% 0.0% -0.6% 10.0% 1.4% 0.0% 0.6% 25.0% 3.5% 0.0% 1.4% 25.0% 0.7% 0.0% 0.5% 25.0% 0.0% 0.0% 0.0%
170
Appendix Table 32 Sensitivity analysis of intangible cost assumptions, 2005/06

Value of statistical life year -10.0% 0.0% -10.0% -2.7% 10.0% 0.0% 10.0% 2.7% 25.0% 0.0% 25.0% 6.8%
Value of statistical life year -10.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 10.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 25.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0%
Source: BERL.
Value of statistical life year -10.0% -10.0% -10.0% -10.0% 10.0% 10.0% 10.0% 10.0% 25.0% 25.0% 25.0% 25.0%
171
11.1
Harmful drug use and consumption decisions
When measuring the social cost of harmful AOD use, known private costs should be excluded (Single et al 2003, Collins and Lapsley 2008). This is because private costs are assumed to be offset by private benefits, so there is no net social cost. For example, a fully informed, consistently rational consumer buying alcohol is assumed to know the negative impacts of drinking, and weighs these against the benefits in making their decision. They will choose to consume up to the point where their benefits match the full, privately borne cost. In the case of harmful drug use, although the consumer is not necessarily behaving irrationally, individual decisions are not necessarily made on a fully informed, rational and consistent basis.
141
Such decisions impose a cost on society. For example, we estimate
that a significant amount of alcohol is consumed by alcohol affected, intoxicated or alcoholdependant individuals. Where these decisions are not fully informed, consistently rational decisions, a consumers private decision may result in social costs, even if this impact is borne by the individual. We follow the standard approach in the cost of substance abuse literature in our estimations. Costs due to decisions that are not fully informed, consistently rational and fully borne by the individual are treated as social costs of harmful use. As Collins and Lapsley note, if the objective were to estimate the total costs of abuse, both private costs/benefits and social costs/benefits should be incorporated into the estimates. As a social cost study; we do not consider private costs of non-harmful private use, nor do we assign private benefits. Were harmful alcohol use to cease, the consequent reduction in consumption would release resources which could be used for other consumption or investment uses (Collins and Lapsley 2008). As such, the 50 percent of harmful alcohol consumption estimated in this study results in a social cost. Similarly, all illegal drug use is assumed to be harmful. While this may overstate the case for social illegal drug users (particularly with regard to assigning their use intangible benefits), frequent drug users are estimated to consume 80-100 percent of most types of illegal drugs, with the exception of LSD where just over half is estimated to be used by frequent users (BERL 2008). Furthermore, in the absence of evidence to support an alternative assumption, we follow the approach set by Collins and Lapsley and treat all illegal drug use as harmful for the estimation of this component.
141
For example, Markandya and Pearce (1989) consider the types of costs that may be constituent parts of the social costs of harmful drug use. They interpret costs, including the cost of production, as social costs even if that individual is rational and they bear that cost privately, where those costs have not been knowingly incurred.
172
Appendix Table 33 Sensitivity analysis of harmful AOD use assumptions, 2005/06

Total social costs ($m) Sensitivity factor Tangible costs Intangible costs Total social costs Tangible costs ($m) Sensitivity factor Labour costs Drug production Crime n.i.e. Health care Road crashes n.i.e. Total tangible costs Harmful alcohol use diverted inputs -25% -3.6% 0.0% -2.6% -10% -2.5% 0.0% -1.8% 10% -1.1% 0.0% -0.7%
Harmful alcohol use diverted inputs -25% 0.0% -30.2% 17.7% 0.0% 0.0% -3.6% -10% 0.0% -25.6% 17.7% 0.0% 0.0% -2.5% 10% 0.0% -19.5% 17.7% 0.0% 0.0% -1.1%
Harmful drug use diverted inputs -10% -1.0% 0.0% -0.7% -25% -2.5% 0.0% -1.8% -50% -4.9% 0.0% -3.6%
Harmful drug use diverted inputs -10% 0.0% -4.0% 0.0% 0.0% 0.0% -1.0% -25% 0.0% -9.9% 0.0% 0.0% 0.0% -2.5% -50% 0.0% -10.5% -10.4% 0.0% 0.0% -4.9%
173
12
Appendix Alternative GP cost estimates
The main GP cost estimates were based on the ratio of people admitted to hospitals with AOD-related conditions, following Collins and Lapsleys (2008) method and advice. An alternative source of data on drug-related GP use is the General Practice Computer Databases. These databases evolved from the Royal New Zealand College of General Practitioners (RNZCGP) Computer Research Network. The Network consists of general practices from throughout urban and rural New Zealand. Data is collected from computerised records of consultation notes, prescriptions, investigations and referral forms. Data from 30 practices between 2005 and 2007 were examined for patients aged 10 years plus treated for alcohol and drug abuse. These patients were identified from GP Read codes, which are a comprehensive clinical coding system, as having an AOD abuse problem. The search results underestimate the number of such patients at these practices having such problems since Read coding only occurs for around 20 percent of consultations. Just over 0.65 percent of 142,353 patients were identified as having AOD problems. Females had an average of 8.4 additional GP visits per annum above the population rate, while males had an average of 5.4 additional GP visits per annum. To get an indication of the cost of AOD-related primary care based on these figures we multiply together: the percentage of people with AOD-related problems visiting a GP age-gender excess GP consultation rates national population age-gender profiles (aged 10 years plus) cost per consult (for those less than 18 years old and for those 18 years or older)
After allowing for savings from premature mortality, this calculation suggests a figure of $145.1 million. This dollar figure is approximately one fifth of the net medical (primary care) services figure estimated in Collins and Lapsley (2008). Collins and Lapsleys estimate suggests that just under five percent of national spending on medical services was due to AODs. Our figure above corresponds to approximately 12.5 percent of New Zealands expenditure on general practitioners. Therefore, the figure is broadly in line with New Zealands smaller population and differing drug use prevalence. However, the figure calculated above is based on a sample of less than 500 people with identified AOD problems. Therefore, the estimate should be regarded as indicative, and it is not used in the main report estimates.
174
All work is done, and services rendered at the request of, and for the purposes of the client only. Neither BERL nor any of its employees accepts any responsibility on any grounds whatsoever, including negligence, to any other person. While every effort is made by BERL to ensure that the information, opinions and forecasts provided to the client are accurate and reliable, BERL shall not be liable for any adverse consequences of the clients decisions made in reliance of any report provided by BERL, nor shall BERL be held to have given or implied any warranty as to whether any report provided by BERL will assist in the performance of the clients functions.
175

BERL - July 2009 - Costs of Harmful Alcohol and Other Drug Use-1

Transféré par

Informations du document

Titre original

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

BERL - July 2009 - Costs of Harmful Alcohol and Other Drug Use-1

Transféré par

Droits d'auteur :

Formats disponibles

Report to:

Ministry of Health and ACC

COSTS OF HARMFUL ALCOHOL AND OTHER DRUG USE

Copyright BERL BERL ref #4577

Costs of Harmful Alcohol and Other Drug Use

3 Literature Review ........................................................................... 14

4 Method and Calculation Summary ............................................... 29

5 Results: the Costs of Harmful AOD Use ...................................... 56

6 Additional Analytical Focuses ...................................................... 67

12 Appendix Alternative GP cost estimates ................................ 174

Alcohol 2,875.1 1,561.9 4,437.1 68.0%

Total 4,561.5 1,963.1 6,524.6 100.0%

Drug prod'n 861.0 19%

Lost output 1951.6 43%

Alcohol 1,440 780 2,220 70% 34%

Total 2,280 980 3,260 100% 50%

Summary Table 3 Injury costs of harmful drug use ($m), 2005/06

Alcohol 624.4 967.5 1,591.9 81.6% 24.4%

Total 729.1 1,222.4 1,951.4 100.0% 29.9%

So, while alcohol use may be benign or harmful in some cases, it is

However, there are intangible benefits,

Table 2.1 below reports the

Grams of alcohol per day

Women 0.0 0-19.99 20-39.99 40+

Men 0.0 0-39.99 40-59.99 60+

In the case of alcohol production, this analysis would be complicated by

Section 4 briefly sets out the broad methods

The summary is based on a separate literature review report, BERL (2008)

Cost categories Tangible costs Cost components Intangible costs

Begg et als introduction to joint risk attribution makes

As noted in the authors earlier

Method and Calculation Summary

Box 1 Harmful AOD use

Source: BERL, StatsNZ

Low risk Hazardous Australia New Zealand

High risk 5.6% 8.7%

Any use 63.2% 87.1%

15-45 y.o. 17.9% 0.2%-0.4% 1.1% 3.4% 0.8% 3.9% 1.8%

Source: BERL, based on SHORE data, Connor et al (2005)

This included estimating the proportion of people

This research indicated that 4.2 percent

This process is described in section 4.3.8.

The cost of collecting alcohol

Source: BERL, based on NZP data

Source: BERL, based on HOI data and NZP advice

For example, just over one

Source: BERL, based on HOI data and NZP advice

We apply this profile to the estimated additional population in

Data on illegal drugs was determined from records on

However, the studys avoidable cost estimate uses a

These methods are detailed in section 3.3.4.

Results: the Costs of Harmful AOD Use

The tangible costs were equivalent to two

Alcohol 2,875.1 1,561.9 4,437.1 68.0%

Total 4,561.5 1,963.1 6,524.6 100.0%

Note: n.i.e. denotes not included elsew here. Source: BERL

Alcohol 1,519.6 42.3 1,561.9 79.6%

Other drugs 386.5 6.0 392.4 20.0%

Joint AOD 0.4 8.3 8.7 0.4%

Total 1,906.5 56.6 1,963.1 100.0%

The latter estimate is based on the ACC data for people

Costs of harmful alcohol use

Alcohol 3,369.8 1,569.5 4,939.3 75.7%