POTENTIAL MEDICATION DOSING ERRORS

IN OUTPATIENT PEDIATRICS
HEATHER A. MCPHILLIPS, MD, MPH, CHRISTOPHER J. STILLE, MD, MPH, DAVID SMITH, RPH, MHA, PHD, JULIA HECHT, PHD, MPH,
JOHN PEARSON, MD, JOHN STULL, MD, KRISTIN DEBELLIS, PHARMD, SUSAN ANDRADE, SCD, MARLENE MILLER, MD, MSC,
RAINU KAUSHAL, MD, MPH, JERRY GURWITZ, MD, AND ROBERT L. DAVIS, MD, MPH

Objective To determine the prevalence of potential dosing errors of medication dispensed to children for 22 common
medications.
Study design Using automated pharmacy data from 3 health maintenance organizations (HMOs), we randomly selected up
to 120 children with a new dispensing prescription for each drug of interest, giving 1933 study subjects. Errors were defined as
potential overdoses or potential underdoses. Error rate in 2 HMOs that use paper prescriptions was compared with 1 HMO that
uses an electronic prescription writer.
Results Approximately 15% of children were dispensed a medication with a potential dosing error: 8% were potential over-
doses and 7% were potential underdoses. Among children weighing <35 kg, only 67% of doses were dispensed within recom-
mended dosing ranges, and more than 1% were dispensed at more than twice the recommended maximum dose. Analgesics were
most likely to be potentially overdosed (15%), whereas antiepileptics were most likely potentially underdosed (20%). Potential
error rates were not lower at the site with an electronic prescription writer.
Conclusions Potential medication dosing errors occur frequently in outpatient
pediatrics. Studies on the clinical impact of these potential errors and effective error
See editorial, p 727.
prevention strategies are needed. (J Pediatr 2005;147:761-7)
From the Departments of Pediatrics and
Epidemiology, University of Washington,
hildren in the United States are evaluated by a physician almost twice a year on and the Center for Health Studies, Group

C average, and receive medications during up to 60% of these visits.1 Data from
case reports, poison control centers, and the Food and Drug Administrations
(FDA) MedWatch system2-4 indicate that serious and even fatal medication errors occur
Health Cooperative, Seattle, Washing-
ton, Meyers Primary Care Institute, Uni-
versity of Massachusetts Medical School
and Fallon Foundation, Worcester, Mas-
sachusetts, Center for Health Research,
Kaiser Permanente Northwest, Portland,
in children each year. Although most medical care for children takes place in the outpatient Oregon, Quality and Safety Initiatives,
setting, very little is known about the frequency and types of medication errors, their Johns Hopkins Childrens Center, Balti-
clinical importance, or effective strategies for error reduction in this setting. more, Maryland, and the Department
of Medicine, Brigham and Womens Hos-
Children are particularly vulnerable to medication dosing errors. Accurate pediatric pital and Harvard Medical School, Bos-
prescribing requires accurate weight, proper conversion of pounds to kilograms, and the ton, Massachusetts.
Supported by a contract to the
choice of an appropriate preparation and concentration. The prescriber must calculate a HMO Research Networks Integrated
daily dose on the basis of weight and then divide it into multiple doses for the appropriate Delivery System Research Network
(AHRQ Contract No. 290-00-0015),
frequency of medication. Their small size makes children more vulnerable than adults and by a grant to the HMO Research
to substantial dosing errors because of a misplaced decimal point or a trailing zero.5-7 Network Center for Education and
Research on Therapeutics (CERTs)
Unfortunately, when medication errors occur, children cannot always demonstrate signs (AHRQ U18HS10391), from the Agency
or communicate symptoms that may alert parents and providers to seek proper care. for Healthcare Research and Quality
(AHRQ U18HS11843-01).
Implementation of computerized physician order entry (CPOE) with decision sup- Submitted for publication Mar 31, 2005;
port has been shown to reduce medication errors in hospitalized children by 40% to 97%.8,9 last revision received Jun 27, 2005;
CPOE, although not well studied in pediatric outpatients, theoretically could reduce most accepted Jul 27, 2005.
Reprint requests: Heather McPhillips,
calculation and decimal point errors. In addition, with appropriate decision support, incor- MD, MPH, 4800 Sand Point Way NE,
rect dosage regimens could be avoided. Mailstop G0061, Seattle, WA 98105.
E-mail: heather.mcphillips@seattlechildrens.
org.
CI Confidence Interval HMO Health maintenance organization 0022-3476/$ - see front matter
CPOE Computerized physician order entry MinRDD Minimum recommended daily dose Copyright 2005 Elsevier Inc. All rights
DOI Drugs of interest MaxRDD Maximum recommended daily dose reserved.
FDA Food and Drug Administration OR Odds ratio
10.1016/j.jpeds.2005.07.043

761
 We conducted a population-based retrospective study
in 3 health maintenance organizations (HMOs) to assess the
prevalence of potential medication dosing errors in ambulatory
pediatrics. We also studied whether CPOE without pediatric
dosing decision support at 1 study site was associated with
fewer potential medication dosing errors.
METHODS
Setting and Eligibility
This study was conducted at 3 HMOs; children were
eligible if they were under the age of 17 years at the dispensing
date and enrolled with pharmacy benefits anytime between
June 1999 and June 2001. The 3 HMOs had approximately
242,000 eligible children enrolled at the study midpoint.
HMO A had a CPOE module within the electronic
medical record used for almost all prescriptions. Although
the CPOE module contained decision support for some
medications, it did not contain child-specific decision support,
nor did it automatically calculate doses or check for errors. The
CPOE module pattern matched to an entered drug and of-
fered a menu from which the prescriber selected the specific
size tablet or concentration of liquid. For some medications
studied, the menu included options that contained the specific
amount of the formulation to be taken at a specified interval.
Once the order was signed electronically, it was retained in
the patients electronic medical record. At HMO B and C,
prescriptions were hand-written and taken by the patient to
a network pharmacy for dispensing.
Study Design
We selected drugs of interest (DOI) using 2 comple-
mentary processes. We selected medications most commonly
prescribed to children at HMOs A and B, as well as those
medications most commonly involved in pediatric error re-
ports to the FDAs Medwatch system. An external panel of
advisors recommended additions to or subtractions from the
list, leading to 14 medications most commonly involved in
pediatric error reports, and 19 most frequently prescribed
medications. Because of some overlap, 22 DOIs were selected
overall (Table I). Prescriptions for these 22 medications
accounted for approximately 33% of all dispensing events to
children at the HMOs during this time period.
For each drug, we randomly selected children dispensed
a new prescription within the study period (index dispensing).
We defined a prescription as new if the child had not received
a dispensing for that medication within the previous 6 months
for children ages 3 and older or the previous 2 months for chil-
dren <3 years. We selected up to 60 children per drug at HMO
A, and up to 30 children per drug at HMOs B and C, for a
possible total of 120 children per medication (or all available,
if there were less than 120 children having been dispensed
a new prescription for the DOI).
To maximize prescriptions where weight-based dosing
would be expected, we limited our selection to children ages
1 day to 12 years of age. For 5 drugs, we included children
762 McPhillips et al
ages 16 and under because these medications were uncom-
monly dispensed to younger children.
For each index dispensing event, we used automated
pharmacy data to calculate the total dose in milligrams (mg)
per day and dosing interval. Administrative databases were
linked to determine the age, sex, and number of diagnostic
codes used during the index visit, as well as the number of
other medications dispensed on the same date. For the 4 anti-
epileptic drugs included in this study, we also determined
from automated data whether laboratory monitoring had
been performed for serum levels of drug within 45 days before
or after the date of the index dispensing event.
Trained medical record reviewers recorded information
on each childs weight, diagnoses at the index visit, and
prescription medication(s), including dose and formulation.
Weights were included in calculation of mg/kg/d if they were
documented within 30 days of drug dispensing for children un-
der 3 years old or within 6 months for children ages 3 years and
older. If weights were not documented, total mg/d was used to
determine whether doses were within the recommended range
for medications where weight-based dosing was not necessary.
For each child, we calculated total mg/d and mg/kg/d, as
well as total number of doses per day (dose/day). We described
these doses as being below minimum recommended daily dose
(MinRDD), within recommended range, or above the maxi-
mum recommended daily dose (MaxRDD) based on health
plan formularies and widely used guidelines.10,11 If differences
existed between formularies and other guidelines, we chose the
smallest minimum dose and largest maximum dose from these
sources and included doses that fell between these 2 numbers
as within the RDD.
For medications commonly prescribed on an asneeded
basis, or prn (ie, analgesics, albuterol, hydroxyzine, and cloni-
dine), we calculated the daily dose as the amount a child would
possibly receive if he or she was given the upper range of the dose
at the most frequent dosing interval. For example, if a prescrip-
tion was written as 1-2 teaspoons every 3-4 hours prn, we cal-
culated the daily dose as the total number of milligrams the child
would have received if given 2 teaspoons every 3 hours for a
24-hour period. While we recognized that this likely overesti-
mated the potential error rate for most prn medications, it
gave an accurate estimate of the potential error rate that would
occur if a parent or guardian administered the highest pre-
scribed dose at the most frequently prescribed interval.
Outcomes
A potential medication error was defined as a medica-
tion dispensed at a dose meeting any of the following criteria:
(1) total mg/kg/d dispensed at 110% or more of the MaxRDD
(potential overdose); (2) total mg/d dispensed at more than the
maximum recommended adult dose (potential overdose); (3)
total mg/kg/d dispensed below 90% of the MinRDD and
below the adult minimum recommended dose in total mg/d
(potential underdose).
For example, 4500 mg/d of amoxicillin dispensed to
a 50-kg child (90 mg/kg/d) was counted as a potential error
because the dose exceeded the MaxRDD for adults (3000
The Journal of Pediatrics  December 2005
Table I. Drugs of interest
Generic drug of interest Total dispensing events (N) % With weight % Potential error

Analgesics
Acetaminophen/codeine 98 89 17
Ibuprofen 75 91 1
Naproxen 97 90 10
Oxycodone* 99 81 28
Antibiotics
Amoxicillin 116 91 3
Azithromycin 94 89 33
Cephalexin 118 93 12
Trimethoprim/sulfa 105 91 1
Antiepileptic drugs
Carbamazepine 48 94 8
Phenytoin* 42 100 41
Phenobarbital* 21 81 17
Valproic acid 57 91 27
Asthma and allergy medications
Albuterol 115 95 16
Hydroxyzine 117 93 35
Prednisolone 116 90 7
Psychotropic medications
Amphetamine 109 79 20
Bupropion 95 91 11
Clonidine 89 94 2
Fluoxetine 80 90 5
Methylphenidate 115 77 26
Paroxetine 109 87 2
Olanzapine* 59 80 20
Other
Isotretinoin* 89 64 9
Total 2063 88 15
*Included dispensing events to children ages 13 to 16.

mg/d),10 despite being within the RDD for weight-based
dosing. A dose of 750 mg/d of amoxicillin prescribed to a
50-kg child (15 mg/kg/d) was counted as appropriate because
the dose was within the adult RDD for mg/day despite being
less than the MinRDD for mg/kg/d (20 mg/kg/d).10 We al-
lowed dosing of 10% above or below weight-based dosing
thresholds to allow for minor rounding errors.
For children <35 kg, where weight-based dosing is most
important, we further characterized the range of mg/kg/d
doses dispensed as percent above the MaxRDD or percent
below the MinRDD.
Statistical Analysis
We used descriptive measures to describe the extent and
types of potential medication dosing errors and logistic regres-
sion to test for factors associated with potential medication
errors and whether CPOE reduced that risk (adjusting for
age, sex, number of medications prescribed, medication class,
and timing of visit). We used 5 or more unique medications
dispensed on the same day as a proxy for children with com-
plex medical conditions.
The sample size for the assessment of CPOE was based
on the assumption of studying 500 children at each of the 2
HMOs without CPOE and 1000 at the site with computer-
ized prescribing, which would give us 80% power to detect
a 50% reduction in error in the site with CPOE with 95%
confidence compared with the 2 sites with hand-written pre-
scriptions. Previous studies demonstrated a medication error
rate between 5% to 6% in inpatient encounters.12 We used
6% as a baseline error rate for our power calculations.
Human Subjects
The Institutional Review Boards of the participating
HMOs approved this study.
RESULTS
Of 2063 children initially identified as eligible for the
study, weights were available for 1810. An additional 123 chil-
dren were included on the basis of total mg/d dosing dis-
pensed, giving a total of 1933 children included in the study.

Potential Medication Dosing Errors In Outpatient Pediatrics 763

Table II. Characteristics associated with potential medication dosing errors
Characteristic % Within RDD (n) % UD (n) % OD (n)

Class of drug*
Analgesics (n = 341) 85 (288) 1 (2) 15 (51)
Antibiotics (n = 396) 88 (351) 9 (35) 3 (10)
Antiepileptic medication (n = 159) 80 (127) 20 (31) 1 (1)
Asthma and allergy medication (n = 324) 81 (261) 7 (24) 12 (39)
Psychotropic (n = 656) 88 (574) 5 (36) 7 (46)
Isotretinoin (n = 57) 91 (52) 9 (5) 0
Prn medications*
Yes (n = 561) 86 (452) 5 (26) 15 (83)
No (n = 1372) 87 (1201) 8 (107) 5 (64)
Formulation
Liquid suspension (n = 744) 84 (623) 8 (63) 8 (58)
Tablet, capsule, other (n = 1189) 87 (1030) 6 (70) 7 (89)
Total (n = 1933) 85 (1653) 7 (133) 8 (147)
UD, Potential underdose; OD, potential overdose.
*P < .05, x2 test.

Of 1933 children with index dispensing events, 15% (280)
contained potential medication errors. Of these 280, 147 pre-
scriptions (8%) were potentially overdosed, and 133 (7%) were
potentially underdosed.
Analgesics were the most frequent class of drugs with
potential overdoses (Table II). Fifteen percent of analgesic dis-
pensing events were above the MaxRDD, with most (28 of 51
potential overdoses) occurring for oxycodone. Antiepileptic
drugs were most likely to be potentially underdosed. Of the
33 new prescriptions for antiepileptic drugs that contained
potential medication errors, only 16 had documented serum
levels obtained within 45 days before or after the date of
dispensing.
Medications commonly prescribed prn were more
than 3 times as likely to be potentially overdosed (RR 3.2,
95% CI 2.3-4.3). Liquid formulations were no more likely
to be associated with a potential medication error than tablet
or capsule formulations. Of the medications prescribed in
liquid formulation, 76% (568 of 744 doses) were prescribed
in teaspoon increments as opposed to mL or cc. There calculation of mg/kg dosing (94% vs 77% at HMO B and 85%
was no difference in potential error on the basis of how the
prescription was dispensed (tsp versus cc).
Overall, 20% of dispensing events to children <4 years
old were associated with potential medication errors, com-
pared with 13% of dispensing events to children ages 4 to 12
(OR 1.7, 95% CI 1.1-2.6) (Table III). Asthma and allergy
medications and antibiotics were most likely to be associated
with potential medication errors in this youngest age group.
For asthma and allergy medications, 24% of dispensing events
(10% underdoses, 14% overdoses) contained potential errors.
Thirteen percent of antibiotics dispensed to children under 4
were potentially underdosed, and 9% were potentially over-
dosed. Adolescents aged 13 to 16 years also had a higher
rate of potential medication errors compared with children
4 to 12 years of age. Of the 18 dispensing events associated
with potential overdosing in adolescents, 17 were for oxyco-
done. For children dispensed 5 or more unique medications
in addition to the index prescription, 9 of 43 index medications
(21%) were potentially overdosed, compared with 63 of 1001
(6%) prescriptions dispensed to children receiving only the in-
dex drug (OR 3.3, 95% CI 1.4-7.7). Approximately 55% of
children in this study received their index dispensed medica-
tion within 2 days of visiting a provider. Those children with-
out a clinic visit 2 or fewer days before drug dispensing were
50% more likely to be dispensed a potential overdose (OR
1.7, 95% CI 1.2 to 2.5).
The HMO with CPOE had similar rates of prescrip-
tions dispensed within the recommended range compared
with the other 2 HMOs (87% vs. 84%). For medications dis-
pensed more than 50% above the MaxRDD in mg/kg/d (2.5%
of all dispensing events), there was no difference in potential
error rates between the HMO with electronic ordering and
the other 2 HMOs. The HMO with the CPOE did have sig-
nificantly more documented weights available for appropriate
at HMO C [P < .01]).
Table IV demonstrates the wide variation in dosing
observed for children weighing <35 kg where weight-based
dosing is most appropriate. Among children <35 kg, only
67% of medications were dispensed within the recommended
dosing range. More than 1% of medications were dispensed at
more than twice the MaxRDD on the basis of total mg/kg/d
for children weighing <35 kg.
DISCUSSION
We found that potential dosing errors in outpatient
medications are common in children, occurring in 15% of pre-
scriptions that we studied. One of every 5 children younger
than 4 years of age receiving any medication, 1 in 5 children

764 McPhillips et al The Journal of Pediatrics  December 2005

Table III. Risk factors for medication errors
Characteristic % RDD (n) % UD (n) % OD (n) OR (95% CI)

Age in years
0 to 3 (n = 336) 80 (268) 10 (33) 10 (35) 1.7 (1.1 to 2.6)
4 to 12 (n = 1404) 87 (1,227) 6 (83) 7 (94) 1.0
13 to16 (n = 193) 82 (158) 9 (17) 9 (18) 1.1 (0.6 to 1.9)
Sex
Male (n = 1132) 85 (967) 6 (66) 9 (99) 1.7 (1.2 to 2.4)
Female (n = 801) 86 (686) 8 (67) 6 (48) 1.0
Number of medications
Index medication only (n = 1001) 86 (859) 8 (79) 6 (63) 1.0
1 to 4 additional (n = 889) 86 (761) 6 (53) 8 (75) 1.4 (1.0 to 2.1)
5 or more additional (n = 43) 77 (33) 2 (1) 21 (9) 3.3 (1.4 to 7.7)
Clinic visit within 2 days
Yes (n = 1031) 88 (904) 6 (62) 6 (65) 1.0
No (n = 836) 82 (689) 8 (68) 9 (79) 1.7 (1.2 to 2.5)
Health plan
HMO A (with CPOE) (n = 1006) 87 (871) 5 (50) 8 (85) 1.0
HMO B (n = 396) 87 (345) 8 (31) 5 (20) 0.7 (0.4 to 1.1)
HMO C (n = 531) 82 (437) 10 (52) 8 (42) 1.1 (0.7 to 1.6)
RDD, % Within recommended daily dosing range; UD, % potentially underdosed; OD, % potentially overdosed; OR (95% CI), odds ratio
and 95% confidence interval for potential overdoses adjusted for class of drug.

Table IV. Weight-based dosing by class of drug for children < 35 kg

Percent (n) of medications dispensed at percentage below MinRDD
or above MaxRDD in mg/kg/d
Drug class <50% 50%-74% 75%-99% RDD 101%-124% 125%-149% 150%-200% .200%

Analgesics (n = 101) 0 (0) 0 (0) 3 (3) 79 (80) 9 (9) 5 (5) 2 (2) 2 (2)
Antibiotics (n = 279) 1 (2) 3 (7) 12 (33) 81 (225) 3 (9) 0 (0) 1 (3) 0 (0)
Antiepileptics (n = 111) 5 (5) 9 (10) 14 (15) 70 (78) 3 (3) 0 (0) 0 (0) 0 (0)
Asthma and allergy (n = 286) 7 (20) 6 (16) 14 (41) 57 (163) 8 (22) 4 (11) 3 (8) 2 (5)
Psychotropic (n = 273) 2 (5) 11 (30) 12 (32) 59 (161) 5 (15) 6 (16) 3 (9) 2 (5)
Total (n = 1050) 3 (32) 6 (63) 12 (124) 67 (707) 6 (58) 3 (32) 2 (22) 1 (12)
RDD, Within recommended daily dosing range.

receiving a prn medication, and 1 in 6 children receiving an
analgesic were potentially receiving an improperly dosed med-
ication. Particularly of concern to us was the high rate of poten-
tial overdoses of analgesics dispensed to children, because
these drugs have a higher likelihood than other drugs of seri-
ous adverse and even fatal events associated with improper
dosing.3 In addition, the children most vulnerable to poten-
tially serious adverse drug events, young and medically com-
plex children, were also those most likely to be dispensed
potential overdoses of medications.
The IOM has identified computerization of medication
prescribing as an important patient safety strategy.13 Comput-
erized order entry, combined with advanced decision support
systems, has been shown to reduce prescribing errors in hospi-
tals.14-18 In our study, the HMO with electronic ordering but
no dosing decision support for children did not have lower
potential medication dosing error rates compared with the
other HMOs. Although the study was not powered to detect
reductions of less than 50%, we had hypothesized that the use
of technology would decrease medication error rates because
of the standardization of information and the prevention of
errors because of mistakes in deciphering handwritten pre-
scriptions. Although providers chose the formulation of the
medication from a drop-down menu, they still had to calculate
and enter weight-based doses for most drugs prescribed. It is
important to realize that many commercially available CPOE
packages currently available are similar in that they do not
contain robust pediatric (weight-based) dosage decision
support and do not contain mechanisms to alert providers to
potential underdosing or overdosing of medications on the

Potential Medication Dosing Errors In Outpatient Pediatrics 765

basis of weight. Although many health care systems are work-
ing toward implementing ambulatory CPOE, it is important
that these systems take into account the complexities around
prescribing medications to children.
We found wide variation in dosing of most medications.
Only 1, trimethoprim-sulfamethoxazole, was dosed within
the recommended weight-based dosing range for all doses
dispensed. This medication has a limited number of com-
monly used outpatient formulations (3) and a simple standard
weight-based dosing range for most diagnoses. The simplicity
in formulation and dosing may have contributed to correct
dosing and suggests that standardization of medication dosing
regimens for children may be an important strategy for pre-
venting ambulatory medication errors.
There are some limitations to this study. This method
did not permit us to know whether prescriptions were written
outside standard dosing recommendations as a result of hu-
man error or as an intentional act by the clinician on the basis
of the clinical scenario. We assessed 22 drugs; hence, our find-
ings cannot necessarily be generalized to all medications used
in pediatrics. Additionally, because we chose this group of
drugs in part because they had been associated previously
with medication errors reported to the FDA, overall potential
dosing errors in ambulatory pediatrics may be less frequent
than we observed. Nevertheless, the medications we studied
accounted for approximately one third of all prescriptions to
children in the 3 health plans, representing a problem with
substantial public health significance.
We did not evaluate whether adverse events were
associated with the potential errors we observed, nor did
we address errors caused by overlooked drug allergies, drug
interactions, or errors in duration. Because we used a different
method to count and report potential errors, our results may
not be comparable with previously published research in this
area. Most potential medication dosing errors we observed
were likely not linked directly to adverse patient outcomes.
Although surrogate measures of patient safetyincluding
doses of medication outside standard recommendationsare
not necessarily evidence of harm in and of themselves, they
are a reasonable means by which to study the performance of
the medical system, and a reasonable end point to use when
devising intervention strategies. Finally, some of the prescrip-
tions we labeled as containing potential errors might have
been clinically necessary. For example, some of the antiepileptic
medications might rightfully have been prescribed with the
intent to titrate the dosage in accordance with the clinical
response. However, roughly half of the new antiepileptic
medication prescriptions that we labeled as potentially errone-
ous did not have evidence of a serum level measured within
45 days of the initial prescription.
HMO enrollees in this study had demographics repre-
sentative of those in their surrounding metropolitan areas.
However, because HMO enrollees have slightly higher educa-
tional attainment and fewer members in the lowest income
brackets, it is possible that our results underestimate rates of
potential errors that would be seen outside of these settings.
Furthermore, there are differences in physician, pharmacy,
766 McPhillips et al
and microsystem characteristics between HMOs and other
methods of delivering ambulatory care including private prac-
tice pediatrics. Therefore these results may not be generaliz-
able to non-HMO settings.
Our findings suggest that children who are prescribed
medication outside of a clinic visit are at higher risk for receiv-
ing a potentially inappropriately dosed medication; however,
we did not gather more specific information about these situ-
ations. We do not know how children were prescribed medi-
cation without a clinic visit or whether providers had access to
patient charts or patient weights in these scenarios. Many chil-
dren likely had a prescription called into the pharmacy for
them without visiting the clinic. It is also possible that a pro-
vider outside of the health plan saw these children, but their
prescriptions were filled at a health plan pharmacy. These chil-
dren also may have been prescribed medication during a clinic
visit of a sibling or other family member. The risk of error may
differ substantially depending on whether the provider has
access to accurate patient weight when writing a prescription
outside of the clinic setting.
Potential medication dosing errors are common in
outpatient pediatrics, occurring in approximately 1 of 7
new prescriptions in this study. These potential errors occur
among all type of medications given to children, including
narcotic analgesics. Health care providers and parents need
to remain vigilant about medication doses prescribed, partic-
ularly for young children and children with complex health
conditions.
REFERENCES
1. Cherry DKB, Catharine W, Woodwell, David A. National Ambulatory
Medical Care Survey: 2001 Summary: Centers for Disease Control and
Prevention; August 11, 2003.
2. Litovitz TL, Klein-Schwartz W, Rodgers GC Jr, Cobaugh DJ, Youniss
J, Omslaer JC, et al. 2001 Annual report of the American Association of
Poison Control Centers Toxic Exposure Surveillance System. Am J Emerg
Med 2002;20:391-452.
3. Phillips J, Beam S, Brinker A, Holquist C, Honig P, Lee LY, et al.
Retrospective analysis of mortalities associated with medication errors.
Am J Health Syst Pharm 2001;58:1835-41.
4. Heubi JE, Barbacci MB, Zimmerman HJ. Therapeutic misadventures
with acetaminophen: hepatoxicity after multiple doses in children. J Pediatr
1998;132:22-7.
5. Koren G, Barzilay Z, Greenwald M. Tenfold errors in administration
of drug doses: a neglected iatrogenic disease in pediatrics. Pediatrics 1986;
77:848-9.
6. Rieder MJ, Goldstein D, Zinman H, Koren G. Tenfold errors in drug
dosage. CMAJ 1988;139:12-3.
7. Perlstein PH, Callison C, White M, Barnes B, Edwards NK. Errors
in drug computations during newborn intensive care. Am J Dis Child 1979;
133:376-9.
8. Potts AL, Barr FE, Gregory DF, Wright L, Patel NR. Computerized
physician order entry and medication errors in a pediatric critical care unit.
Pediatrics 2004;113(Pt 1):59-63.
9. King WJ, Paice N, Rangrej J, Forestell GJ, Swartz R. The ef-
fect of computerized physician order entry on medication errors and
adverse drug events in pediatric inpatients. Pediatrics 2003;112(Pt 1):
506-9.
10. The Harriet Lane handbook: a manual for pediatric house officers. 16th
ed. Philadelphia: Mosby; 2002.
11. Practitioners guide to psychoactive drugs for children and adolescents.
The Journal of Pediatrics  December 2005
12. Kaushal R, Bates DW, Landrigan C, McKenna KJ, Clapp MD,
Federico F, et al. Medication errors and adverse drug events in pediatric
inpatients. JAMA 2001;285:2114-20.
13. Kohn LC, JM; Donaldson MS, ed. To err is human: building a safer
health system. Washington: National Academy Press; 2000.
14. Wilson AL, Hill JJ, Wilson RG, Nipper K, Kwon IW. Computerized
medication administration records decrease medication occurrences. Pharm
Pract Manag Q 1997;17:17-29.
15. Bates DW, Leape LL, Cullen DJ, Laird N, Petersen LA, Teich JM,
et al. Effect of computerized physician order entry and a team intervention
on prevention of serious medication errors. JAMA 1998;280:1311-6.
16. Bates DW, Teich JM, Lee J, Seger D, Kuperman GJ, MaLuf N,
Boyle D, et al. The impact of computerized physician order entry
on medication error prevention. J Am Med Inform Assoc 1999;6:
313-21.
17. Kaushal R, Barker KN, Bates DW. How can information technology
improve patient safety and reduce medication errors in childrens health
care? Arch Pediatr Adolesc Med 2001;155:1002-7.
18. Raschke RA, Gollihare B, Wunderlich TA, Guidry JR, Leibowitz AI,
Peirce JC, et al. A computer alert system to prevent injury from adverse
drug events: development and evaluation in a community teaching hospital.
JAMA 1998;280:1317-20.

50 Years Ago in The Journal of Pediatrics

PATHOLOGIC PHYSIOLOGY IN SOME DISTURBANCES OF CARBOHYDRATE METABOLISM
Hartmann A. F. J Pediatr 1955;47:537-70

It required a celestial force with a sense of humor to have this published presentation on hypoglycemia given as the
Clifford Sweet Lecture. In it, Hartmann reviews the state of knowledge of glucose homeostasis and its perturbations.
For me it gave new meaning and context to the notion of a generation gap. Older active physicians were, or nearly
were, contemporaries of an era in which our understanding of glucose balance, glucose mobilization and diabetes was
far more rudimentary. A modern reader might be amused to realize that there was no appreciation for gluconeogenesis
and that all lactate was thought to come from glucose or glycogen. This reader might gloat smugly over their superior con-
temporary understanding. The older reader might wonder about how much of our current knowledge and wisdom might
seem antiquated and nave in 50 years, or even less. Sic transit gloria.
This presentation emphasizes a number of truths completely unspoken today and probably unknown to the modern
pediatrician. A milk-drinking infant has a considerable amount of free galactose in their blood, often accounting to as
much as 40% of the total reducing sugar content of the blood. This information is transparent to the contemporary medical
practice, since we now routinely measure glucose directly, rather than as reducing substances. In this context, finding
galactose in the urine should come as no surprise, since portal circulation removal of galactose is so incomplete. Yet a report
of galactose in urine may today provoke a call to a metabolic specialist. These galactose levels were discovered as a conse-
quence of seeing hypoglycemic symptoms in infants with normal blood sugars, an insight that helps to explain the regard
with which Alexis Hartmann is held in Pediatrics and Metabolism. Here too, the earliest realization that the newborn, and
particularly the premature newborn, was prone to hypoglycemia, presumably due to lactase deficiency, when milk was
introduced after the bowel was rested for severe diarrhea.
Equally dramatic is realization that knowledge of the biochemistry glycogen synthesis, degradation and storage was
as clearly understood then by Carl and Gerti Cori, as it is now. Finally, the absence of any references in the paper is note-
worthy. This is particularly dramatic against the relief of modern practice and belief that the quality of scholarship is
measured by the length of the reference list and eschewing any but the original source.
Stephen Cederbaum, MD
David Geffen School of Medicine at UCLA
Los Angeles, CA 90095-7332
YMPD1923
10.1016/j.jpeds.2005.11.021

Potential Medication Dosing Errors In Outpatient Pediatrics 767