TOXICOLOGY/ORIGINAL CONTRIBUTION
Adverse Drug Events in Emergency Department
From the Department of Emergency
Medicine, OSF Saint Francis Medical
Center,* and the Department of
Biomedical and Therapeutic Sciences,‡
University of Illinois College of
Medicine at Peoria, Peoria, IL.
Author contributions are provided
at the end of this article.
Received for publication
March 26, 2001. Revision received
September 28, 2001. Accepted for
publication October 30, 2001.
Presented in poster form at the
Society for Academic Emergency
Medicine annual meeting, Chicago,
IL, May 1998 (Hafner J Jr, Bucheit K,
Squillante M, et al. Adverse drug
events in emergency department
patients. Acad Emerg Med.
1998;5:528).
Address for reprints: John W.
Hafner, Jr., MD, Department of
Emergency Medicine, OSF Saint
Francis Medical Center, 530 NE
Glen Oak Avenue, Peoria, IL 61637;
309-655-2553, fax 309-655-2602;
E-mail mlhafner@home.com;
jhafner@pol.net.
Copyright © 2002 by the American
College of Emergency Physicians.
0196-0644/2002/$35.00 + 0
47/1/121401
doi:10.1067/mem.2002.121401
2 5 8
Patients
John W. Hafner, Jr., MD* Study objective: Adverse drug events (ADEs) have been
Steven M. Belknap, MD‡
studied in hospitalized patients. Less is known about this com-
Marc D. Squillante, DO*
mon type of injury in emergency department patients. This
Kay A. Bucheit, MD*
study seeks to measure the risks, incidence, severity, and costs
of ADEs in an ED population.
Methods: ED charts of visits to a university-affiliated tertiary-
care ED occurring between March 1 and May 31, 1997, were
retrospectively reviewed. The main outcome measures were
ADE incidence, severity, and total cost. Visits identified by
investigators as containing a suspected ADE were further
assessed by using the Naranjo Adverse Drug Reaction (ADR)
probability scale. Events judged as probable ADEs (Naranjo
ADR probability scale score of >4) were compared with ED con-
trol visits best matched by age for disposition, survival, sever-
ity, payer, sex, race, age, number of drugs, and total cost.
Results: Of 13,602 visits, 13,004 records were available. Three
hundred twenty-one had suspected and 217 had probable ADEs
(1.7% of evaluable encounters); these were compared with vis-
its by 217 age-matched control patients. Insulin and warfarin
were the most commonly responsible drugs. Patients with
ADEs were older (mean age 45.1 versus 36.8 years; mean dif-
ference 8.3; 95% confidence interval [CI] 3.7 to 12.9), were
more often women (odds ratio [OR] 1.48; 95% CI 1.01 to 2.16),
took more drugs (mean number of drugs 4.1 versus 1.9; mean
difference 2.2; 95% CI 1.7 to 2.8), and were hospitalized more
frequently (OR 2.29; 95% CI 1.33 to 3.94) than control patients.
Conclusion: ADEs encompassed an important segment of ED
encounters and annual health care costs. ED screening may
provide useful information about the epidemiology of outpa-
tient ADEs.
[Hafner JW Jr, Belknap SM, Squillante MD, Bucheit KA.
Adverse drug events in emergency department patients. Ann
Emerg Med. March 2002;39:258-267.]
ANNALS OF EMERGENCY MEDICINE 39:3 MARCH 2002
ADVERSE DRUG EVENTS
Hafner et al
INTRODUCTION
The publication of the Institute of Medicine report on
medical errors exposed preventable adverse drug events
(ADEs) as an important and previously underrecognized
cause of medical injury.1 An ADE is an injury (noxious or
harmful effect) resulting from medical intervention
related to a drug.2 Studies of the incidence, severity, and
cost of ADEs in hospitalized patients have found that
between 2.4% and 6.5% of hospitalized patients have an
ADE (>770,000 US hospital patients annually), with
direct costs between US$1.56 and $4.2 billion annually
and an estimated total cost of $12.2 billion in 1996 dol-
lars.2-5
Although much ADE research thus far has focused on
hospital inpatient populations, less is known about emer-
gency department ADEs. Previous reports have noted that
ED patients are at high risk for adverse drug interactions
(ADIs) and that drug-related illness is not uncommon in
the ED.6-17 However, the full scope of ADE-related ED
visits and the factors that put patients at risk for ADEs
remain largely unknown. Inpatient ADEs are most com-
mon in the ICU, perhaps because these patients are sicker
and require more medications.2,4,18-20 EDs also evaluate
many patients with high-severity illnesses who take mul-
tiple medications, and we hypothesize that ADEs are also
common in this population.
The objective of this study was to measure the inci-
dence of ADEs in an ED population, classify the severity
and disposition of ADE visits, quantitate the costs of
ADEs in the ED, and identify risk factors for ADEs within
an ED population.
M AT E R I A L S A N D M E T H O D S
With more than 59,000 annual ED visits and 731 inpa-
tients beds, the study site ED and medical center is the
primary teaching hospital for a university medical school
and serves as both a primary and tertiary referral center
for a surrounding urban and rural population of greater
than 2 million persons. Board-certified attending emer-
gency physicians, emergency resident physicians, and
rotating off-service resident physicians collaboratively
care for a mixed payor population in the ED.
All patients registered at the study site ED between
March 1 and May 31, 1997, were eligible for this case-
control study. Clinicians involved in the direct care of
these patients were unaware of the study. All available ED
charts (physician records, nursing notes, emergency
medical services logs, and discharge instructions) were
MARCH 2002 39:3 ANNALS OF EMERGENCY MEDICINE
retrospectively hand reviewed by either of 2 investigators
(JWH or KAB) to identify any suspected ADEs. Investiga-
tors were trained in predefined data-extraction criteria
and the classification of ADEs before the review. Each
investigator reviewed approximately one half of the avail-
able ED charts in no set order, and data were recorded on
computerized abstraction forms (Microsoft Excel 97,
Microsoft Corporation, Redmond, WA). Ambiguous or
conflicting events were reviewed by a third investigator
(SMB), and conflicts were resolved by consensus. Identi-
fied events included those formally diagnosed by ED staff,
as well as those found by investigators during the chart
review.
Each visit identified as containing a possible ADE was
further evaluated by using the Naranjo Adverse Drug
Reaction (ADR) probability scale score21 (Figure 1) by 2
independent investigators. Visits that were assigned a
total Naranjo ADR probability scale score of greater than
4 were then compared with randomly selected controls
visits. Control visits were randomly selected by using a
hospital database computer from all ED visits occurring
during the study period (excluding the previously identi-
fied patients with ADEs). Control patients’ ages were
directly compared with ages of patients with ADEs, but
for purposes of all additional comparisons, patients with
ADEs were best matched with control patients by age. A
community institutional review board representing the
area medical school and local hospitals approved this
study.
The term “adverse drug event” has become popular in
describing drug-related injury because of the ease in cate-
gorizing events by means of this definition compared with
previous classifications. ADEs encompass all drug-related
injuries that result from medication errors, drug–drug
interactions, or ADRs. An ADR is defined as any noxious
change in a patient’s condition that a physician suspects
may be caused by a drug occurring at dosages normally
used in human patients and that (1) requires treatment,
(2) requires a decrease or cessation of therapy with the
drug, or (3) suggests that future therapy with the drug
carries an unusual risk in this patient.22 A medication
error is any preventable event that may cause or lead to
inappropriate medication use or patient harm (eg, wrong
drug, wrong dose, wrong route, wrong dosing
schedule).23
We studied ADEs rather than simply ADRs because
ADRs exclude medication errors.24 Medication errors are
often difficult to exclude in outpatients because of a lack
of clinical monitoring and scanty documentation. ADEs
involving either prescription or over-the-counter drugs
2 5 9
ADVERSE DRUG EVENTS
Hafner et al

were included. Overdoses of prescribed drugs, whether
intentional or not, were included. Potential ADEs, defined
as incidents with a potential for drug-related injury but
causing no apparent harm, were excluded. Events related
to therapeutic failures (eg, a seizure in a patient who had
stopped anticonvulsant therapy) were not included. Events
caused by nutritional supplements, illicit drugs, nicotine,
or ethanol were not included because of inadequate stan-
dardization, inadequate verification of reported sub-
stances, or both.
Age, sex, insurance status, race, date of the ED visit,
disposition, drug class, and total number of patient drugs
were recorded for both the ADE and control populations.
Patient disposition was coded as released, admitted (ob-
servation and full admission), ICU admission, left against
medical advice or without discharge, or ED death. Insurance
status was recorded as insured, Medicare, Medicaid, or
self-pay. Race was recorded from the admission demo-
graphic database and thus represents the race assigned by
ED personnel rather than that reported by the patient.
(The validity of the medical use of race has been chal-
lenged, but it is commonly recorded.25) ED supporting
diagnoses for both the ADE and control populations
were assigned to 1 of 13 investigator-derived system and
toxidrome–based categories. Multiple ED diagnoses sup-
porting an ADE were accepted, but chronic or clearly
unrelated ED diagnoses were not. International Classifica-
tion of Diseases, 9th Revision, Clinical Modification (ICD-9-
CM) diagnoses of all ED patients seen during the study
period were identified from a hospital database. 26
Similar ICD-9-CM diagnoses were grouped into cate-
gories and ranked according to frequency.
The total number of hospitalized days and hospital
charges were obtained by means of review of the ADE and
control population’s permanent medical records and
billing statements. Charges were converted to costs by
multiplying by hospital-specific, cost-specific ratios of
costs to charges.19
Two investigators (JWH and KAB), each blinded to the
other’s assessment, evaluated every suspected event and
assigned an individual total Naranjo ADR probability
scale score. Investigators were trained in using the
Naranjo ADR probability scale score before study initia-
tion by using a training sample of medical records, and
any general coding conflicts were resolved during
research team meetings by group consensus. In contrast
with other ambiguous methods,27 the Naranjo ADR
probability scale score is inexpensive, simple, and repro-
ducible and has high concordance with the Bayesian
Adverse Reaction Diagnostic Instrument.28 The Naranjo
ADR probability scale score ranges from –4 to 13, with 0
or less considered doubtful, 1 to 4 considered possible, 5
to 8 considered probable, and 9 or greater considered
definite (Figure 1).21,28

Figure 1.
Yes No Unsure Score
Naranjo ADR probability scale. With per-
mission from Naranjo CA, Busto U, Sellers Are there previous conclusive reports on this reaction? +1 0 0
EM, et al. A method for estimating the prob- Did the adverse event appear after the drug was administered? +2 –1 0
ability of adverse drug reactions. Clin
Pharmacol Ther. 1981;30:239-245. Did the adverse reaction improve when the drug was +1 0 0
discontinued or a specific antagonist was given?
Did the adverse reaction appear when the drug was +2 –1 0
readministered?
Are there alternative causes (other than the drug) that –1 +2 0
could have on their own caused the reaction?
Did the drug reaction appear when a placebo was given? –1 +1 0
Was the drug detected in the blood (or other fluids) in +1 0 0
concentrations known to be toxic?
Was the reaction more severe when the dose was +1 0 0
increased or less severe when the dose was decreased?
Did the patient have a similar reaction to the same or +1 0 0
similar drugs in any previous exposure?
Was the adverse event confirmed by any objective evidence? +1 0 0
TOTAL SCORE

2 6 0 ANNALS OF EMERGENCY MEDICINE 39:3 MARCH 2002

ADVERSE DRUG EVENTS
Hafner et al

ADE severity was ranked as fatal, life threatening, sig-
nificant, or insignificant. Events involving patients who
had an altered level of consciousness or significantly
altered vital signs (respiratory rate, pulse rate, blood pres-
sure, and temperature) or who required resuscitation or
acute stabilization, invasive monitoring, or constant one-
to-one nursing were considered life threatening. Events
involving patients who were symptomatic but stable,
required a limited acute intervention, or both were con-
sidered significant ADEs. Other events were considered
insignificant.
ADEs were classified according to their relation to the
chief complaint. A visit to the ED primarily for problems
related to the ADE was considered directly related. A visit
that included multiple complaints and diagnoses, as well
as a symptomatic ADE, was considered moderately re-
lated. A visit in which an ADE was diagnosed incidentally
during an evaluation for a separate problem or in which
an ADE occurred in the ED was considered unrelated.
Statistical calculations were performed with Mathe-
matica 4.0 (Wolfram Research, Champaign, IL) running
on a Macintosh 2300 with MacOS 9.0.4 software (Apple
Computer, Cupertino, CA). Interrater concordance was
calculated with the Cohen κ statistic.29 Nominal vari-
ables (disposition, payer, sex, race, and severity) were
pairwise compared with those of randomly selected con-
trol patients best matched for age by using odds ratios
(ORs); significance and 95% confidence intervals (CIs)
were calculated by using the Tango-McNemars method.30
Payer was paired as drug benefit (insurance and Medi-
caid) versus no drug benefit (Medicare and self-pay).
Race was paired as white versus nonwhite. Disposition
was paired as admitted (hospital and ICU admission) and
released (released, left against medical advice, and ED
death). Ratio variables (number of drugs, age, and cost)
were compared with those of control patients by using the
Welch t test.
R E S U LT S
During the study period, 13,602 ED visits occurred, with
13,004 (95.6%) ED charts available for review and 321
(2.5% of reviewed ED charts) suspected ADE visits iden-
tified. Evaluation of the suspected ADE visit group yielded
217 (1.7% of total reviewed ED charts) visits with a Naranjo
ADR probability scale score of greater than 4; these visits
were classified as probable ADEs and were compared with
control visits. Investigators were highly concordant in
assigning Naranjo ADR probability scale scores to the
ADEs (κ=0.68). Most (68.7%) patients with ADEs were
released from the ED. ADE and control visits were similar
in terms of race and financial class (Table 1). Compared
with the control population, the patients with ADEs were
older, were more likely to be women, and used a higher
mean number of drugs compared with control patients
(Table).
Although the majority of patients were discharged to
home in both groups, significantly more patients with

Table.
Comparison of ADE and control populations.

Patients With ADEs Control Patients Odds Ratio Difference of Means

Characteristic (n=217) (n=217) (95% CI) (ADE Versus Control) (95% CI)

Mean age, y 45.1 (range 0–89) 36.8 (range 0–88) 8.3 (3.7–12.9)
Female sex 132 (60.8%) 111 (51.1%) 1.48 (1.01–2.16)
Race
White 177 (81.6%) 171 (78.8%) 1.23 (0.74–2.05)
Nonwhite 40 (18.4%) 46 (21.2%)
Payer
Drug benefit 134 (61.8%) 142 (65.4%) 1.42 (0.87–2.30)
No drug benefit 83 (38.2%) 75 (34.6%)
Disposition
Released 150 (69.1%) 179 (82.5%) 2.29 (1.33–3.94)
Admitted 67 (30.9%) 38 (17.5%)
Mean total No. of drugs 4.13* (range 1–27) 1.9† (range 0–10) 2.2 (1.7–2.8)
*
Eighty (36.9%) patients with ADEs took >4 drugs.
†
Eighty-eight (40.6%) control patients took no drugs.

MARCH 2002 39:3 ANNALS OF EMERGENCY MEDICINE 2 6 1

ADVERSE DRUG EVENTS
Hafner et al
ADEs were admitted to the hospital (Table 1). Two deaths
occurred during ED care in the control population. There
was no higher severity demonstrated among patients with
ADEs compared with control patients when ICU admis-
sion and patient death were compared with hospital
admission and release from the ED (OR 3.0; 95% CI 0.75
to 13.9).
Thirteen ADEs were caused by ED therapeutic inter-
ventions, representing 6% of total ADEs. Two patients
had ADEs in the ED and were also hospitalized (0.9% of
total ADEs and 1.5% of total ADE admissions). One woman
with dehydration and pyelonephritis had urticaria caused
by ED-administered ciprofloxacin. Another elderly woman
with syncope and transient hypotension had multifocal
atrial tachycardia caused by ED-administered dopamine,
necessitating a non-ICU hospital telemetry admission.
Dystonic reactions represented the most common diag-
nosis for ED-induced ADEs (5 visits or 38.5%).
The most common diagnoses among the patients with
ADEs were hypoglycemia, coagulopathy-hemorrhage,
and rash. The most common diagnoses among the control
patients were chest pain, laceration, and abdominal pain
(Figure 2). The drugs most commonly causing ADEs were
insulin, warfarin, and furosemide (Figure 3). Drug cate-
gories most commonly causing ADEs were antihyper-
glycemics, analgesics, antibiotics, and anticoagulants
(Figure 4). Thirty-nine (95%) of the 41 antihyper-
glycemic ADEs were caused by insulin. One third of the
27 antibiotic ADEs were caused by amoxicillin. Twelve
(44%) of the 27 analgesic ADEs were caused by cyclooxy-
genase inhibitors, and 10 (37%) were caused by opioid
analgesics. All (100%) of the 23 anticoagulant ADEs were
caused by warfarin. ADIs were responsible for 16 (7.4%)
of all ADEs, with mental status changes being the most
common ADI (31.3%).
Two (0.9%) of the ADEs were fatal, 19 (8.8%) were life
threatening, 184 (85%) were significant, and 12 (5.5%)
were insignificant. One patient taking warfarin died from
a retroperitoneal hemorrhage. Another patient taking a
diuretic died from dehydration, pneumonia, and septic
shock. Both deaths occurred during inpatient hospital-
ization. Of the ADEs resulting in hospital admission, 45
(67.2%) were significant, 17 (25.4%) were life threaten-
ing, and 3 (4.5%) were insignificant.
In 81.6% of ADE visits, the ED presentation was directly
related to the identified ADE, whereas 17 (7.8%) were
moderately related, and 23 (10.6%) were unrelated. Of
admitted patients with ADEs, 50 (74.6%) of the hospital
and ICU admissions were primarily for the ADE, 8 (12%)
were moderately related to the ADE, and 9 (13.4%) were
2 6 2
unrelated to the ADE. Fourteen (21%) admissions in-
volved an intentional overdose, and 9 (64.3% of inten-
tional overdose admissions) were admitted primarily for
this ADE. Seven visits were classified as unintentional
overdoses.
Of all the ED visits occurring during the study, the
most frequent ICD-9-CM–coded diagnoses were lacera-
tion-abrasions (7.6%), chest pain not otherwise specified
(4.4%), and abdominal pain of undetermined cause
(4.2%). Comparison of the ED ICD-9-CM codes between
the study center and the 1997 US average reveals similar
top ED diagnoses, although differing in rank and stratifi-
cation.31 Analysis of the ADE incidence compared with
ICD-9-CM code frequency indicates that ADEs collec-
tively were as common as the 14th most prevalent ED
diagnosis during the study period and as common as the
15th most prevalent ED diagnosis nationally in 1997.31
ADEs caused more study-site ED visits than pneumonia,
pharyngitis, syncope, or congestive heart failure.
The mean cost of ADE visits was higher than the mean
cost of control visits ($1,764 versus $1,133) but was not
statistically significant (difference in mean costs $631.17;
95% CI $–196.47 to $1,458.80). From annualization of
our 3 months of data, ED ADEs cost the study center $1.63
million in 1997. ADE costs varied by disposition. Patients
discharged to home were less costly on average than hos-
pitalized patients ($247.68 [range $20.40-$2,989.46]
versus $5,162.19 [range $91.27 to $43,312.96]).
DISCUSSION
ADEs comprise an important proportion of ED visits,
ranking higher than common ED diagnoses, such as syn-
cope, pneumonia, and pharyngitis. ADEs usually occurred
before ED evaluation, and most patients presented pri-
marily for ADE symptomatology. Few identified ADEs
were fatal or life threatening, but most were considered
significant. Compared with a control population, patients
with ADEs were older, took more medications, and were
more frequently hospitalized. ADE visits resulted in a
diagnosis most often of a metabolic, hematologic, or gas-
trointestinal nature, whereas musculoskeletal, miscella-
neous, and gastrointestinal diagnoses were common in
the control visits.
Prince et al10 reported that 2.9% of ED visits were
caused by drug–related illness. However, their definition
of drug–related illness included drug abuse, toxicity, drug
interactions, and suboptimal medication taking. Ethanol,
cocaine, and heroin represented 23% of all drug–related
illness and caused more drug–related ED illness than
ANNALS OF EMERGENCY MEDICINE 39:3 MARCH 2002
ADVERSE DRUG EVENTS
Hafner et al

therapeutic drugs. Schneitman-McIntire et al14 found
1.7% of ED visits to a Health Maintenance Organization–
based hospital were caused by “medication misadven-
tures,” defined as unfavorable medication effects, poor
compliance, inappropriate self-medication, inappropri-
ate prescribing, and drug interactions. Others have re-
ported a drug-related ED visit incidence of between 0.86%
and 3.9% but have also used a variety of classifications for
drug-related illness.15-17
Raschetti et al13 prospectively analyzed 5,497 ED
patients and documented an annual ADE incidence of
4.3%. However, they considered suboptimal medication
use to be an ADE; 31% of their recorded ADEs were thera-
peutic failures caused by suboptimal medication taking,
representing the main cause of drug-related hospital
admissions. Substance abuse, suboptimal medication
taking, and ADEs are distinct problems resulting from
different pathologic, sociologic, and health care system

Figure 2. Diagnoses of Patients With ADEs Diagnoses of Control Patients

Diagnoses for patients with
ADEs and control patients. 0 5 10 15 20 25 30 35 0 5 10 0
Hypoglycemia Chest pain Scabies
Coagulopathy or hemorrhage Laceration Ruptured viscus
Rash Abdominal pain Renal colic
Mental status changes Neck strain Rash
Hypokalemia Fracture Positive culture
Nausea or vomiting Contusion Pharyngitis
Dystonic reaction Pneumonia Pelvic inflammatory disease
Neutropenia Muscle strain Ocular foreign body
Anticholinergic syndrome Upper respiratory tract infection Nausea or vomiting
Antibiotic colitis Otitis media Motor vehicle crash
Respiratory depression Headache Leg pain
Fever Gastroenteritis Knee strain
Tachycardia Urinary tract infection Joint effusion
Hypotension Ethanol intoxication Intradermal inclusion cyst
Hyperkalemia Cellulitis Inguinal hernia
Gingival hyperplasia Cardiac arrhythmia Hypotension
Gastritis Back strain Hyponatremia
Dehydration Asthma HIV/pneumonia
Constipation Viral illness Hemorrhoid
Bradycardia Sprain Hemorrhage
Anaphylaxis Hematuria Gastrointestinal hemorrhage
Abdominal pain Bronchitis Furuncle
Urinary retention Thoracic strain Foreign-body removal
Tinnitus Streptococcal pharyngitis Foreign-body ingestion
Respiratory distress Seizure Fever
Psychosis Nasal fracture Esophagitis
Pruritis Mental status changes Drug screen
Pancytopenia Incomplete abortion Dog bite
Myalgia Hypoxia Deep venous thrombosis
Mental status change Gastritis Crohns disease
Local edema Conjunctivitis Constipation
Hyponatremia Cardiac arrest Congestive heart failure
Headache Anxiety Colitis
Gynecomastia Well-child examination Coagulopathy
Edema Weakness Chronic pain
Dysrhythmia Warts Chest-wall pain
Dizziness Vertigo Cervical strain
Congestive heart failure Urinary incontinence Catheter placement
Cardiac arrest Ultraviolet keratitis Carpal tunnel syndrome
Candidiasis Thrombophlebitis Burn
Ataxia Threatened abortion Bartholin’s gland cyst
Arrhythmia Syncope Antibiotic administration
Angioedema Suture removal Ankle sprain
Stroke Animal bite
Small-bowel obstruction Anal fissure
Sinusitis Alcoholism
Sepsis Acute coronary syndrome
Schizophrenia

MARCH 2002 39:3 ANNALS OF EMERGENCY MEDICINE 2 6 3

ADVERSE DRUG EVENTS
Hafner et al

processes and are best addressed separately. In our defini-
tion of ADE, we excluded substance abuse and subopti-
mal medication taking. We considered intentional over-
doses of therapeutic drugs to be ADEs because the intent
of the prescribing physician was to make a medical inter-
vention related to the use of a drug. We classified inten-
tional overdoses as a medication error, specifically as a
patient-initiated wrong-dose error, although we acknowl-
edge that the intent of the patient was not to make a medi-
cal intervention. Previous studies of inpatient admissions
report ADE incidences of 2.43 to 6.5 per 100 admis-
sions.2,4,19 The lower ADE incidence in our ED patients
likely reflects lower severity of illness, fewer drugs, and
less comprehensive charting than in hospitalized patients.
Much ED drug morbidity research has focused on
ADIs. Potential ADIs have been recorded with the assis-
tance of ED-based pharmacists and computer programs
in 13% to 47% of ED patients, but few were clinically sig-
nificant.7-9,11 In our population, ADIs were responsible
for 7.4% of ADEs compared with the 0.3% to 4% reported
previously.12,13 Medication errors are a common cause of
hospital ADEs.2,3,20,30,31 Unintentional overdose did
occur in 3.4% of our study, but this likely underestimates
the medication error rate in the ED population because
other types of medication errors are difficult to detect in
this setting.
Evaluation of ICD-9-CM codes of all study site ED
patients seen during the study period suggests that many

Figure 3. 0 5 10 15 20 25 30 35 0
Drugs that caused ADEs. DPT, Insulin Nalbuphine
Diptheria pertussis and tetanus Warfarin Morphine sulfate (extended release)
immunization; OPV, oral polio Furosemide Metoclopramide
vaccine. Chemotherapy Methadone
Digoxin Mepivicaine
Amoxicillin Medroxyprogesterone acetate (Aepot)
Prochlorperazine Lisinopril/hydrochlorothiazide
Ibuprofen Ketoprofen
Diphenhydramine Ketamine
Risperidone Indapamide
Phenytoin Hydroxychloroquine sulfate
Hydrochlorothiazide Hydrocodone bitartrate/acetaminophen
Metronidazole Hydrochlorothiazide/triamterene
Codeine/acetaminophen Hepatitis vaccine
Cefaclor Haloperidol
Vancomycin Glycerine
Prednisone Fluvoxamine maleate
Medroxyprogesterone acetate (Depot) Fentanyl patch
Glyburide Fenfluramine/phentermine
DPT and OPV vaccines Erythromycin
Clonidine DPT vaccine
Carbamazepine Doxycycline
Azithromycin Dopamine
Amoxicillin/clavulanate Dirithromycin
Amitriptyline Dimenhydrinate
Zolpidem Diltiazem
Trimethoprim/sulfamethaxosole Diclofenac potassium
Tramadol Dexamethasone
Ticlopidine Desogestrel/ethinyl estradiol
Theophylline Conjugated estrogens
Tetanus toxoid Codeine
Temazepam Ciprofloxacin
Sertraline Chlorambucil
Quinapril Cephalexin
Propranolol Cefprozil
Propoxyphene napsylate/acetaminophen Benzocaine topical gel
Potassium chloride Atenolol
Plasma (fresh frozen) Aspirin/butalbital/caffeine
Phenylephrine hydrochloride/guaifenesin (extended release) Aspirin
Phenobarbital Amlodipine
Penicillin Amitriptyline/perphenazine
Nefazodone hydrochloride Acetaminophen
Naproxen

2 6 4 ANNALS OF EMERGENCY MEDICINE 39:3 MARCH 2002

ADVERSE DRUG EVENTS
Hafner et al

ADEs are coded under alternate codes (ie, an ED visit are correct and generalized to all 1997 US ED patients,
receiving an ICD-9-CM code of gastritis rather than ery- $2.85 billion was spent on management of ADEs, with
thromycin–induced gastritis). Although an ADE may be $276 million for patients discharged to home and $2.58
the primary cause for the ED visit, most ICD-9-CM codes billion for patients admitted to the hospital.
simply reflect the condition treated without reference to Recorded ED diagnoses represented a wide range of
the ADE. Although it may be an oversimplification to disorders distributed among 12 of 13 system and tox-
directly compare our ADE incidence with reported ICD- idrome–based categories. ADE visits appeared to present
9-CM diagnosis incidence, it nonetheless appears that a different spectrum of disease compared with the control
ADEs represent the primary causes of various ED diag- visits. Although ADE visits contained a large number of
noses. metabolic diagnoses, control visits were often caused by
In 1997, the National Center for Health Statistics re- musculoskeletal disorders. Gastrointestinal diagnoses
ported an estimated 94.9 million annual ED visits, or 35.6 were common to both populations. This differs from the
visits per 100 persons.33 By extrapolating these data to pattern in inpatient studies, in which central nervous sys-
the proportion of study visits in which ED presentation tem ADEs were more common than metabolic or gastro-
was directly related to an ADE and that received a Naranjo intestinal ADEs.19,34,35
ADR probability scale score of greater than 4, we estimate In our study, patients with ADEs were an older and
that 1.58 million US patients were evaluated for an ADE more often female population that used more medica-
during 1997. On the basis of our disposition proportions, tions compared with control patients. Hanlon et al,36 in a
we project that 1.1 million of the ED patients were randomized cohort study, documented that 35% of
released from the ED and 474,000 were admitted to the ambulatory patients older than 65 years had experienced
hospital (54,000 to an ICU). If our mean total ADE costs an ADE during a 1-year period. Although not specifically
examined, these authors suggest that the high average
number of daily drugs contributed to the increased num-
ber of ADEs. In the study by Carbonin et al34 of risk fac-
Figure 4. tors for ADRs, the incidence was highest between ages 70
Drug categories that caused ADEs. and 79 years (6.5%). However, after multivariate logistic
regression analysis, advanced age was not an independent
predictor of ADR, although taking 4 or more drugs, lengthy
0 5 10 15 20 25 30 35 40
Antihyperglycemics hospital stay, and multiple medical problems were demon-
Antibiotics strated to be predictive of ADRs. In our study, ED patients
Analgesics
Anticoagulants with ADEs were taking nearly twice as many drugs as con-
Diuretics trol patients, which is consistent with results of previous
Antineoplastics studies.5,10,11,18,34,37-41 Previous studies, in addition to
Cardiac inotropes
Antinauseants ours, have reported that ADEs are more common in
Anticonvulsants women.14,15,35,42,43 However, some authors have sug-
Antipsychotics
Antihistamines gested that this may simply be the consequence of women
Antidepressants taking more drugs.34,35 We found that among the patients
Vaccines with ADEs, women took an average of 3.7 drugs compared
β-Adrenergic blockers
Contraceptives with 4.0 drugs taken by men, suggesting that polyphar-
Steroids (systemic) macy alone does not account for the greater incidence of
Hypnotics
Calcium channel blockers ADEs in women.
Platelet inhibitors Several limitations are inherent to this study. The study
Mucolytics
Methylxanthines
group was identified by a retrospective review of ED
Laxatives (osmotic) charts, and the recorded medical record restricts ex-
Hormones tractable data. Documentation of ADEs by clinicians
Electrolytes
Catecholamines likely underestimates ADE incidence.44-48 Although we
Blood products focused on ADEs that were clearly documented, we
Antirheumatics
Anorexiants included events that were present on chart review but not
Angiotensin-converting enzyme inhibitors necessarily diagnosed or coded by ED personnel. It is
possible, despite evaluating possible events with a vali-

MARCH 2002 39:3 ANNALS OF EMERGENCY MEDICINE 2 6 5

ADVERSE DRUG EVENTS
Hafner et al
dated ADE scoring instrument, that some incidents were
a result of causes other than ADEs. However, regardless of
these restrictions, these incidence measures are probably
conservative. Approximately 5% of charts were unavail-
able for review, but we are aware of no reason why these
charts would be distinctly different from those sampled.
Our control group was randomly selected from ED visits
occurring during the same study period and then best
matched by patient age to ADE visits. Although this strat-
egy allowed for a direct comparison of the ADE and con-
trol group’s ages, it limited how precisely the control visits
could be matched to ADE visits by age, possibly introduc-
ing bias. An alternative approach would use 2 separate
control groups, one for direct age comparison and the
other for all additional comparisons. Additionally, our
definition of ADE included drug overdoses and excluded
therapeutic failures, differing from some previously re-
ported classifications.
Although our ED is both a community and a tertiary-
care facility representing a wide mix of financial classes,
these results may not be applicable to the strictly rural or
inner-city ED. Our reported ADE incidence, cost data,
and extrapolation to national statistics must be taken in
context. Our study population was limited to a single ED,
and the ADE incidence is based on the particular combi-
nation of patients presenting to our institution. Future
prospective studies should sample ADEs from a larger
cross-section of EDs and urgent care facilities in a com-
munity so as to avoid systematic biases caused by varia-
tion in patient mix at different institutions.49
Monitoring ED patients for ADEs may be helpful for
future medication-injury research. One study of 1,000
patients in an office-based general internal medicine
practice found 42 (0.42%) ADEs, of which 23 (0.23%)
were considered preventable.50 Our study found 217
(1.7%) ADEs in 13,004 ED patients, an ADE capture rate
fourfold higher than that of the office-based practice study.
Although we acknowledge that general office patients are
a different population than ED patients, screening ED
patients for ADEs may be a more effective community
ADE-monitoring strategy than sampling office-based
practices. Furthermore, the severity of ADEs is likely to
be higher in an ED, possibly reflecting more clinically
important events. For example, ongoing screening of
ADEs occurring in a diverse sample of EDs might shorten
the delay between introduction of a new drug and recog-
nition of serious adverse effects of that drug unrecognized
in initial clinical trials. Current methods of capturing out-
patient ADEs rely on spontaneous medical provider
reporting and grossly underestimate the actual ADE rate.
2 6 6
Alternatively, these methods might be helpful at the local
level, providing guidance to the local medical commu-
nity’s health care improvement efforts by identifying
which preventable ADEs are most common. We believe
that our results justify further study of ED ADE screening
as a source of information for reducing the rate of outpa-
tient ADEs and medication errors.
In summary, ADEs are not uncommon in the ED and
may represent substantial annual costs. The majority of
ADEs in the ED are significant, represent the presenting
chief complaint of the patient, and are more common in
older patients and in those taking multiple drugs. Com-
munity ED–based screening may have promise as a means
of monitoring outpatient ADEs.
Author contributions: SMB and MDS conceived the study and SMB, MDS, and JWH devel-
oped its design. JWH and KAB acquired the data. SMB and JWH managed the data, and SMB
oversaw quality control. SMB, JWH, and MDS analyzed and interpreted the data and SMB
provided statistical advice. JWH drafted the manuscript, and SMB and MDS contributed sub-
stantially to its revision. JWH and SMB take responsibility for the paper as a whole.
REFERENCES
1. Kohn LT, Corrigan JM, Donaldson M, eds. To Err Is Human: Building a Safer Health System.
Washington, DC: Institute of Medicine; 1999.
2. Bates DW, Cullen DJ, Laird N, et al. Incidence of adverse drug events and potential adverse
drug events: implications for prevention. JAMA. 1995;274:29-34.
3. Leape LL, Brennan TA, Laird N, et al. The nature of adverse events in hospitalized patients:
results of the Harvard Medical Practice Study II. N Engl J Med. 1991;324:377-384.
4. Classen DC, Pestotnik MS, Evans RS, et al. Adverse drug events in hospitalized patients:
excessive length of stay, extra costs, and attributable mortality. JAMA. 1997;277:301-306.
5. Thomas EJ, Studdert DM, Newhouse JP, et al. Costs of medical injuries in Utah and Colorado.
Inquiry. 1999;36:255-264.
6. Smith KM, McAdams JW, Frenia ML, et al. Drug-related problems in emergency department
patients. Am J Health Syst Pharm. 1997;54:295-298.
7. Karas S. The potential for drug interactions. Ann Emerg Med. 1981;10:627-630.
8. Beers MH, Storrie M, Lee G. Potential adverse drug interactions in the emergency room. An
issue in the quality of care. Ann Intern Med. 1990;112:61-64.
9. Herr RD, Caravati EM, Tyler LS, et al. Prospective evaluation of adverse drug interactions in
the emergency department. Ann Emerg Med. 1992;21:1331-1336.
10. Prince BS, Goetz CM, Rihn TL, et al. Drug-related emergency department visits and hospital
admissions. Am J Hosp Pharm. 1992;49:1696-1700.
11. Goldberg RM, Mabee J, Mammone M, et al. A comparison of drug interaction software
programs: applicability to the emergency department. Ann Emerg Med. 1994;24:619-625.
12. Goldberg RM, Mabee J, Chan L, et al. Drug-drug and drug-disease interactions in the ED:
analysis of a high-risk population. Am J Emerg Med. 1996;14:447-450.
13. Raschetti R, Morgutti M, Menniti-Ippolito F, et al. Suspected adverse drug events requiring
emergency department visits or hospital admissions. Eur J Clin Pharmacol. 1999;54:959-963.
14. Schneitman-McIntire O, Farnen TA, Gordon N, et al. Medication misadventures resulting in
emergency department visits at an HMO medical center. Am J Health Syst Pharm.
1996;53:1416-1422.
15. Stoukides CS, D’Agostino PR, Kaufman MB. Adverse drug reaction surveillance in an emer-
gency room. Am J Hosp Pharm. 1993;50:712-714.
16. Aparasu RR. Drug-related-injury visits to hospital emergency departments. Am J Health
Syst Pharm. 1998;55:1158-1161.
ANNALS OF EMERGENCY MEDICINE 39:3 MARCH 2002
ADVERSE DRUG EVENTS
Hafner et al

17. Dennehy CE, Kishi DT, Louie C. Drug-related illness in emergency department patients. Am 45. Keith MR, Bellanger-McCleery RA, Fuchs JE. Multidisciplinary program for detecting and
J Health Syst Pharm. 1996;53:1422-1426. evaluating adverse drug reactions. Am J Hosp Pharm. 1989;46:1809-1818.
18. Cullen DJ, Sweitzer BJ, Bates DW, et al. Preventable adverse drug events in hospitalized 46. Classen DC, Pestotnik SL, Evans RS, et al. Computerized surveillance of adverse drug
patients: a comparative study of intensive care and general care units. Crit Care Med. events in hospitalized patients. JAMA. 1991;266:2847-2851.
1997;25:1289-1297. 47. Cullen DJ, Bates DJ, Small SD, et al. The incident reporting system does not detect adverse
19. Bates DW, Spell N, Cullen DJ, et al. The costs of adverse drug events in hospitalized patients. drug events: a problem for quality improvement. J Qual Improvement. 1995;21:541-548.
JAMA. 1997;277:307-311. 48. Figueiras A, Tato F, Fontainas J, et al. Influence of physicians’ attitudes on reporting
20. Bates DW, Leape LL, Petrycki S. Incidence and preventability of adverse drug events in hos- adverse drug events: a case-control study. Med Care. 1999;37:809-814.
pitalized patients. J Gen Intern Med. 1993;8:289-294. 49. United States General Accounting Office, Health, Education, and Human Services Division.
21. Naranjo CA, Busto U, Sellers EM, et al. A method for estimating the probability of adverse GAO Report to Congressional Requesters. Washington, DC: United States General Accounting
drug reactions. Clin Pharmacol Ther. 1981;30:239-245. Office, Health, Education, and Human Services Division; 2000. Publication No. GAO/HEHS-00-21.
22. Koch-Weser J. Definition and classification of adverse drug reactions. Drug Information 50. Burnum JF. Preventability of adverse drug reactions [letter]. Ann Intern Med. 1976;85:80-81.
Bull. 1968;2:72-78.
23. National Coordinating Council for Medication Error Reporting and Prevention. About medi-
cation errors. Available at: www.nccmerp.org. Accessed January 14, 2001.
24. Bates DW, Boyle DL, Vander Vliet MB, et al. Relationship between medication errors and
adverse drug events. J Gen Intern Med. 1995;10:199-205.
25. Witzig R. The medicalization of race: scientific legitimization of a flawed social construct.
Ann Intern Med. 1996;125:675-679.
26. Public Health Service and Health Care Financing Administration. International Classification
of Diseases, 9th Revision, Clinical Modification. Washington, DC: Public Health Service; 1991.
27. Koch-Weser J, Sellers EM, Zacest R. The ambiguity of adverse drug reactions. Eur J Clin
Pharmacol. 1977;11:75-78.
28. Lanctôt KL, Naranjo CA. Comparison of the Bayesian approach and a simple algorithm for
assessment of adverse drug events. Clin Pharmacol Ther. 1995;58:692-698.
29. Cohen J. A coefficient of agreement for nominal scales. Educ Psychol Meas. 1960;20:37-46.
30. Tango T. Equivalence test and confidence interval for the difference in proportions for the
paired-sample design. Stat Med. 1998;17:891-908.
31. Lesar TS, Lomaestro BM, Pohl H. Medication-prescribing errors in a teaching hospital: a 9-
year experience. Arch Intern Med. 1997;157:1569-1576.
32. Lesar TS, Briceland L, Stein DS. Factors related to errors in medication prescribing. JAMA.
1997;277:312-317.
33. Nourjah P. National Hospital Ambulatory Medical Care Survey: 1997 Emergency Depart-
ment Summary. Advance data from vital and health statistics; no. 304. Hyattsville, MD: National
Center for Health Statistics; 1999.
34. Carbonin P, Pahor M, Bernabei R, et al. Is age an independent risk factor of adverse drug
reactions in hospitalized medical patients? J Am Geriatr Soc. 1991;39:1093-1099.
35. Domecq C, Naranjo A, Ruiz I, et al. Sex-related variations in the frequency and characteris-
tics of adverse drug reactions. Int J Clin Pharmacol. 1980;18:362-366.
36. Hanlon JT, Schmader KE, Koronkowski MJ, et al. Adverse drug events in high risk older
outpatients. J Am Geriatr Soc. 1997;45:945-948.
37. Nolan L, O’Malley K. Prescribing for the elderly, part I: sensitivity of the elderly to adverse
drug reactions. J Am Geriatr Soc. 1988;36:142-150.
38. Montamat SC, Cusak B. Overcoming problems with polypharmacy and drug misuse. Clin
Geriatr Med. 1992;8:143-158.
39. Stewart RB, Cooper JW. Polypharmacy in the aged: practical solutions. Drugs Aging.
1994;4:449-461.
40. Nelson KM, Talbert RL. Drug-related hospital admissions. Pharmacotherapy. 1996;16:701-707.
41. Bigby JA, Dunn J, Goldman L, et al. Assessing the preventability of emergency hospital
admissions: a method for evaluating the quality of medical care in a primary care facility. Am J
Med. 1987;83:1031-1086.
42. Caranasos G, Stewart RB, Cluff LE. Drug-induced illness leading to hospitalization. JAMA.
1974;228:713-717.
43. Hurwitz N, Wade OL. Factors predisposing to adverse drug reactions. BMJ. 1969;1:536-539.
44. Reinstein PF, Robinson JI. Annual adverse drug experience report: 1994. Rockville, MD:
Division of Epidemiology and Surveillance, Food and Drug Administration, US Department of
Health and Human Services; undated.

MARCH 2002 39:3 ANNALS OF EMERGENCY MEDICINE 2 6 7