Vous êtes sur la page 1sur 101

Principles of Pharmacoeconomics

 Assessing the clinical effectiveness of any new health care intervention,

including medications.
 Determining the role of the new intervention in clinical practice. Sometimes the
new interventions may provide only a modest advantage (or no advantage) over
existing treatment, usually at a higher cost.
 In the case of pharmaceutical interventions, pharmacoeconomics attempts to
measure if the added benefit of one intervention is worth the added cost of that
 Pharmacoeconomics has been defined as the description and analysis of the
costs of drug therapy to health care systems and society. It identifies,
measures, and compares the costs and consequences of pharmaceutical
products and services.
 Clinicians and other decision makers can use these methods to evaluate and
compare the total costs of treatment options and the outcomes associated with
these options.

Figure 1. Basic pharmacoeconomic equation. Pharmacoeconomic studies compare the

costs (left box) associated with providing a pharmacy product or service (represented as
Rx) to the outcome of the product or service.

The left-hand side of the equation represents the inputs (costs) used to obtain and
use the pharmaceutical product or service. The right-hand side of the equation
represents the health-related outcomes produced by the pharmaceutical product or
service. The center of the equation, the drug product or service being assessed, is
symbolized by Rx.
 If just the left-hand side of the equation is measured without regard for outcomes,
it is a cost analysis (or a partial economic analysis).
 If just the right-hand side of the equation is measured without regard to costs, it
is a clinical or outcome study (not an economic analysis).
To be a true pharmacoeconomic analysis, both sides of the equation must be
considered and compared. Theoretically, at least two options must be compared in
pharmacoeconomics, but some assessments consist of a "with or without"
comparison, estimating what would occur if the product or service was provided
(e.g., immunization or pharmacy clinic services) compared with no provision of the
product or service.

Pharmacoeconomics may help to know:

1-What drugs should be included on the hospital formulary?
2-What is the best drug for a particular patient?
3-What is the best drug for this particular disease?
4-What is the best drug for a pharmaceutical manufacturer to develop?
5- How do two clinical pharmacy services compare?
6-Will patient quality of life is improved by a particular drug therapy decision?
7-What are the patient outcomes of various treatment modalities?
In essence, pharmacoeconomic analysis uses important tools for examining the
outcomes or impact of drug therapy and related healthcare interventions.

Relationship of Pharmacoeconomics to Other Research

 Unlike in other scientific fields, there is no standardized training for pharma-

coeconomists, and it is a multidisciplinary field. The specific field of
pharmacoeconomics is relatively new—the term first appeared in the literature in
the mid-1980s—yet the concepts and methods are borrowed from other, more
established disciplines and research areas.

 Pharmacoeconomics overlaps with both health care economics and

pharmacy-related clinical or humanistic outcomes research, as illustrated in
Figure 2.
This figure illustrates that pharmacoeconomics incorporates methodologies
from outcomes research and the discipline of health care economics.
 Health care economics encompasses a broad range of topics, including supply
and demand for health care resources, the effects of health insurance, and
manpower supply.
 Clinical or humanistic outcomes research is defined as the attempt to identify,
measure, and evaluate the end results of health care services.
 It may include not only clinical and economic consequences but also outcomes
such as patients' health status and satisfaction with their health care.
 Pharmacoeconomics is a type of outcomes research, but not all outcomes
research is pharmacoeconomic research.
 If the research involves economic and clinical outcome evaluations and
comparisons of pharmacy products or services, it can be termed a
pharmacoeconomic study (overlapped area of Figure 2).

 Similar to costs, the outcomes or consequences of a disease and its treatment are
an equally important component of pharmacoeconomic analyses.
 The manner in which consequences are quantified is a key distinction among
pharmacoeconomic methods because the assessment of costs is relatively
Depending on perspective, the outcomes of health care are multidimensional. The
clinician has traditionally been most concerned with clinical outcomes of
treatments. More recently, healthcare payers and administrators have focused on
the resource use or economic outcome of healthcare decisions. Patients, on the other
hand, are becoming increasingly knowledgeable and involved in decisions regarding
their own health care and are seeking more information regarding the humanistic
outcomes of therapy.
Patients want to know how their quality of life will be affected or how satisfied other
patients with their condition have been with various treatments.
Accordingly, the consequences (or outcomes) of medical care also can be
categorized. One approach is to separate outcomes into three categories: economic,
clinical, and humanistic.
 Economic outcomes are the direct, indirect, and intangible costs compared with
the consequences of alternative medical treatment.
 Clinical outcomes are the medical events that occur as a result of disease or
treatment (e.g., safety and efficacy end points).
 Humanistic outcomes are the consequences of disease or treatment on patient
functional status or quality of life along several dimensions (e.g., physical
function, social function, general health and well-being, and life satisfaction).
Assessing the economic, clinical, and humanistic outcomes (ECHO) associated with
a treatment alternative provides a complete model for decision making.

 Clearly, cost-containment is an important objective.

 However, successful healthcare management as measured by the objectives of
patients, physicians, and other healthcare providers, as well as by societal
expectations, requires that the quality of care also be maintained.
 Outcomes measurement must take into account economic considerations while
recognizing that acceptable clinical and humanistic outcomes are also important

 The true value of healthcare interventions, programs, and policy can be assessed
only if all three dimensions of outcomes are measured and considered.

Positive Vs negative consequences

 These consequences (outcomes) can be further categorized as positive or
 An example of a positive outcome is a desired effect of a drug (efficacy or
effectiveness measure), possibly manifested as cases cured, life-years gained
 Since all drugs have adverse effects, negative consequences also can occur with
their use. A negative outcome is an undesired or adverse effect of a drug, possibly
manifested as a treatment failure, an adverse drug reaction (ADR), a drug
toxicity, or even death.
 Pharmacoeconomic evaluations should include assessments of both types of
outcomes. Evaluating only positive outcomes may be misleading because of the
potential detriment and expense associated with negative outcomes. Thus the
balancing of positive and negative consequences is important in any
pharmacoeconomic evaluation.

Assessment of costs (in healthcare)

Identifying costs:
 It is important to take into account all the costs associated with an intervention,
not just the market prices. Calculation of true economic cost is difficult, but it is
essential to make sure that cost information reflects true economic cost as closely
as possible.

 This is not usually straightforward in healthcare because normal markets and

pricing mechanisms are not necessarily operating. For example, prescribing the
highly effective antipsychotic drug clozapine for a person with schizophrenia
does not just incur the costs of buying the drug. Clozapine can have serious side
effects, and so regular blood monitoring tests have to be carried out in all patients.
Therefore, these monitoring costs must be taken into account when the economic
implications of using clozapine are being assessed. A very small number of
patients go on to experience serious side effects that require hospitalization and
treatment, and these costs must also be identified and measured.

Types of costs
Economic studies use a range of costs. It is important to be able to distinguish
between the different types of costs that are used. Costs are usually divided into
direct, indirect and intangible (Figure 3) and table 1.

Figure 3 types of costs

Type of Cost Category

Direct medical costs
• Medications
• Medication monitoring
• Medication administration
• Patient counseling and consultations
• Diagnostic tests
• Hospitalizations
• Ambulance services
• Nursing services
Direct nonmedical costs
• Travel costs to receive health care (bus, taxi)
• Hotel stays for patient or family for out-of-town care
• Child care services for children of patients
Indirect costs
• Lost productivity for patient
• Lost productivity for unpaid caregiver (e.g., family member, neighbour, friend)
• Lost productivity because of premature mortality
Intangible costs
• Pain and suffering • Fatigue • Anxiety

Direct Medical Costs

Direct medical costs are the most obvious costs to measure. These are the medically
related inputs used directly to provide the treatment. include the costs associated
with the pharmaceuticals, diagnostic tests, physician visits, pharmacist visits,
emergency department visits, and hospitalizations.
Example: during chemotherapy treatment, direct medical costs may include the
chemotherapy products themselves, other medications given to reduce side effects
of the chemotherapy, intravenous supplies, laboratory tests, clinic costs, and
physician visits.

Direct Nonmedical Costs

Direct nonmedical costs are costs to patients and their families that are directly
associated with treatment but are not medical in nature. These include the cost
of traveling to and from the physician's office, clinic, or the hospital; child care
services for the children of a patient; and food and lodging required for the patients
and their families during out-of-town treatment.
Using the example of chemotherapy treatment, patients may have increased travel
costs related to traveling to the clinic or hospital. They may also have to hire a
babysitter for the time they are undergoing treatment.

Indirect costs
Indirect costs are incurred by the reduced productivity of a patient and their family,
resulting from illness, death or treatment. They may include time off work or; time
spent going to healthcare providers; time spent caring for the patient by relatives;
time forgone from leisure. Only the following indirect costs can be calculated
reliably from data:
A-Time off work due to sick leave
B-Early retirement
C-Reduced productivity at work.

The significance of indirect costs depends upon the particular illness and treatments
involved. Diseases such as asthma, migraine and depression affect working age
groups, whereas other diseases, such as Alzheimer's, do not.
Indirect costs are difficult to measure, so that they are not often included in economic
studies. However, it is likely that most interventions will affect indirect costs, so they
should always be considered, if not measured.

Intangible costs
Intangible costs are difficult or impossible to measure, but they still occur and it is
of value to identify them. They can include anxiety, pain or suffering from an
illness or treatment.

Incremental costs and marginal costs

The terms incremental cost and marginal cost are often used in economic
evaluations. An incremental cost is the difference in overall costs between two
alternatives. A marginal cost is the cost of carrying out one more intervention, or
expanding a programme or service (e.g. increasing the length of stay in hospital by
one day).
Opportunity costs
Represent the economic benefit forgone when using one therapy instead of the next
best alternative therapy. Therefore, if a resource has been used to purchase a program
or treatment alternative, then the opportunity to use it for another purpose is lost. In
other words, opportunity cost is the value of the alternative that was forgone.

Imagine we have a choice of two effective treatments, A and B, but only enough
money for one of them. If treatment A is funded rather than treatment B, the
opportunity cost benefits of funding A is in not choosing B.
Consider, for example, two possible interventions: a cancer screening program
(intervention A) and a smoking cessation program (intervention B). If only one of
these interventions can be funded then the opportunity cost of funding A can be
thought of as the number of life years that would have been gained is more than
through the smoking cessation program.

Average cost
Average cost is calculated by dividing the total costs for the intervention by the total
quantity of treatment units provided, such as the number of patients receiving a
course of antibiotics.

How are costs valued?

The two ways of collecting costs are either 'top down' or 'bottom up' (also called
 Top-down studies use the total budget to produce average costs per patient. This
method is the quicker one, but assumes that all patients have the same diagnosis,
severity of illness and treatment.
 Bottom-up studies measure resource use by individual patients and so are able to
detect treatment differences between patients. This method produces much better
quality costs, but can be time-consuming and expensive.

When to stop collecting costs

It is important to carry on collecting information about resource use until it is clear
that no more events are likely to occur related to the original intervention.
These events may be further interventions, such as treatment, or they can be side
effects of the intervention, and consequences of failed treatment.
 The study must carry on collecting cost information until all these events have
stopped. This may be a month, a year, or longer, depending on the type of
intervention or illness being treated.
 The time when costs are no longer collected should coincide with when outcomes
are also no longer going to be collected and is called the time horizon.
Often, trials do not continue with such a long-term follow-up period because of
expense, and a shorter time horizon will be employed.

Timing Adjustments for Costs

Bringing Past Costs to the Present: Standardization of Costs
When costs are estimated from information collected for more than 1 year before the
study, adjustment of costs is needed; this is also referred to as standardization of
costs. If retrospective data are used to assess resources used over a number of years
back, these costs should be adjusted.
If you compared costs for patients who received treatment in 2000 with patients who
received treatment in 2005, the comparison of resources used would not be a fair
comparison because treatment costs tend to go up each year, so patients who
received the same treatment in 2000 would have lower costs than those who received
the treatment in 2005. Adjustment of the 2000 costs to the amount they would have
cost in 2005 is needed before a direct (fair) comparison can be made between these
For example, if the objective of the study is to estimate the difference in the costs of
chemotherapy regimens, information on the past use of these two treatments might
be collected from a review of medical records. If the retrospective review of these
medical records dates back for more than 1 year, it may be necessary to standardize
the cost of both medications by calculating the number of units (doses) used per case
and multiplying this number by the current unit cost for each medication due to
frequent use).
Table 2.2 illustrates an example of adjustment using this first method to estimate the
treatment costs for a mild infection.
Another method used to standardize past costs is to multiply all of the costs from the
year the data were collected by the medical inflation rate for that year (Table 2.3).
Medical Consumer Price Index (CPI) inflation rates can be found at the Bureau of
Labor Statistics' website (www.bls.gov) and have been between 4% - 5% each year
since 2000.

Bringing Future Costs (Benefits) to the Present: Discounting

If costs are estimated based on dollars spent or saved in future years, another type of
modification, called discounting, is needed. There is a time-value associated with
money. People (and businesses) prefer to receive money today rather than at a later
time. Therefore money received today is worth more than the same amount of money
received next year.
Modifications for this time value are estimated using a discount rate. From this
parameter, the present value (PV) of future expenditures and savings can be
calculated. The discount rate generally accepted for health care interventions is
between 3% and 6%.
The discount factor is equal to 1/ (1 + r) , where r is the discount rate and t is the
number of years in the future that the cost or savings occur.
For example, if the expenses of cancer treatment for the next 3 years are $5,000 for
the first year, $3,000 for year 2, and $4,000 for year 3, discounting should be used
to determine total expenses in present value (PV) terms.
If one assumes that the expenses occur at the beginning of each year, then first year
costs are not discounted (see Table 2.4).

It is equally acceptable to assume that expenses occur at the end of the first year (12
months later) and therefore, they are discounted (see Table 2.5).
This example illustrates the impact of discounting the costs of an intervention using
information from a theoretical hip replacement operation compared to drug
treatment for a patient with osteoarthritis.

Costs (£) arising during
Year O Year 1 Year 2 Year 3 Year 4
Hip replacement 5,000
Painkillers (undiscounted) 1,000 1,000 1,000 1,000 1,000

1-What are the total costs for a hip replacement?

The total cost for the hip replacement operation is £5,000.
2-What are the total costs for drug treatment?
The total cost for the drug treatment is £5,000.
3-What is the impact of discounting the costs?
The costs for the hip replacement all occur in the first year (now). This means that
they should not be discounted.

The costs for drug treatment occur over a 5-year period. This means that they must
be adjusted for discounting.
Using the above formula and discount rate of 3.5% we can calculate the discount
factor for each year.
The discounted costs are calculated by multiplying the discount factor by the costs
for each year.

This example shows that when discounting is taken into account, drug treatment over
5 years period is cheaper than hip replacement.
Therefore, future costs have to be discounted in any economic evaluation. Studies
will generally discount costs if the intervention lasts more than 1 year.

Cost of illness
The cost of illness (COI) is the personal cost of acute or chronic disease. The cost to
the patient may be economic, social or psychological………etc.
It differs from healthcare costs, meaning the societal cost of providing services
related to the delivery of healthcare, rather than personal impact on individuals.
The direct medical costs of illness include diagnosis, medical treatment, surgery and
follow-up care. Indirect costs of illness are costs attributable to loss of productivity
of patients with that disease or condition.
Intangible costs of illness are the pain and suffering associated with illness.
COI studies are used to indicate the magnitude of resources needed for a specific
disease or condition, and they may be used to compare the economic impact of one
disease versus another (e.g., costs of schizophrenia versus costs of asthma) or the
economic impact of a disease on one country compared with another (e.g., costs of
D.M in Iraq versus costs of D.M in Jordan).
The point of treating a patient with a disease is to reduce these costs of illness.
Measuring patient outcomes for use in economic evaluations
 An economic evaluation looks at all the implications of deciding to choose one
way of providing care over another, not just the costs.
 This means that any factor affect the service, good or bad, has on the patient or
customer needs to be investigated.

Benefits, outcomes and consequences refer to the effect on the patient, not the
effect on people providing the service. Cost is not an outcome measure.
 In order to assess the benefit of healthcare, correct identification and
measurement of patient outcomes is essential in health economics. An
understanding of these outcomes is a prerequisite for economic evaluation.
 This lecture defines and describes the main categories of outcome measure used:
effectiveness, quality of life, utility, and expressing benefits as monetary
values, and appraises the importance of using the appropriate outcome measure.

Effectiveness is the outcome of an intervention or service measured in natural units.
These can be general outcome measures, such as:
 Cases successfully diagnosed.
 Cases successfully treated.
 Life years saved.

It is also possible to use clinical indicators, such as:

 Number of asthma attacks avoided
 Pain-free days
 Change in infection rate
 Percentage reduction in blood pressure
 Effect on nausea and vomiting frequency.

These measures are relatively simple to use and are often reported in clinical trials
of interventions. They are therefore the most common type of outcome measure
and the most frequently used in economic evaluation.
They are sometimes called intermediate outcome measures because the
implication that changes in them will extrapolate to an effect on the patient's ultimate
health status. For example, a study of diet-based lipid-lowering therapy in the
prevention of coronary heart disease could use the drop in plasma cholesterol level
to assess the effectiveness of this intervention. It is assumed that by reducing a
person's cholesterol level, the risk of developing coronary heart disease is reduced.
We know that this assumption is true because there is epidemiological evidence to
support the link between cholesterol levels and coronary heart disease.

Example of effectiveness
Table 4.1 reports the results from a trial that compared daily recombinant human
deoxyribonuclease (rhDNase), alternate-day rhDNase and hypertonic saline for
treating children with cystic fibrosis. The primary outcome measure was forced
expiratory volume in one second (FEV-1, a measure of lung function).
The results show that rhDNase was more effective at improving lung function than
hypertonic saline, but there was no difference between the two different dosage
schedules for rhDNase.

Effectiveness versus efficacy

Efficacy is the consequence (benefit) of a treatment under ideal and controlled
conditions (ideal trial study). However, real life does not behave like an ideal trial
study. In practice, different types of patients from those in the trial may receive the
Effectiveness is the therapeutic consequence of a treatment in real-world
conditions. The effectiveness of a treatment or service is often lower than its
efficacy, and so using information from an ideal clinical trial may overestimate the
impact of the intervention. Many researchers are attempting to overcome this
problem by designing clinical trials that reflect practice more closely.

Limitations of effectiveness measures

A problem with effectiveness measures is that:

1-There may be more than one outcome reported for a particular treatment or
service. For example, an analgesic may be more effective than another at relieving
pain, bur cause more nausea.
Does one choose the analgesic with more power, or the one with a lower incidence
of nausea? So, effectiveness measures are limited because they only measure one
part of an outcome and may not reflect the overall impact of the intervention
on the patient's health-related quality of life (HRQoL).
For example, in the cystic fibrosis study reported above, we know that FEV1 was
improved by the interventions. However, we do not know whether that improvement
was large enough to improve the patient's HRQoL, so it is an intermediate
measure of out-come, where we are assuming that improvements in lung function
can improve overall health status. Also, we do not know whether the treatments were
associated with unwanted side effects. There may be side effects that are
sufficiently bad to affect the patient's health status, or so bad that they refuse to have
the treatment at all. Finally, we do not know whether the improvement in lung
function now will continue into the future and extend the life expectancy of the

2- Many effectiveness measures are disease specific, such as lung function, and so
cannot be used to compare outcomes for different disease states. For example, it
would not be possible to use lung function to assess the outcome of an operation for
a ruptured Achilles tendon.

Mortality used as an effectiveness measure

Mortality has been used to measure the effectiveness of treatments in patients.
Examples of this are the studies looking at the use of aspirin after myocardial
infarction, lipid-lowering agents in coronary heart disease, and the treatment of
hypertension in patients with diabetes. Mortality is a useful outcome measure
because it is objective and easy to measure. However, there are problems associated
with it:

1- People may die from causes other than that of interest to the study, which can
mask mortality linked to the intervention of interest and thus confound the results.
2- Most illnesses affect quality of life rather than mortality, and so quality of life
improvements due to interventions will not be detected or included in the economic
3- Third, mortality requires a study with many patients and need follow up over
a long period of time.
4- People of different ages and sex have different risks of mortality, so it is
important that patient groups have similar age and sex profiles if they are to be

Example of mortality used as an effectiveness measures

Table 4.2 reports the results from a trial that examined the mortality of men aged 45-
54 with pre-existing coronary heart disease, and assessed the impact of lipid-
lowering drugs on that mortality. This table shows that using these drugs reduces
mortality in this patient group. If 1000 men were treated for 10 years, three deaths
would be prevented.

The limitations of the use of mortality can be seen in this example. First, it takes up
to 10 years' follow-up in a very large group of patients to detect a change in
mortality. Also, people may die from another cause, which can confound the
It is likely that treating patients with these drugs reduced the progression of many of
them to coronary heart disease. Therefore, the health status of many patients was
probably improved by the intervention, such that they could carry on working or
their normal daily activities. This important effect of the treatment is not picked
up by measuring mortality only.
Everybody dies eventually, so it is actually premature mortality we are trying to
reduce, such that individuals are able to live out an acceptable lifespan. Therefore,
if a premature death is prevented — for example by lipid-lowering drugs - we have
saved years of life for an individual. Therefore, mortality can be converted into life-
years saved or life-years gained.

Worked example A theoretical worked example of using effectiveness

Patients with chronic renal failure who are on haemodialysis suffer from profound
anaemia, which is often extremely debilitating. This is due to a reduction in their
production of erythropoietin and loss of blood during haemodialysis. Historically,
these patients have been managed by the use of blood transfusions. Now, synthetic
erythropoietin is available. It is considered to be highly effective, but is very
expensive. So, the alternatives are either to give erythropoietin or to give blood
transfusions when the patient's haemoglobin level is below 8g/dl.
Effectiveness data for the two alternatives available from the literature suggest that
erythropoietin can maintain haemoglobin levels above 8g/dl for 91 % of the year,
whereas blood transfusions maintain levels for 76% of the year.

1. What is the implicit assumption being made by the use of this outcome
The assumption is that this is a desirable outcome because the reversal of anaemia
will increase the patient's energy levels and hence their quality of life.
2-What is the difference in effectiveness of the two alternatives, for 1000
Erythropoietin keeps the Hb level >8g/dl for 15% more of the year than do blood
transfusions = 54.75 days per patient per year = 54 750 days per 1 000 patients per

2-Quality of life
 Most modern medicine improves quality rather than quantity of life. The
limitations associated with effectiveness measures have led researchers to
develop ways of measuring the whole impact of a disease or treatment on a
 Definitions: it may be helpful to distinguish between the terms quality of life
(QoL) and health-related quality of life (HRQoL). The first term, QoL, is a broad
concept with many aspects that measures people's overall perception of their life.
QoL includes both health-related and non-health-related aspects of their lives
(e.g., economical, political, cultural).
 HRQoL is the part of a person's overall QoL that "represents the functional
effect of an illness and its consequent therapy on a patient, as perceived by
the patient."
 In health-related research and articles, the term QoL is often used interchangeably
with the term HRQoL, and both might be used to indicate the narrower definition
pertaining to a person's health.

Importance: The use of HRQoL measures has been increasing since the mid-1980s.
Traditionally, health has been considered from a biomedical point of view. From this
viewpoint emphasis is placed on activities associated with repairing injury and
reducing the impact or length of illness.
 Although this approach is essential, it does not include all of the aspects that are
important to health. A broad definition of health proposed by the World Health
Organization more than 50 years ago is: "Health is a state of complete physical,
mental and social well-being and not merely the absence of disease or
 Many health care providers and health services researchers have adopted this
expanded view of health and now include measures of the overall impact of
diseases and their treatments.
 In addition to physical function, the overall concept of HRQoL includes other
aspects of health, called domains, such as psychological and social functioning,
that are important to the patient.

Measuring QoL
Measuring QoL is methodologically complex. There are many functional, social,
psychological, cognitive and subjective factors that affect QoL.
QoL measures can be divided into generic and disease specific.

A-Disease-specific measures have been developed for patients with chronic

diseases; such measures can only be used to assess patients or treatments within
those disease states.
It is often necessary to focus on the impact that a certain disease or condition has on
patients. In this case, condition- or disease-specific measures are often used to collect
more narrowly focused patient views on the impact of the disease.
Examples of specific areas investigated with disease-specific questionnaires include
nausea and vomiting for cancer treatment, and range of movement for arthritis
Examples of disease-specific health status instruments are found in Table 8.4, and
some questions from Asthma Quality of Life Questionnaire are listed in Table 8.5.
The Quality of Life Instruments Database website (http://www.qolid.org)
summarizes many of the health status instruments.

Table 8.4. Examples of disease-specific HRQOL measures

Table 8.5. Examples of items from the asthma quality of life questionnaire (AQLQ):

B-Generic measures are more useful when looking at groups of patients who
may have different illnesses, and can be used to compare outcomes in different
patient groups. One of the most widely used is the Short Form (SF)-36 health survey.
This looks at:
• Physical functioning
• Physical role
• Bodily pain
• General health
• Vitality
• Social functioning
• Emotional role
• Mental health.
Typical questions asked include: Does your health limit you in these activities?

Table 8.3: Examples of items from the SF-36 general HRQOL instrument:

Using these types of tool can give a much better indication of the impact of the
treatment or service on the patient's QoL than using effectiveness measures.
In the example of the study of lipid-lowering drugs in men aged 45-54 with pre-
existing heart disease, a tool such as the SF-36 health survey could assess the effect
of reducing the incidence of coronary heart disease on the QoL of the patients who
have the treatment.
 Utility is a measure of the relative preference for various options.
 Utility is the value attached by an individual to a specific level of health or a
specific health outcome.
 Different individuals may attach different values to the same health state. For
example, some people may be prepared to tolerate a lot of nausea to allow them
to be pain free. Others may prefer to tolerate more pain and reduce the level of
 The important concept here is that utility measurement allows patients to
value their health status based on their own preferences.
 Like generic QoL measures, utility can be used when looking at groups of
patients who may have different illnesses, and can be used to compare
outcomes in different patient groups.
 Utility measures go beyond generic quality of life measures because they enable
quantitative comparison. Simply, utility is used to "attach a numerical value
to the value a person has for a particular health state.
 Imagine that treatment A improves a group of patients' health by an average of 6
points on a utility scale, and that treatment B improves a group of patients' health
by an average of 3 points. Treatment A can be said to be twice as effective as
treatment B. However, treatment A might be surgery for a ruptured Achilles
tendon and treatment B might be rhDNase for cystic fibrosis.
 This example shows that utility can be used to compare outcomes for very
different treatments in very different patient groups.

The specific methods used to derive utility are complex and are still under
development. They are preference-based, which means that they allow individuals
to indicate the direction and strength of their preference for a particular health state.
Attaching values to health states can be carried out using standard gamble or time
trade-off methods, or a rating scale. The last is rarely used.
A- Rating scale
To understand these methods it is necessary to be familiar with visual analogue
scales (VAS). Figure 4.1 illustrates a typical VAS.
To help you understand this, try the following:
1- Mark on the scale where you think indicates how you feel now.
2- Mark on the VAS where you would value your health state if you had pneumonia.
The difference between (1) and (2) is the difference in your health state, as valued
by you. So, if you had pneumonia, and you were given some antibiotics to cure it,
that difference in health state would be the health gain obtained by the drugs.

B-Standard gamble
The standard gamble is considered by some health economists to be the gold
standard for utility valuation. In this approach, an individual is asked to choose
between the following:
1-The certainty of surviving for a fixed period in a defined health state
(Choice A—livings in health state i (a chronic health state between perfect health
and death) with certainty).
2- Choice B: A gamble between a probability (p) of surviving for the same period
without disability or a probability (1 — p) of immediate death.
The probability (p) is varied until the person shows no preference (is indifferent)
between the certain option and the gamble.

As an example, a person considers two options: a kidney transplant with 20%

probability of dying (80% chance of returning to normal health) during the operation
(alternative B) or certain dialysis for the rest of his or her life (alternative A).
If the person says he or she would have the operation if the chance of the successful
operation p is 80% (chance of immediate death 20%), the percent chance of success
is decreased until the person reaches his or her point of indifference (the point
where the two options are nearly equal and the person cannot decide between
the two).
If the person says he or she would not have the operation if the percent chance for
success was 80% (chance of dying, 20%), the percent chance of success is increased
until the person reaches his or her point of indifference.
Let us say that the first person chooses a 70% chance (p) of a successful operation
(with a 30% chance [1-p] of immediate death) as the point of indifference between
having a kidney transplant and living with kidney dialysis for life.
The utility score for this person for this disease state or condition (kidney
dialysis) would be calculated as the probability (p) of living a normal life after
the operation, is 0.7.

The limitations of the standard gamble are that it is time-consuming, that people
have difficulty understanding probabilities, and that how people value health
states can be influenced by how the questions are phrased or presented.
A variety of other problems with the gamble have become apparent. For example,
treatment of most chronic diseases does not approximate the gamble. There is
no known product that will cure a patient with arthritis or one that is likely to kill
him or her. In other words, the decision-making experience of the patient is not likely
to include an option that has a realistic gamble.
C-Time trade-off (TTO)
The third technique for measuring health preferences or utilities is the TTO method
Figure 6.3. Again, the subject is offered two alternatives.
Alternative 1 is a certain disease state for a specific length of time (t), the life
expectancy for a person with the disease, and then death.
Alternative 2 is being healthy for time x, which is less than t. Time x is varied until
the respondent is indifferent between the two alternatives. The utility score for the
health state is calculated as x divided by t.
For example, a person with a life expectancy of 50 years is given two options:
Alternative 1 is being blind for 50 years, and alternative 2 is being healthy (including
being able to see) for 25 years followed by death. If the person says he or she would
rather be blind for 50 years than sighted for 25 years, the number of years (x) of sight
(healthy state) is increased until the person is indifferent between the two
alternatives. If the person would rather be sighted for 25 years than blind for 50
years, the number of years (x) of sight is decreased until the person is indifferent
between the two alternatives. Let us say that for a person who expects to live 50
more years, the person's point of indifference is 40 years of sight versus 50 years of
being blind. The utility score would be x/t = 40/50 or 0.8.

Figure 6.3: Time trade-off (TTO). This TTO schematic represents the choice a respondent makes
about trading off years of life for better health for a shorter period of time. The respondent is given
the choice of living a full life (to time t) with a specific condition or living fewer years (to time x)
without this condition (being healthy). The time of living healthy is varied until the respondent is
indifferent between living in full health x years and living with the condition for t years. The utility
calculated for the condition is x/t.
5-Preference-based multi-attribute health status measurement
Three main methods are used: Euroqol (EQ-5D), Quality of Well-Being (QWB)
and Health Utilities Index (HUI). EQ-5D is described below.
Euroqol (EQ-5D)
The EQ-5D (1997) is a standardized instrument for use as a measure of health
outcome. Applicable to a wide range of health conditions and treatments.
EQ-5D has five dimensions: mobility, self-care, usual activities, pain/ discomfort
and anxiety/depression. There are three levels per dimension and
respondents/patients describe themselves within this system. The EQ-5D
dimensions are shown in Table 4.3:
EQ-5D Quality of Life instrument (from http://www.euroqol.org):

This means there are 243 possible health states plus unconscious. This is what some
of these health states look like:
Health State 11111 Health State 21111
No problems walking about Some problems walking about
No problems with self-care No problems with self-care
No problems performing usual activities No problems performing usual activities
No pain or discomfort No pain or discomfort
Not anxious or depressed. Not anxious or depressed

The EQ-5D health state may be converted to a score.

5-Quality-adjusted life years
All the approaches above give utility measures that range from 0 to 1 where the
higher the score, the better the health-related quality.
These utility measures are used to generate quality adjusted life-years (QALY). A
QALY combines survival periods (quantity of life) with health status valuations
(quality of life) to provide a standard unit for measuring health gain. One QALY is
1 year in perfect health. One QALY could also be 2 years at 'half perfect health.
A treatment that moves a patient from 0.5 to 0.75 produces the equivalents of 0.25
QALY if it is maintained for 1 year. If this applied to 4 individuals, and duration of
the treatment effect is 1 year, the effect of the treatment would be equivalent to 1
completely well-year life. If you value health states using QALYs, you can
compare different treatments. This method would let you compare the health gain
from hip replacements with those due to antidepressant treatment, even though the
clinical indicators for these conditions are very different.

A theoretical worked example using QALYs

Utility data for the two alternatives available from the literature suggest that patients
maintained on erythropoietin value their health states at a higher level than those
maintained on blood transfusions.
In this study, 100 patients stated their utility for a treatment period of 10 years. The
utility value for each year (when valued from 0 to 1) on erythropoietin was 0.80,
whereas on blood transfusions it was 0.62. Figure 4.4 shows how those QALYs can
be represented graphically.
What is the difference in utility production of the two alternatives, i.e. how
many extra QALYs are produced by erythropoietin per year of treatment, for
1000 patients?
Incremental difference in utility = 0.80 - 0.62
= 0.18 QALYs per patient per annum
= 180 QALYs per 1000 patients per annum

Whose utility values should be used?

Utility values can be obtained from healthcare professionals, patients and the
general public. There are advantages and disadvantages associated with each group.
• Healthcare professionals are more informed about the health states and
interventions but may provide a biased value owing to their continued exposure to
that illness or intervention. Healthcare professionals have been shown to assign
lower ratings than patients or the general public.
• Patients are informed about the health status and interventions that they have
experienced. They will not be informed about interventions they have not
experienced. Patients tend to attach higher value of health status than do
healthcare professionals and the general public. This may be because people in a
health state gradually develop ways of dealing or coping with that health state,
whereas the general public are less informed or are valuing their fear of
experiencing that he state.
• Sometimes the health state has to be valued by proxy, such as the health state of a
newborn baby or a person with advanced Alzheimer's disease. Therefore, it is
important to remember -proxy values may be lower than the patient's values would

Expressing benefits as monetary values

Another method of measuring outcome is to convert these benefits to a monetary
value. The 'willingness to pay' (WTP) method, elicits monetary values for items not
typically traded in private markets, such as health.
WTP: is the use of survey to find the maximum persons who are prepared to
pay for a service that has been described using hypothetical (imaginary)
In simple terms, this method seeks to elicit how much an individual would be willing
to pay to avoid an illness or obtain the benefits of a treatment. WTP is
increasingly being applied to elicit preferences regarding the use of medicines, for
example in hypertension, lipid-lowering and depression. WTP has been used to elicit
preferences for the avoidance of side effects with antidepressants, identifying those
'most troublesome' to patients. Blurred vision and tremor were the side effects
considered most troublesome and were associated with the highest WTP values to
avoid them.

Example of willingness to pay

Imagine you have a headache. You can have medicine A or medicine B. You are
given the following information:
Medicine A and medicine B are equally effective for alleviating headache.
Medicine A makes 1 in 10 people feel sick.
Medicine B makes 3 in 10 people feel sick.
Which medicine do you prefer?
How much would you be 'willing to pay' to have medicine A?
This exercise is not asking you to guess how much medicine A or medicine B costs:
it is asking you to put yourself into the situation (health state) resulting from taking
one of the two medicines. Both will cure your headache, but medicine B has a higher
risk of nausea associated with it. What value, in dollars, do you attach to the
reduced risk of feeling sick?
How would your selection and willingness to pay change if you were told that
medicine A cures 50% of headaches and medicine B cures 90% of headaches? Now
you will have to decide whether you are willing to risk an increased chance of nausea
for an increased chance of cure (make a trade-off) (see Worked example 4.4).

EXAMPLE 4.4 Using WTP.

Let us go back to the erythropoietin example we looked at earlier.
A willingness-to-pay study for the two alternatives available from the literature
suggests that patients maintained on erythropoietin are 'willing to pay' for the extra
perceived health benefits over blood transfusions. Fifty patients in a study stated that
they would be willing to pay a mean of £2,000 a year for the extra health benefits
associated with erythropoietin.
What is the difference in benefit between the two alternatives, expressed in
monetary terms, i.e. how much are patients willing to pay for the health benefits
of erythropoietin per year of treatment, for the 1000 patients?
Change in benefit = £2,000 more benefit per annum per patient when given
Change in benefit = £2,000,000 more benefit per annum per 1000 patients when
given erythropoietin.

Hypothetical nature of WTP

 The primary concern for both advocates and critics of WTP is the hypothetical
nature of the scenarios, and hence the valuations elicited. Great efforts are made
in studies to develop realistic scenarios with understandable language and
minimum bias.
However, there is concern that respondents have difficulty responding because of
the hypothetical nature of the questions. Till present there is no healthcare study that
has compared hypothetical WTP responses with actual market rates, and so the
debate must remain unresolved.

Pharmacoeconomic Methodologies:
 There is often more than one way of doing something in healthcare.
 For example, there may be two different drugs that can be used to treat
depression, or two surgical techniques for the management of dysmenorrhoea.
 Note that interventions may be compared against each other (for example
antibiotic A against antibiotic B) or against a 'do nothing' scenario.
 There are different ways in which we can choose one of these options. We may
decide to pick the more effective surgical technique, or we may decide to select
the less costly antidepressant.
 Economic evaluation is a generic term for techniques that are used to identify,
measure and value both the costs and the outcomes of healthcare interventions.
 Economic evaluation is concerned with identifying the differences in costs and
outcomes between options.
 It can be defined as a study that compares the costs and benefits of two or
more alternative interventions; so, the main components are costs and benefits.
Types of economic evaluation
There are four main types of economic evaluation (table 1):
• Cost- minimization analysis (CMA)
• Cost-effectiveness analysis (CEA)
• Cost-utility analysis (CUA)
• Cost-benefit analysis (CBA).
Although they employ similar methods to define and evaluate costs, they differ in
the methods used to estimate the benefits from a program or intervention.

The components of economic evaluation

It is clear that economic evaluations can be understood in terms of the inputs (costs)
and outputs (benefits or outcomes) of a healthcare intervention (Figure 5.1).

Therefore, an economic evaluation requires the systematic identification of costs

and consequences of the healthcare interventions to be compared (Figure 5.2). Any
healthcare intervention can be seen as a process in this way.

Figure 5.3 illustrates how an operation can be shown as a process with inputs
(resources consumed) and outputs (effect on the patient).
Cost-minimization analysis (CMA)
In CMA, the outcome of the treatments being compared is the same. Having ensured
that the outcomes between the comparators are equivalent, then the approach used
is to consider the costs of each option. The preferred option is the cheapest.
 Example: suppose two antibiotics, G and C, are equally effective in the treatment
of Pseudomonas pneumonia, according to the current evidence. Therefore, we
should use the least costly alternative. The following assumptions have been
• The two interventions are equally effective.
• We have included all the costs.
 A recent economic evaluation of Atenolol versus Captopril in hypertensive
patients with type 2 diabetes reported that there was no statistically significant
difference in life expectancy between groups (Gray et al., 2001). However, the
mean cost per patient over the trial period was £6,485 in the Captopril group and
£5,550 in the Atenolol group. The reduction was statistically significant and
was due partly to differences in drug acquisition prices, and partly due to fewer
and shorter hospitalizations in the Atenolol group. The results from this CMA
would suggest that Atenolol should be used in preference to Captopril in this
group of patients.
 Another common example of CMA is comparing medications that are the same
chemical entity, the same dose, and have the same pharmaceutical properties
(i.e. they are bioequivalent) such as brand versus generic or generic made by one
company compared with generic made by another company. In these cases only
the cost of the medication itself needs to be compared because outcome should
be the same.
 Another example of CMA includes measuring the cost of receiving the same
medications in different setting. For example researchers could measure the
costs of receiving I.V. antibiotics in a hospital and compare this with receiving
the same antibiotics (at same doses) at home via a home health care service.

CMA is the simplest of the four types of pharmacoeconomics analysis because the
focus is on measuring the left-hand side of the pharmacoeconomic equation (the
cost) and the right –hand side of the equation (outcomes) is assumed to be the same.
This method is limited in use because it can only compare alternatives with the
same outcomes.

Cost-effectiveness analysis
Cost-effectiveness analysis is a technique designed to assist a decision-maker in
identifying a preferred choice among possible alternatives.
 Generally, cost-effectiveness is defined as a series of analytical and
mathematical procedures that aid in the selection of a course of action from
various alternative approaches.
 Cost-effectiveness analysis has been applied to health matters where the
program's inputs can be readily measured in dollars, but the program's outputs
are more appropriately stated in terms of health improvement created (e.g., life-
years extended, clinical cures).

Outcome measures in cost-effectiveness analysis

In CEA, outcomes are reported in a single unit of measurement, and are given in
natural units, for example mmHg for blood pressure reduction, or life years gained
by transplantation. The outcome measure is common to both alternatives, but
may be achieved to different degrees (i.e. there is a difference in effectiveness).
 An economic evaluation could examine the use of coronary artery bypass graft
(CABG) surgery for ischemic heart disease compared with medical (drug therapy
only) management.
 The effectiveness of both treatment methods can be measured using mortality at
10 years. Evidence suggests that it is likely that mortality will be lower if CABG
is used.
 Therefore, cost-effectiveness analysis is the appropriate method to use because
the outcome is common to the two alternatives, but there is a difference in

Cost-Effectiveness Ratios
 Results from a CEA are typically expressed as a cost-effectiveness (C/E) ratio;
the numerator of the ratio reflects total costs, while the denominator is the
expression of the outcome variable.
 Two forms of the C/E ratio exist: (1) average, or simple, and (2) incremental
(ICER). The average/simple C/E ratio is a straightforward approach, defined as

Although average C/E ratios provide useful information for analysts, incremental
analyses are considered a hallmark of CEA.
 In cost-effectiveness analysis (and cost-utility analysis) you will come across the
regular use of incremental economic analysis. This is a systematic method for
identifying the difference (increment) in costs and outcomes between two
healthcare interventions.
 The following questions are always asked:
 What is the difference in cost between the interventions?
 What is the difference in outcome between the interventions?

The answers to these questions allow the explanation of the incremental cost-
effectiveness ratio (ICER). Incremental cost/outcome ratios may be calculated using
the following equation:
 Outcome1 is the number of patients successfully treated with intervention 1.
 Outcome 2 is the number of patients successfully treated with intervention 2.
 Cost 1 is the cost of treating patients with intervention 1.
 Cost 2 is the cost of treating patients with intervention 2.

The ICER expresses the cost required to achieve each extra unit of outcome.
When one alternative is more effective but requires more resources, the ICER must
be calculated. In the situation when one alternative is more effective and less costly,
this alternative is the dominant therapy. When there is dominance, ICERs do not
need to be generated (see Worked example 5.1).

EXAMPLE 5.1 incremental economic analysis:

An economic evaluation could examine the first-line management of community-
acquired pneumonia using antibiotics A or C. The effectiveness for both treatment
methods can be measured using 'infections successfully treated first line'.
The incremental economic analysis is carried out in the following way:
What are the costs associated with treatment with
a) Antibiotic A [Cost A]
b) Antibiotic C [Cost C]
What are the outcomes associated with
a) Antibiotic A [Outcome A]
b) Antibiotic C [Outcome C]
What is the difference in cost between using antibiotic A and antibiotic C?
[Cost A – Cost C]
What is the difference in outcome between using antibiotic A and antibiotic C?
[Outcome A – Outcome C]
ICER for treating community-acquired pneumonia with antibiotic A instead of
antibiotic C
Ways of using outcome data in incremental economic analysis
 ICERs present as the cost per unit of outcome.
 This could be cost per life year gained, cost per death averted, cost per case
successfully diagnosed, or cost per patient successfully treated.
 It can also be cost per mmHg reduction or cost per percentage point improvement
in lung function.
 For example, in the cost-effectiveness analysis of rhDNase in children with cystic
fibrosis, the outcome measure was percentage improvement in FEV1. In this
study, it's reported that the ICER to be £200 per 1% gain in FEV1. This approach
is often used in CEA.
 The advantage of this outcome measure is that it is objective and generally
accepted as 'true'.
 However, sometimes it is not easy to interpret the clinical significance of this
type of ICER. An improvement of 1% in FEV1 is not likely to be clinically
significant. Therefore, we might be more interested in knowing the ICER for an
improvement of 10% or 20% in FEV1 which is more likely to be clinically
significant. The ICER will also be correspondingly larger.

Another approach would be to convert the results into numbers of patients

successfully treated:
 We can easily explain this if we consider the cost-effectiveness analysis of
rhDNase in children with cystic fibrosis again.
 Clinicians working in the management of people with cystic fibrosis consider that
a clinically significant improvement in FEV1 is 10% or greater.
 Therefore, an alternative approach would be to examine the change in FEV1 of
each patient in the study and allocate them to 'successfully treated' or
'unsuccessfully treated'.
 Then the ICER could be calculated on the basis of patients rather than
percentage change in FEV1.
 The disadvantage of this method is that it may introduce subjective judgments
about whether a patient has been treated successfully or not.

Using an incremental cost-effectiveness ratio to make a decision

 After reading Worked example 5.1, you may now ask yourself: which antibiotic
do you think should be chosen by the healthcare decision maker?
 This example illustrates that carrying out incremental cost-effectiveness
analysis does not necessarily provide an obvious option: this will only happen
when dominance occurs, where it is clear that the more effective, less costly
option should be selected.
 However, the decision-maker usually has to select between the more costly, more
effective option and the less costly, less effective option.

 The generation of the ICER allows us to see how much extra cost is incurred
for the extra benefit.
 It is then left to the decision-maker to make a value judgment as to whether the
extra benefit is worth the extra cost.
 In the example above, the decision-maker must decide whether they think that
the extra case of pneumonia successfully treated with antibiotic C is worth £200.
This is a disadvantage of cost-effectiveness analysis.

Should the incremental cost-effectiveness ratio be large or small?

 The larger the ICER, the more money is required to buy each unit of outcome.
Therefore, as an ICER becomes larger, the intervention is said to be less cost-
 For example, a CEA of reducing cholesterol concentration with statins generated
cost per life year gained for different types of patients in which the drugs may
be indicated. Table 5.1 shows the different ICERs generated for these different
groups of patients.
 It can be seen that different groups of patients had very different ICERs.
 To generate one additional life year women aged 45-54 with a history of angina
and cholesterol 5.5-6.0 mmol/1 it would cost £361,000. This is 60 times what it
would cost to generate one additional life year in men aged 55-64 with a history
of myocardial infarction and cholesterol above 7.2 mmol/1.
Therefore, although statins are effective and safe in all these groups of patients, they
have very different levels of cost-effectiveness. It is generally accepted that
healthcare providers cannot afford to treat all patients in whom statin treatment is
likely to be effective. The implications of these different magnitudes of ICERs,
in a resource-constrained healthcare system are that they may be used to
prioritize which patients receive statins.

Cost-Effectiveness Grid
A cost-effectiveness grid can be used to illustrate the definition of "cost-
effectiveness" (Figure 5.1). To determine if a therapy or service is cost effective,
both the costs and effectiveness must be considered. Think of comparing a new drug
with the current standard treatment. If the new treatment is:
1) Both more effective and less costly (cell A),
2) More effective at the same price (cell B), or
3) Has the same effectiveness at a lower price (cell D),
The new therapy is considered cost effective.

On the other hand, if the new drug is:

1) Less effective and more costly (cell I),
2) Has the same effectiveness but costs more (cell F), or
3) Has lower effectiveness for the same costs (cell H).
Then the new product is not cost effective.

For the middle cell E, other factors may be considered to determine which
medication might be best. For the other two cells (C and G), an ICER is calculated
to determine the extra cost for each extra unit of outcome.

 Dominant strategies are defined as offering both lower cost and higher
effectiveness compared with an alternative, while a dominated strategy is one
that costs more than the comparator and is less effective.
 Options requiring a trade-off include technologies that present a higher cost with
higher effectiveness or lower cost with lower effectiveness relative to

Cost-Effectiveness Plane
A graphical depiction of cost-effectiveness comparisons is also sometimes seen in
the literature. Figure 5.3 is a cost-effectiveness plane.
 The point on the plane where the x and y axes cross indicates the starting point
of costs and effectiveness for the standard comparator.
 A point is placed in the plane for each alternative to the standard comparator by
indicating how much more or less it costs than the starting point (y-axis) and how
much more or less effective it is than the starting point (x-axis).
 If an alternative is more expensive and more effective than the standard
comparator, this point will fall in quadrant I, and a Trade-off is required (i.e., is
the increased cost of the new treatment worth the increase in effectiveness?)
 If an alternative is less expensive and more effective, the point would fall in
quadrant II, and the alternative would dominate the standard comparator.
 If an alternative was less costly and less effective, the point would fall in quadrant
III, and again a tradeoff would have to be considered. (Do the costs savings of
the alternative outweigh its decrease in effectiveness?).
 If an alternative was more expensive and less effective, the point would fall in
quadrant IV, and the alternative would be dominated by the standard
To illustrate the differences between the simple/average and incremental C/E ratios,
Example One presents two treatments as follows:
Treatment One (a baseline comparator) costs $500 to treat 100 patients and
treatment Two (a new innovation) costs $750 to treat the same number of
The effectiveness measure is a final outcome, cure rate; where in 95 percent of the
patients are cured with treatment One and 97 percent of the patients are cured with
Treatment Two.
Calculating the average C/E ratio yields a relatively small difference between
the average costs to treat the patient groups between the two alternatives,
However, an incremental analysis finds the cost of treatment two to be
substantially higher, costing $12,500 to obtain one additional cure.
Cost-Utility Analysis (CUA)
What Is Cost-Utility Analysis?
CUA is a formal economic technique for assessing the efficiency of healthcare
interventions. It is considered by some to be a specific type of cost-effectiveness
analysis in which the measure of effectiveness is a utility- or preference-adjusted
Utility is the value or worth placed on a level of health status, or improvement in
health status, as measured by the preferences of individuals or society.
Measurement of health state utilities or preference values is necessary for calculation
of the most commonly used outcome measure in this type of analysis: quality-
adjusted life-years (QALYs) gained.
CUA has some distinct advantages over cost-effectiveness analysis:
 Cost-effectiveness analysis is limited by the inability to simultaneously
incorporate multiple outcomes from the same intervention or to compare
interventions with different outcomes.
 In cost-effectiveness analysis, although the outcome measure is in natural units
(eg, life years saved), no attempt is made to value the consequence or outcome in
terms of quality or desirability.
 In contrast, CUA incorporates the quality of (or preference for) the health
outcome achieved.
CUA, using QALYs gained as the outcome measure, is the most common approach
to combining quantity and quality-of-life outcomes in economic evaluations.
In a cost-utility analysis (CUA) the outcomes of the two alternatives are measured
using utility values, that is, the value attached to the health states produced by the
two interventions. The value may be attached by patients, health professionals or
the general population, but the last is preferred. CUA is actually a form of cost-
effectiveness analysis but utility is used instead of natural units to measure

When Is Cost-Utility Analysis Appropriate:

There are several circumstances in which CUA may be the most appropriate analytic
1. When health-related quality of life is the important outcome—for example, when
comparing interventions that are not expected to have an impact on mortality, but
a potential impact on patient function and well-being (eg, treatments for
2. When health-related quality of life is an important outcome—for example,
evaluation of the outcomes associated with the treatment of acute myocardial
infarction. Not only a life saved an important outcome measure, but also the
quality of the lives saved.
3. When the intervention affects both morbidity and mortality and a combined
unit of outcome is desired—for example, evaluation of a therapy, such as
estrogen use by postmenopausal women, that can improve quality of life, may
reduce mortality from certain conditions (eg, heart disease), but may increase
mortality from other conditions (eg, uterine cancer).
4. When the interventions being compared have a wide range of potential
outcomes and there is a need to have a common unit of outcome for
comparison. This is most commonly the case when a decision-maker must
allocate limited resources among interventions that have different objectives and
resultant benefits—for example, the choice between providing increased prenatal
care or expanding a hypertension screening and treatment program.
5. When the objective is to compare an intervention with others that have already
been evaluated in terms of cost per QALY (or equivalent) gained.

WORKED EXAMPLE 6.1 (ex: 5.2 contin…)

Economic evaluation of management of anaemia in haemodialysis patients:
Patients with chronic renal failure who are on haemodialysis suffer from profound
anaemia, which is often extremely debilitating. This is due to a reduction in the
production of erythropoietin in these patients, and loss of blood during
haemodialysis. Historically, these patients have been managed by the use of blood
transfusions. Now, synthetic erythropoietin is available. It is considered to be highly
effective, but is very expensive. So the alternatives are to either give erythropoietin
or to give blood transfusions when the patient's "haemoglobin level is below 8g/dl.

Cost information:
Total costs to manage the 1000 patients for 1 year using blood transfusions:
Total costs to manage the 1000 patients for 1 year using erythropoietin: £5,547,100.

Outcome information:
Utility data for the two alternatives available from the literature suggest that patients
maintained on erythropoietin value their health states at a higher level than those
maintained on blood transfusions. In a study, 100 patients stated that for a treatment
period of 10 years, their utility value for each year (when valued from 0 to 1) on
erythropoietin was 0.80, whereas on blood transfusions it was 0.75.
1. What is the difference in cost between the two alternatives for the 1000
2. What is the difference in utility production of the two alternatives, i.e. how
many extra QALYs are produced by erythropoietin per year of treatment, for
the 1000 patients?
Change in utility = 0.80 - 0.75
= 0.05 QALYs per patient per annum
= 50 QALYs per 1000 patients per annum.

Figure 6.1 illustrates the difference in utility production for the two alternatives.
Figure 6.2 Cost-effectiveness plane for erythropoietin Vs blood transfusions

Results from a study suggest that the practice will have cost reductions and the
patients will have improved outcomes (Table 6.1)
The practice has to decide whether to reduce emergency admissions due to asthma
by 10 a year and save £1,000, or reduce emergency admissions due to chest pain by
25 a year at no change in costs to the practice.

How can the GP objectively compare and choose between improving the health
of asthma and IHD patients?
Groups A and B elicit utility values from the 200 asthma and 250 IHD patients. Time
trade-off was used to elicit the utility values and these were used to calculate
The groups obtain the following results:
Asthma patients IHD patients
Mean QALYs before intervention 0.75 0.60
Mean QALYs after intervention 0.85 0.75
Incremental QALY change caused 0.10 0.15
by intervention

The results refer to a 1-year period. The asthma patients improved their quality of
life per year by 0.10 QALYs each. The IHD patients improved their quality of life
per year by 0.15 QALYs each.
Incremental cost-effectiveness ratio (ICER):

If the GP practice funds the IHD service it will cost them £1,000 per year more than
the asthma service, but they will obtain 17.5 more QALYs for their patients.

EXERCISE 1: Calculating a cost per QALY:

Several treatments exist to treat fungal toenail infections. Four oral medicines used
are drugs A, B, C and D. The table below shows the costs (£) associated with treating
one patient with each of these four treatments:

Drug A Drug B Drug C Drug D

Total costs (£) 1,301 1,503 1,570 1,200

You then find some evidence to suggest that two of these agents have differing
effects on patients' quality of life owing to difference in their side-effect profiles.
This evidence is summarized below:

Agents Increase in QALYs per patient per year

Drug C 0.10
Drug D 0.05
Which treatment will you recommend to your Trust, and why?
Either could be recommended, depending on the driving force for the choice. Is cost
containment most important? Then choose drug D. Is improved patient outcome
most important? Then choose drug C.

Cost-benefit analysis
In a cost-benefit analysis (CBA) the outcomes of the two alternatives are measured
using monetary values, that is, the monetary value attached to the health states
produced by the two interventions. The value may be attached by patients, health
professionals or by the general population, but the last is preferred.
Advantages and Disadvantages of Cost-Benefit Analysis
An advantage of this type of analysis is that many different outcomes can be
compared as long as the outcomes measures are valued in monetary units. The
disadvantage is that placing economic values on medical outcomes is not an easy
task and there is no universal agreement on one standard method for accomplishing

To illustrate the advantage of CBA compared with cost-effectiveness analysis

(CEA), Table 7.1 shows examples of various programs and interventions and their
corresponding cost-effectiveness and cost-benefit ratios. (CBA ratios are expressed
as benefit-to-cost ratios, where the higher the number, the more cost benefit.

Assume you are a decision maker and you must choose one program from Table 7.1
to implement in your organization.
Assume that you only had cost-effectiveness ratios available to help make the
choice. How would you choose? One can quickly see that it would be difficult to
compare the programs using only cost-effectiveness ratios because of the varying
outcomes (e.g., case prevented, life years saved).

On the other hand, the benefit-to-cost ratios can be ranked, and programs with
similar, as well as dissimilar, outcomes can be compared.

If only the cost-effectiveness ratios were available, it would be more difficult to

compare the value of the various interventions.

The financial pressures are forcing decision makers to consider the following
questions: Do the benefits of a program or intervention outweigh the costs? And
which program will provide the greatest benefit? CBA is a tool that can be used to
address these questions.

Conducting a Cost-Benefit-Analysis

The first step in a CBA is to determine the type of program or intervention to be

considered. The second step is to identify alternatives. In many cases, the alternative
is to "do nothing." In other cases, the alternative could be to implement a similar
program that is smaller or larger in scale or to implement a different program.

For example, a clinical pharmacist would like to start an asthma clinic. The
alternative could be to compare the costs and benefits of having an asthma clinic
with not having an asthma clinic. Another alternative could be to compare
implementing an asthma clinic for all persons who had an asthma-related emergency
department visit. A third alternative could be to compare implementing an asthma
clinic with implementing a diabetes clinic.
To illustrate the components of a CBA, we will use the example of an asthma clinic.
The clinic will focus on people with asthma who have had an asthma-related
emergency department visit. These people would automatically be referred to a
clinical pharmacist, who would provide education on managing asthma. These
clinical pharmacy services could include education on triggers, medication
adherence, and the use of peak flow meters and inhalers. In this example, the
alternative will be no asthma clinic. After the program or intervention and
alternatives are identified, the next step is to identify the costs and benefits.

Figure 7.1. Components of cost-benefit analysis (CBA). This schematic represents the types of
costs measured when conducting a CBA. Input costs (the numerator) usually consist of direct
medical and direct nonmedical costs. The benefits of alternatives can include measures of direct
medical and nonmedical costs avoided, indirect costs avoided (measured by human capital [HC]
or willingness-to-pay [WTP] methods) and intangible costs avoided (measured by patient
preferences or WTP methods).

Figure 7.1 shows the basic components of CBA. As shown, there are two categories
of costs, direct medical and direct nonmedical, and three categories of benefits,
direct benefits (both medical and nonmedical), indirect benefits (productivity), and
intangible benefits. CBA can incorporate as few as one category of benefits or as
many as all three of the benefit categories.

Before starting any pharmacoeconomic analysis, it is important to determine the

perspective of the study. Because of its focus on social welfare and policy and the
incorporation of indirect (productivity) or intangible benefits, economists
recommend that CBAs should be conducted from the societal perspective.

Difference between Costs versus Benefits

In CBA, both costs and benefits are measured in dollar values. For example, in the
asthma program, a cost to the program could be an increase in medical costs related
to visits to the pharmacy. A "cost saving" or benefit as a result of the program could
be a reduction in medical costs for asthma- related emergency department visits.

Measuring Indirect and Intangible Benefits

Various methods have been developed to estimate the monetary value of health
benefits. The two most common methods seen in the pharmacoeconomic literature
are the human capital (HC) approach and the willingness to pay (WTP) approach.

Human Capital Method

As indicated previously, indirect benefits are increases in productivity or earnings

because of a program or intervention. The HC approach is one way to measure
indirect benefits. HC estimates wage and productivity losses because of illness,
disability, or death. The HC approach assumes that the value of health benefits
equals the economic productivity that they permit. There are two basic components
to calculating HC: wage rate and missed time (days or years) because of illness.
Because the HC approach is based on wages, it is necessary to have some estimate
of income. Income estimates can be obtained from several sources (self-report; or
any other data source that provides income estimates based). Missed time (days or
years) because of illness can be obtained by self-report.

Wage Rate Calculations

Depending on the type of study, a yearly wage rate or a daily wage rate can be
calculated. A yearly wage rate (income per year) would be calculated for a program
or intervention that would reduce long-term disability or death. For example, a
pneumococcal vaccination program might result in preventing premature death.
Thus, it would be appropriate to use a yearly wage rate and assess the value of the
number of years saved because of the intervention.

A daily wage rate (income per year ÷ by number of days worked per year) may be
calculated for a program or intervention targeted at an acute or chronic illness with
short-term disability. A person may not be adversely affected by the disease state
on a continual basis, but he or she may have short-term periodic disability. For
example, asthma, a chronic disease state, may include episodic asthma attacks. Thus,
a person may only experience problems with the disease state on a periodic basis.
For this type of disease state, a daily wage rate would be calculated. To calculate a
daily wage rate both income and number of days worked per year must be assessed.

We may assume that the average person works 240 days a year when accounting for
weekends, vacation, and sick leave. A formula to calculate number of days worked
per year is.

Number of days in a year (365) — Number of weekend days (104) — Number of

vacation days (14) - Number of sick-leave days (7) = 240.
Missed Time (Days or Years) Because of Illness:

If a yearly wage rate is calculated, then assessment of the number of years lost
because of a disease or illness must be made. If a daily wage rate is calculated, an
assessment of the number of missed days because of illness must be calculated.
Because many pharmaceutical interventions involve chronic disease states with
intermittent episodes, we will use an example calculating the daily wage rate and
number of missed days. Missed days because of illness can fall into four groups,
(Table 7.2)

Notice that for housekeeping and child care, estimates of productivity loss are
estimated (imputed) even though no payments are directly associated with these

Using the asthma clinic example, we will calculate an indirect benefit. Assume that
the population served by the clinic is made up of adults with an average income of
$40,000 and 240 days worked per year. The daily wage rate (average income/number
of days worked per year) would be $40,000/240 = $167/day. An average of 20 days
a year were missed from work before participating in the asthma clinic, and an
average of 7 days a year were missed from work after participating in the asthma
clinic. Multiplying the daily wage rate times the number of missed days results in
the value of lost productivity. In other words, the value of 20 days lost from work is
$3340, and the value of 7 days lost from work is $1169. The difference between
before and after the program is $2171, which is the cost savings or the indirect
benefit of the program or intervention (see Table 7.3 for the calculation).

Advantages and Disadvantages of the Human Capital Method

Measuring indirect benefits using the HC approach has several advantages:

 It is fairly straightforward and easy to measure.

 Income estimates can be obtained or estimated from publicly available sources.
 Days lost from illness can be readily obtained from the patient or another
secondary source.
The HC approach also has several disadvantages:

 The primary concern with using the HC approach is that it may be biased against
specific groups of people, namely unemployed individuals.
 It assumes that if a person is not working, he or she has little or no economic
 Children and unemployed elderly individuals are two groups with which bias
can occur.
 The HC assumption that the value of health benefits equals the economic
productivity they permit may also be biased. The earnings for some individuals
may not equal the value of their output.
For example, there is a large difference between the daily wage rates of a
professional football player compared with that of an elementary school teacher.

 Some contend that because the underlying goal of using CBA is to measure
the effect of an intervention on society, the HC approach is meant to measure
the loss of productivity to society.
 Thus, wage rates should be based on those of the average population, not the
specific patients included in a study.
 Although using general wage rates would not represent actual productivity
losses or benefits to a specific group of patients, it would decrease some of the
limitations of inequity already mentioned.
 The HC method also does not incorporate values for pain and suffering if these
values do not impact productivity. There may be certain disease states or
conditions (e.g., menopause, hair loss) that may not impact productivity but do
have an impact on a person's health-related quality of life. For example:
 Many women experience problems with menopause, including moodiness,
hot flashes, and irregular cycles.
 Although this condition may have a significant impact on quality of life,
most women do not miss many days of work because of complications from
 Thus, the HC method would not be sensitive enough to capture the benefits
of a pharmacist-provided menopause clinic.
 But although biases exist with this method, it is the most commonly used
method to measure indirect benefits.
Willingness to Pay Method

 The WTP method can value both the indirect and intangible aspects of a disease
or condition.
 The WTP method determines how much people are willing to pay to reduce the
chance of an adverse health outcome.
 The WTP method is grounded in welfare economic theory, and it incorporates
patient preferences and intangible benefits such as quality of life differences.
 WTP values can be collected through face-to-face interviews, mail, telephone, or
via the Internet.
 Respondents are then asked to value the health care intervention in a dollar
 Measuring WTP should include two general elements, a hypothetical scenario
and a bidding vehicle.

Hypothetical Scenario

 The hypothetical scenario should include a description of the health care program
or intervention (e.g., medication therapy, management program, or new drug
 The aim of the scenario is to provide the respondent with an accurate description
of the service that he or she is being asked to value. In addition, the scenario
should detail the amount of time the person should expect to spend, as well as the
benefit (e.g., percent improvement in the condition) of the intervention.
 An example of a hypothetical scenario for the asthma clinic might read:
Asthma Clinic Scenario
Patients with asthma have improved their condition by learning more about their
disease and by taking their medications as directed. Pharmacists can help people
with asthma understand their condition and the medications used to treat it. In
addition, they can:

1-Help you learn how to use a peak flow meter and an inhaler.

2-Help you better manage the medications used to treat asthma.

3-Help you recognize and handle situations when asthma attacks occur.

4-Monitor your asthma by keeping a record on file and following up with you-
on a regular basis to assess your progress.

5-Contact your doctor and report any changes in your health.

An initial visit with your pharmacist would include an educational program on

managing your disease state. This type of service is available by appointment only
and would last approximately 1 hour. Assume that the program would result in a
50% improvement in your asthma.

Bidding Vehicles

After the program or intervention has been adequately described, respondents are-
then asked to "bid," or place a value on the program or intervention. Bids can be
obtained through a variety of formats, such as open-ended questions, closed-ended
questions, a bidding game, or a payment card. Below is a brief description of each
of the methods.
Open-Ended Questions Open-ended questions simply ask respondents how
much— they would be willing to pay for the program or intervention. This question
would immediately follow the hypothetical scenario. Here is an example:

What is the maximum amount that you would be willing to pay for a 1-hour
consultation with a pharmacist? --------------------------------------------.

The respondent would then write in their maximum WTP amount.

This method is used the least because it results in WTP values that vary widely.
Many people do not know how to value health care programs because they do not
normally pay the full amount out of pocket. The other methods discussed later
provide respondents with more guidance in determining their maximum WTP.

Closed-Ended Questions Closed-ended questions are also called "take-it-or-leave-

it" questions. Respondents are asked whether or not they will pay a specified dollar
amount for the program or intervention.
Here is an example: Would you be willing to pay $60 for a 1-hour consultation
with a pharmacist? _________________ .Yes or No

This method more closely resembles the marketplace. When consumers shop
for products, they must decide based on the price of the product whether to "take-it-
One drawback to this method is that only one question is asked, so only one
WTP value can be elicited from a respondent. Thus, a very large sample would be
required to determine the overall WTP value.

Bidding Game The bidding game resembles an auction in that several bids are
offered to reach a person's maximum WTP. Before soliciting a second response, the
bids are adjusted based on the first response. This iteration could go on a number of
times, but it is suggested that three times is optimal. Here is an example:
Would you be willing to pay $60 for a 1-hour consultation with a pharmacist?
__________________ If yes, ask: "Would you be willing to pay $80?"
__________________ If no, ask: "Would you be willing to pay $40?"
This method is useful to try to arrive at a person's maximum WTP value. It is time
consuming and is best conducted via a face-to-face interview or over the Internet. In
addition, the WTP values can be biased depending on how high (or low) the first bid
is. This is called "starting point bias".
Payment Card The payment card method provides the respondent with a list of
possible WTP amounts (i.e., payment card) to choose from. Here is an example:
What is the maximum amount that you would be willing to pay for a 1-hour
consultation with a pharmacist? Please circle your choice.
$150 $90 $30
$130 $70 $10
$110 $50 $0
This method is very easy to use and it provides respondents with a range of
values to choose from. The advantages of the method can also result in
Providing respondents with a range of values can bias their WTP values. The
range provided can "suggest" the value of the intervention and can influence what
respondents say.
Also, "range bias" can influence the WTP amount. For example, if the range
of values was from $0 to $75 versus $0 to $150, the respondents' WTP amount can
vary depending on which range or starting point was provided.

Advantages and Disadvantages of the Willingness-to-Pay Method

The main advantage of the WTP approach is that it is a method to place a dollar
value on intangible benefits.
However, there are several disadvantages to the WTP methodology. It is
difficult for people to place a dollar value on a health benefit or an increase in health-
related quality of life or satisfaction.
Because a "hypothetical" or artificial scenario is presented, it is possible that
respondents might give a "hypothetical response" or that the respondent!" may not
understand the value of the market (e.g., pharmaceutical care program" being
Calculating Results of Costs and Benefits
After all costs and benefits have been identified and quantified, the results of
the analysis must be presented in ways that help decision makers understand the
value of the program or intervention.
CBA can be presented in the following three formats: net benefit calculations,
benefit-to-cost ratios, and internal rates of return (IRR).
When evaluating interventions, it is important to consider the time horizon for
the project. If retrospective data are collected for more than 1 year or if the project
inputs or outcomes are estimated for more than 1 year into the future, it is important
to adjust or discount these costs one point in time.

Net Benefit (or Net Cost) Calculations.

The net benefit (or net cost) calculation simply presents the difference between
the total costs and benefits.
Net benefit = total benefits — total costs;
Net cost = total costs — total benefits.
Interventions would be considered to be cost beneficial if:
Net Benefit > 0 or Net Cost < 0

Benefit-to-Cost (or Cost-to-Benefit) Ratio Calculations.

CBA results can also be calculated by summing up the total benefits and
dividing by the total costs. The ratio may be expressed as a benefit-to-cost ratio or a
cost to-benefit ratio. Depending on how the ratio is calculated, interventions are cost
beneficial if:
Benefit-to-cost > 1 or Cost-to-benefit < 1

Example Using Different Calculation Techniques.

Suppose a decision maker had to choose between two proposals for implementation.
Also assume that the projects are for 1 year, so discounting is not needed.
Proposal A: Cost = $1000; Benefit = $2000
Proposal B: Cost = $5000; Benefit = $7500

Table 7.4 shows the net and ratio calculations for both proposals. Although four
calculations are shown in the table, the benefit-to-cost ratio (when compared with
the cost-to-benefit ratio) and the net benefit calculation (when compared with the net
cost calculation) are used most often because the higher the result, the more cost
beneficial an option becomes.
Using the criteria outlined above for cost-beneficial programs, it is apparent
that both programs are cost beneficial using both the net and ratio methods of
calculations. However, when comparing net calculations, proposal B is more cost
beneficial than proposal A (net benefit = $2500 versus $1000), but proposal A is
more cost beneficial than proposal B (benefit-to-cost ratio = 2.0 versus 1.5) when
using ratio calculations.
In this example, in which both proposals are cost beneficial, the decision maker may
consider other issues, such as the amount of money available for investment.
Whereas A would require $1000 input costs, proposal B would require $5000.
Another consideration may involve the return on investment. Proposal A, with a
2:1 benefit-to-cost ratio, has a higher return than proposal B (i.e., 1.5:1 benefit-to-
cost ratio).
‫ يعود دوالرين‬Proposal A ‫اي ان الدوالر الواحد المصروف على‬
‫ يعود دوالر ونصف‬Proposal B ‫وان الدوالر الواحد المصروف على‬
A third consideration is the actual net benefit amount. Proposal B has a higher net
benefit than proposal A ($2500 versus $1000).
Using cost-benefit analysis to allocate resources to different services
Worked example 7.1 shows how WTP can be used in a CBA to generate net
benefit. However, in that example we were trying to decide between two ways of
treating the same illness. We could also have used a common outcome such as the
presence of anemia, and the resulting economic evaluation would have been a CEA
(see 5, worked examples).
It is more difficult to compare two healthcare interventions that do not have
comparable outcome measures. CEA cannot be used in this situation, and we have
to use an outcome measure that can be used across different diseases, such as WTP.
CBA can be used to generate net benefit in different disease areas, and so diseases
with different clinical outcomes can be compared. This means that CBA can be used
to allocate resources to different services. Worked example 7.2 illustrates how this
can be carried out.
Cost-Benefit or Cost-Effectiveness Analysis
Cost-benefit and cost-effectiveness analyses are useful tools for assessing the
clinical economic impact of medical care programs or interventions. There are,
however, several important distinctions between the two approaches.

First, CBA may be applied to single or multiple programs, while cost-effectiveness

analysis is applied to multiple programs.
Second, CBA may be used to compare programs with disparate outcomes. In
contrast, cost-effectiveness analysis is a method for identifying the least costly
approach to achieving a single outcome.
A third distinction is that CBA requires that all the outcomes or benefits be assigned
a dollar value. The outcome or effect is not valued in cost-effectiveness analysis.
Which approach should you use in the pharmacy arena? It depends. A general
guideline is that cost-effectiveness analysis is most appropriate when a single effect
or outcome can be defined.
CBA is usually most appropriate when a single program is to be evaluated or
when budget allocation decisions must be made among programs with unrelated
The use of decision analysis in economic evaluations:
Decision analysis is a technique that can be used to incorporate information and
estimates in a systematic way to compare different options. Decision analysis is
being used more commonly in pharmacoeconomic evaluations.
Decisions are made all the time in healthcare where the results of those
decisions are not certain. For example, if you treat a person with pneumonia using
an antibiotic, there may be an 80% chance it will work. However, you cannot be
100% sure that it will work, so there is uncertainty about the effectiveness of that
antibiotic. Therefore, the decision to treat that patient is made under conditions of

What are the sources of uncertainty?

When a decision is made in healthcare, there is uncertainty around the outcome of
that decision. There are lots of ways in which uncertainty can exist:
 Diagnosis: for example, diagnostic tests do not always give the correct result.
 Natural history of the disease: not everyone with the same disease will feel the
same way or will suffer the same ill effects of that illness; for example, not
everyone with hypertension will eventually have a myocardial infarction.
 Treatment efficacy and effectiveness: no treatment is 100% effective, so there
is always a chance it will not work in everyone; for example, as antidepressant
 The development of adverse events: some people show side effects or allergic
responses to drugs, such as penicillins.
 Unit cost of resources: the price of services changes over time and between
places; for example, drugs, staff time.
The common theme in these examples is that you can never predict exactly what is
going to happen when you make a decision, but a decision must be made.
This is what is meant by decision-making under conditions of uncertainty. In fact,
most decisions we make are made under conditions of uncertainty, whether they are
about healthcare or other aspects of our lives.

Probability: the language of uncertainty

It is not possible to remove uncertainty from the decision-making process, but it is
possible to quantify it. Knowing the level of uncertainty can influence our decision.
For example, if you were about to cross a road and you were told that you had a 20%
risk of death, would you still cross? In decision analysis we represent uncertainty
by using probabilities.
Probability can have many related meanings:
• Number between 1 and 0 expresses likelihood of event: for example, what is
the probability of successfully treating a urinary tract infection (UTI) with
• Probability as proportion in a population: for example, what is the probability
of having appendicitis if you are between 11 and 16 years old?
• Probability as a measure of strength of belief: we all use the words 'possibly'
and 'probably' as subjective measures of probability.

Principles of probability
The sum of probabilities of all possible outcomes of a chance event is always 1.
 If the probability of an antibiotic successfully treating a case of pneumonia is 0.8,
then the probability that it will not work must be 0.2.
 There may be three or more possible outcomes of a decision. For example, the
possible outcomes of a total hip replacement operation may be survival with
improved mobility, survival with no improvement in mobility, or
perioperative death.
 If the probability of perioperative death is 1% (0.01) and the probability of
survival with improved mobility is 85% (0.85), then the probability of survival
with no improvement in mobility must be 14% (1 - (0.01 + 0.85) = 0.14).


What is the probability of picking a man from the population in Baghdad who is
both hypertensive [p = 0.09) and has arthritis (p = 0.02)?
The probability of this happening is 0.09 X 0.02 = 0.0018 (0.1 8%).
What is the probability of picking a man from the population in Baghdad who is not
hypertensive (p = 1 — 0.09) and has arthritis (p = 0.02)?
The probability of picking a man from the population in Baghdad who is not
hypertensive (p = 1 — 0.09) and has arthritis (p = 0.02) is:
(1 -.0.09) x 0.02 = 0.0182 (1.82%).

Use of decision analysis to design economic evaluations

The aim of economic analysis of an intervention is to determine whether, under
specified conditions, it is cost-effective. Economic evaluation can be considered to
consist of many stages.

1-Step 1: Identify the research question.

The specific decision to be evaluated should be clearly defined by answering the
questions: What is the objective of the study?
For example the decision is whether to add a new antibiotic to an institutional
formulary to treat infections.
Step 2: Specify Alternatives
 Ideally the most effective treatments or alternatives should be compared.
 In pharmacotherapy evaluations, makers of new products may compare or
measure themselves against a standard (i.e., older, more well-established)
 Decision analysis could compare more than two treatment options (e.g., it could
compare the five most common statins) or an intervention versus no intervention
(e.g., a diabetes clinic versus no clinic).
 For the example problem, the use of the new medication (antibiotic A) will be
compared with that of the current standard (antibiotic B).

Step 3: Draw the Decision Analysis Structure or tree

Once these stages have been completed, we can build our decision tree.
A decision tree has five principal components:
1. Starting point: at which point in the process we begin the evaluation of the
2. All treatment alternatives under investigation: the different strategies under
3. Decision nodes ■: there should only be one decision node: the policy decision
of whether to use one strategy or the other.

4. Chance nodes : these are uncertain events and will have probability values

attached to them.
5. Outcome/time horizon ◄: the outcome being used must be defined and the
point at which evaluation ends (time horizon).

The decision tree in Figure 8.1 shows two alternatives for treating urinary tract
infections (UTIs).

The starting point is the patient group who has been diagnosed with a UTI that
now needs to be treated. At the decision node, the policy decision is whether to treat
this group of patients with the standard current treatment (drug T) or whether to use
a newer, more costly agent, drug C.
There is no 'do nothing' option here because current practice is to treat
symptomatic UTIs, to alleviate symptoms, and also to prevent complications such
as pyelonephritis. It would, however, be possible to include more antibiotics in the
model and have more arms in the tree, if it were felt to be necessary.
The probabilistic event here is whether or not the antibiotic is successful in
treating the infection.
Figure 8.2 show decision tree for the treatment of depression using selective
serotonin-reuptake inhibitors (SSRIs) or tricyclic antidepressants (TCAs).

It would also be possible to expand the tree by including the probabilistic events
of side effects and withdrawal from treatment. The endpoint of the evaluation is
whether the antibiotic is successful or not, and the time horizon would probably be
quite short -about 7 days in this intervention.

Another example:
Step 4: Specify Possible Costs, Outcomes, and Probabilities
 For each option, information should be obtained for the probability of occurrence
and the consequences of the occurrence.
 Probabilities are assigned for each branch of the chance nodes, and the sum of
the probabilities for each branch must add up to 1.00.
 Consequences are reported as monetary outcomes, health-related outcomes, or
Table 9.1 lists these data for the antibiotic example.
Table 9.1. Estimates for the antibiotic example
Antibiotic A Antibiotic B
Probability of clinical success 90% 80%
Cost of antibiotic per course of therapy $600 $500
Probability or adverse events 10% 15%
Cost of treating adverse events $1000 $1000

Step 5: Perform Calculations

 At each terminal node, the probability of a patient having that outcome is
calculated by multiplying the probability of each arm from the choice node to the
terminal node.
 The total costs for each terminal node are calculated by adding up the costs over
all of the branches from the choice node to the terminal node. The product of the
costs multiplied by the probability (C X P) is calculated for each node, and then
summed for each option.
 In our example, each of the two options (antibiotic A versus antibiotic B) has
four possible terminal endpoints: success/no adverse events, success/adverse
events, failure/no adverse events, failure/adverse events.
 Table 9.2 and Figure 9.2 show the calculations used to estimate the average
expected cost per treatment. Note that the sum of the probabilities for the four
terminal endpoints equals 1.00

For patients taking antibiotic A, the costs can range from $600 (for medication and
no adverse events) to $1600 (for medication and treatment of adverse events), the
average cost is $700 per patient. Similarly, for patients taking antibiotic B, the costs
can range from $500 (for medication and no adverse events) to $1500 (for
medication and treatment of adverse events), and the average cost is $650 per patient.
These calculations show that antibiotic B is less expensive even when including the
costs of treating adverse events. But because antibiotic A is a better clinical option
(higher probability of success and lower probability of adverse events), decision
makers could use the incremental cost-effectiveness ratio (ICER) to determine
whether to add antibiotic A to the formulary. The calculated ICER would be:

If it is decided that each extra successful outcome is worth at least $500 (patient
discharged from the hospital faster, prevention of second round of treatment costs
with another antibiotic, and so on), then antibiotic A would be added to the

Step 6: Conduct a Sensitivity Analysis

Because some uncertainty surrounds the estimates used to construct these models, a
sensitivity analysis is conducted.

WORKED EXAMPLE 8.8 Theoretical example of use of a decision tree in an

incremental economic analysis of the treatment of blue finger syndrome
Several drugs are indicated for a very distressing condition known as 'blue finger
syndrome', where patients' fingers turn dark blue for no reason. The drugs used are
associated with lots of side effects. Some people cannot tolerate the drugs (this
means they have side effects that are so bad they have to stop taking them) or the
drugs may not work. If these things happen, they will be referred to their local
outpatient clinic to see a finger specialist. This may be an important factor in the
ultimate choice of therapy. You are the adviser to your local healthcare provider.
They want to know whether to use indigociliin or navytriptyline when a person
presents to their GP with blue finger syndrome. About 200 people per year present
with blue finger syndrome.

The additional cost of treating a patient who withdraws from either drug per year is
£ 500.
Figure 8.3 shows the decision-analytical model for this intervention. The
information above allows us to calculate how many of the 200 patients will go down
each arm of the model. We also know how much each arm costs for one patient.
Therefore, we can calculate how much each arm costs in total
Markov modelling
A simple decision tree may not be capable of modelling chronic disease states. A
model trying to represent a chronic disease, such as relapsing-remitting multiple
sclerosis must be capable of reflecting changes in and out of health states. These
may be referred to as random processes that evolve over time. They are random
because we do not know when they will occur in the disease progression. Markov
models are particularly useful for representing the use of interventions to manage
chronic health states.
A decision-analytical model may become unnecessarily complex, as patients will
move in and out of health states many times. An alternative method for presenting
these events is shown in Figure 8.4.
This shows a simplified version of what can happen to a person with relapsing-
remitting multiple sclerosis. When they are symptom free there is a probability they
will have a relapse, stay symptom free or die. When they are experiencing
symptoms, there is a probability they will become symptom free, the relapse may
continue, or they may die. When a patient dies, they cannot return to the other health
states. Therefore, death is referred to as the 'absorbing' state.
Markov models therefore simulate the natural history of a chronic illness such as
multiple sclerosis in a population of patients over a period of time, and its associated
risk of relapse, remission and death.
The population of patients moves through the model over time. The model will
estimate how many patients are in remission, have relapsed or have died at any given
time. Probabilities of moving from one state to another will be obtained from clinical
data sources.