The ability of a hospital's adverse drug reaction (ADR) database to identify common
and repeated patterns of preventable adverse drug events (ADEs) was analyzed.
ADR reports collected from 1994 through 2000 were extracted from a teaching hospital's
ADR database. Reports were assessed concurrently in accordance with seven previously
published explicit criteria for preventability. Only cases considered clinically
significant were included in this analysis. Events that occurred in the ambulatory
care setting were excluded. Preventable ADEs were categorized by drug or drug
class, type of medication error, and the subsequent adverse outcome. Novel in
this analysis was the linking of these three descriptors.
Of the 2571 ADR reports assessed, 415 ADEs were deemed preventable. Of the preventable
ADEs, 98 were not analyzed because they occurred in the ambulatory care setting,
leaving 317 preventable ADEs in 275 inpatients (mean age ± S.D., 48.5 ±
23.9 years) for analysis. Although 93 drugs were associated with these ADEs, only
10 drugs accounted for more than 60% of the events. Analysis and categorization
by type of error and outcome suggested that three high-priority preventable ADEs
accounted for 50% of all reports: (1) overdoses of anti-coagulants or insufficient
monitoring and adjustments (according to laboratory test values) were associated
with hemorrhagic events, (2) overdosing or failure to adjust for drug drug interactions
of opiate agonists was associated with somnolence and respiratory depression,
and (3) inappropriate dosing or insufficient monitoring of insulins was associated
Analysis of a hospital ADR database identified prevalent and preventable clinically
The widely recognized Institute of Medicine (IOM) report on patient safety has
challenged medical facilities to systematically investigate and reduce medical
errors. It recognizes that preventable adverse drug events (ADEs)
are some of the most common consequences of medical errors. Inappropriate drug
therapy and medication errors are responsible for at least 3% and as many as 9%
of hospital admissions (median, 4.3%) and result in preventable ADEs in up to
4% of patients during their hospital stay.
The IOM mandate for a reduction of medical errors is reflected in new patient
safety standards established by the Joint Commission on Accreditation of Healthcare
Organizations (JCAHO). These standards require institutions to ensure
that the actual performance of processes identified as "error-prone" or "high-risk"
regarding patient safety are measured and analyzed and, when significant variation
is identified, appropriate corrective actions are taken to enhance the system
Determining institution-specific high-risk or high-priority areas for preventable
ADEs can be difficult for health care organizations. Preventable ADEs can result
from errors involving a large number of processes related to patient care, involve
various drugs, and manifest in a variety of adverse patient outcomes. A study
by Bates and colleagues found that no single drug accounted for
more than 9% of all ADEs. These investigators concluded that interventions for
preventing drug-related injuries must target many drugs to have a major impact
on the overall number of events. Another study by Hallas et al.
stated that a major concern is the diversity of preventable ADEs, as "a high degree
of intervention will be required to produce a measurable change."
Published studies that involved a systematic analysis of ADEs have lacked the
necessary detail to describe the nature and frequency of specific preventable
ADEs. For example, many studies discuss only major drug classes that caused preventable
ADEs (e.g., cardiovascular drugs) or major organ systems affected (e.g., gastrointestinal
problems). Other means to find more specific information are sentinel event alerts,
as published by the Institute of Safe Medication Practices and JCAHO, but these
events may be rare or may not pose a significant problem for an individual institution.
Thus, by focusing improvement strategies on sentinel event alerts alone, health
systems might neglect a considerable number of more prevalent, more severe, or
more costly preventable ADEs. Medication safety officers or committees in health
systems face a very difficult task in describing and selecting high-risk areas
for preventable ADEs that could be targeted for most efficient quality improvement
Spontaneous reporting of adverse drug reactions (ADRs) and, more recently, medication
errors provides an opportunity for institution-specific analyses. Although spontaneous
ADR reporting is not comprehensive and may be biased, it probably generates the
most common and comprehensive database on ADEs.[6, 7] FDA has promoted
the voluntary reporting of ADRs since the 1960s. In the 1970s, the
regulatory mandate of JCAHO to define and review all significant ADRs made ADR
reporting mandatory in the hospital setting.
The objectives of JCAHO's and FDA's reviews of ADRs are fundamentally different.
FDA is primarily interested in serious reactions, which are often idiosyncratic
(unpreventable), whereas JCAHO stresses quality improvement. Therefore, monitoring
potentially preventable ADEs is more the purview of the JCAHO mandate. Consequently,
many medical care facilities began emphasizing preventability as a key quality
indicator for ADRs in the 1990s. As a result, current ADR databases
not only include idiosyncratic reactions (i.e., adverse reactions caused by medications
that were appropriately given) but also ADEs following medication errors or inappropriate
care, such as drug overdoses, contraindications, or insufficient monitoring.
Previous studies have analyzed hospital ADR databases to estimate ADR frequency
and costs and to identify high-risk drugs; however, these studies did not focus
on preventable events[11, 12] or did not provide sufficient detail
to identify highly prevalent preventable ADEs.[13, 14] For example,
while published studies may report major drug classes and major organ systems
affected, rarely do such reports link these two components or attempt to describe
the associated errors that caused the event. We used a novel approach to analyze
preventable ADE reports by linking types of drugs, types of associated adverse
outcomes, and types of errors to organize ostensibly unrelated and disparate preventable
This study explored whether a systematic analysis of data extracted from an existing
ADR database could help identify common and repeated patterns of preventable ADEs
during a patient's hospital stay. The specific objectives were to develop a systematic
method for analyzing ADR reports that could be adopted by other institutions and
identify high-priority areas of preventable ADEs in health systems.
This is a descriptive analysis of preventable ADE reports accumulated in an ADR
database at an academic teaching hospital in Florida. The reporting hospital is
a universitybased, tertiary referral center with approximately 25,000-28,000 inpatient
admissions annually that specializes in oncology, cardiovascular medicine and
surgery, neurologic medicine and surgery, pediatrics, and transplantation.
The ADR database was developed in 1988 to enable systematic analyses of reported
ADRs. The database includes ADRs collected at the institution's
drug information center from spontaneous reports by health care professionals
and from a tracer drug search (i.e., screening of pharmacy medication databases
for orders of antidotes or drugs commonly used to treat ADRs). Spontaneous reports
are submitted on an institution-specific ADR report form, which is available online,
or via telephone hotline. Most of the spontaneous reports (>92%) were submitted
by pharmacists. The pharmacy also has an active tracer drug program with which
the pharmacy medication database is daily screened for patients who received specific
drugs that are frequently used to treat ADRs and drug toxicities.[15, 16]
Whether detected patients truly had an ADR is verified by medical chart review,
usually conducted concurrently by pharmacy students and supervised by pharmacists
or pharmacy residents. The tracer drug list includes antidotes (e.g., naloxone,
phytonadione) and drugs used to treat allergies, hypotension, and hypoglycemia
Each report is reviewed by a drug information pharmacist, who determines whether
the ADR was reported to FDA, classifies each report by type of
drug and organ system affected, and assesses causality with a modification of
the Jones algorithm. Since January 1994, each ADR has been further
categorized as to whether it was potentially preventable with a modification of
the criteria published by Schumock and Thornton. The explicit descriptions
of the modified criteria and their frequencies for all analyzed ADRs are listed
in Table 1.
The database includes all variables required by FDA's MedWatch program.
Additional information collected includes the area where the error occurred, the
medical service responsible for the error, profession of the reporter (e.g., nurse,
pharmacist, physician), and the reason why the reaction was reported.
Because the database does not formally differentiate between ADEs that resulted
from ambulatory care versus inpatient treatment, two strategies were used in the
present study to exclude ADEs that occurred before or led to hospitalization.
The full-text description of each report was searched for the term "ED" or "emergency,"
and respective cases were reviewed by the authors. ADEs occurring in the ambulatory
care setting were excluded from the analyses. Further, if the reports indicated
that the ADR was the reason for admission, the case was excluded. A new variable,
"inhouse (Y/N)," was added to the database in 2001 to facilitate future analyses.
Only cases considered clinically significant were included in the analyses. Relevant
cases included the following variables: patient died (3 cases), reaction was permanently
disabling (2 cases), reaction was life-threatening (34 cases), reaction prolonged
hospitalization (42 cases), reaction required change in drug therapy (123 cases),
and reaction required additional therapeutic intervention (252 cases). Drug serum
concentrations beyond the target range and other associated abnormal laboratory
test results were considered clinically significant because they represented an
increased risk for patient injury. For example, an International Normalized Ratio
(INR) of 6 may in itself not indicate an adverse event, but it indicates a patient's
high propensity for bleeding. For all other cases, adverse outcomes were associated
with significant symptoms or complications.
To develop a systematic method for analyzing the database cases, the available
data elements and theoretical work that provided definitions for preventable ADEs
were reviewed. ADEs have been defined as injuries related to the use or nonuse
of medications. They are considered preventable when they are preceded
by an inappropriate pattern of care or medication error that could have been detected,
predicted, controlled, and avoided. Using these definitions, we
chose the following unique descriptors of a preventable ADE to categorize the
extracted reports: (1) the type of drug or drug class involved in the preventable
ADE, (2) the type of medication error or inappropriate care that caused the preventable
ADE, and (3) the type of adverse outcome or patient injury associated with the
medication error. The designation of "types of medication errors" was based on
the preventability criteria listed in Table 1. Medications
were organized into drug classes by using a modification of the American Hospital
Formulary Service (AHFS) code. If a drug class was represented
by only one drug, the specific drug was reported. This study was approved by the
institution's investigational review board.
Of 2571 ADR reports collected over the six-year study period, 415 (16.1%) ADRs
were considered potentially preventable; 98 cases were excluded because the ADR
occurred in the ambulatory care setting and led to a hospital admission. Therefore,
317 reports of preventable ADEs in 275 patients were analyzed. The mean ±
S.D. age of the patients who had these preventable ADEs was 48.5 ± 23.9 years
(median, 54 years), and 52% were female.
Analysis by drug class. A total of 93 different drugs were associated with
the 317 preventable ADEs reviewed; however, 10 drugs accounted for more than 60%
of all reports. They included the anticoagulants warfarin (28.7% of reports) and
heparin (3.8%), the opiate agonists morphine (8.5%) and meperidine (1.9%), insulin
(6.9%), midazolam (3.5%), digoxin (2.2%), phenytoin (2.2%), cyclosporine (1.6%),
and promethazine (1.6%).
Almost three quarters of all reports were accounted for by 10 major drug classes
(Table 2). Organizing drugs into major drug classes did
not affect the high-priority areas: anticoagulants, opiate agonists, insulins,
benzodiazepines, hydantoins, and digoxin continued to cause the most problems.
Excessive anticoagulation or hemorrhagic events accounted for one third of all
preventable ADEs, and another third was described by central nervous system problems,
such as excessive sedation or respiratory depression. Reports associated with
inappropriate use of anticoagulants were the only cases that frequently listed
abnormal laboratory values (elevated activated partial thromboplastin time [aPTT]
or INR values) without other clinical manifestations. However, even when minor
events (e.g., echymosis, skin lesions) and major hemorrhagic events (e.g., gastric
or intracranial hemorrhage) were considered separately, anticoagulants still remained
within the top 10 problem areas.
ADR tracer program. Almost two thirds (62.5%) of all ADRs reviewed were
identified by the ADR tracer program. The remainder of cases were generated by
spontaneous reports to the drug information center. Even though the proportion
of preventable ADEs that could be explained by the top 10 drug classes became
smaller when the reactions identified by the tracer program were deleted, the
most prevalent classes remained the same when only spontaneous reports were considered:
opiate agonists, benzodiazepines, hydantoins, digoxin, anticoagulants, cyclosporine,
antineoplastics, cephalosporins, aminglycosides, and tacrolimus were most often
associated with preventable ADEs (in descending order). The only difference between
cases identified by tracer searching or spontaneous reporting was the absence
of insulin-induced hypoglycemia within spontaneous reports. Only one insulin-induced
hypoglycemic event was spontaneously reported, while 21 cases were identified
through the ADR tracer program (using a trigger for chart review of patients who
received 50% dextrose injection).
Analysis by type of error and preventability code. A total of 436 preventability
codes were assigned to the 317 reports (Table 1). When
multiple codes were assigned to single cases, they were often code 2 or 3. Code
2 (inappropriate route, dose, or frequency) was assigned to 68.1% of all ADRs.
These ADRs included a variety of different drugs and outcomes and did not show
a specific pattern. In contrast, preventability code 3 (therapeutic drug monitoring
and laboratory tests not performed or performed infrequently) was dominantly associated
with inappropriate monitoring or testing of INR and aPTT with anticoagulants (70%),
serum creatinine with potentially nephrotoxic drugs (10%), blood glucose with
insulin (6%), and digoxin serum concentrations (6%). Of these four high-risk areas,
three were already highlighted by the top 10 drug classes identified. The lack
of renal monitoring for patients receiving potentially nephrotoxic drugs, such
as aminoglycosides, vancomycin, methotrexate, and cyclosporine, was an additional
high-risk area for preventable ADEs.
In addition, six reports (5.0%) of preventable ADEs were caused by health care
providers not considering a patient's allergy history when choosing drug therapy
(preventability code 4). Drug types varied in this category, with the largest
proportion of preventable allergies associated with antimicrobials.
Of 80 ADEs resulting from drug-drug interactions, 30 were associated with psychoactive
drugs and led predominantly to somnolence or respiratory depression (opiate agonists,
benzodiazepines, antidepressants, hydantoins, skeletal muscle relaxants, promethazine,
The other preventability codes were rarely used.
Analysis by adverse outcome. Analysis by adverse outcome (Table
3) confirmed the same high-risk circumstances identified by analysis of major
drug categories. The most frequent outcomes were excessive anticoagulation and
major and minor bleeding events, which were mainly associated with anticoagulants
(three with nonsteroidal antiinflammatory drugs [NSAIDs], two with alteplase and
abciximab). The second most frequent outcomes categories were respiratory depression
or arrest and excessive sedation or somnolence associated with opiate agonists,
benzodiazepines, and other psychoactive drugs. Renal problems were associated
with aminoglycosides, vancomycin, methotrexate, cisplatin, angiotensinconverting-enzyme
inhibitors, and tacrolimus. Analysis by outcome identified two additional preventable
ADE high-priority circumstances that were not highlighted by the analysis by drug
class or preventability code. The first was seizures associated with imipenem
(three patients), cyclosporine (three), tacrolimus (two), ciprofloxacin, theophylline,
meperidine, desmopressin, epoprostenol, and fluconazole (one each). Preventability
codes associated with seizures involved drug-drug interactions or patient histories
of seizures that were not considered. The second most frequent preventable ADEs
not identified by previous analyses were extravasation, necrosis, and edema associated
with inappropriate administration and infiltration of sympathomimetics, amrinone,
Table 4 lists the 10 most common preventable ADEs identified
by the ADR report database by type of drug or drug class, underlying medication
error, and associated adverse outcome. Almost 251 (80%) of all ADEs could be explained
by these 10 unique preventable ADE high-priority areas. Of these, 88 events (35%)
were asymptomatic and only identified by abnormal aPTT and INR values. The 251
high-priority preventable ADEs identified were identified and reported in approximately
150,000 admissions between 1994 and 2000, indicating a preventable ADE (from spontaneous
reporting) rate of 0.2%, or 2 of every 1000 hospitalized patients.
Unlike previously reported systematic analyses of preventable ADEs, our findings
suggest that specific high-priority preventable ADEs can be identified and targeted
for quality-improvement efforts. Over-anticoagulation associated with overdoses
of warfarin and heparin and insufficient monitoring, oversedation and respiratory
depression associated with overdoses and drug- drug interactions of opiate agonists
or benzodiazepines, and hypoglycemia associated with insulin overdoses and insufficient
monitoring accounted for more than half of the preventable ADEs recorded in this
The generalizability of the identified high-priority preventable ADEs to other
hospitals is indirectly supported by previous studies that identified preventable
ADEs in hospitals with systematic expert chart review.[4, 22, 23] Even
though these studies do not provide explicit descriptions of the identified preventable
ADEs, comparison with the reported drug classes supports our findings. Analgesics
and opiate agonists (12-29%) and sedatives and psychotropics (11-23%) ranked within
the top three drug classes associated with ADEs in all three studies. Bates and
colleagues found that analgesics, sedatives, antimicrobials, antipsychotics,
and diabetes medications are most often associated with preventable ADEs.
However, the ADR database does not represent the universe of preventable ADEs
and may omit other equally important areas. Moreover, ADR reports may represent
only about 5-10% of all ADR occurrences.[24, 25] The decision to report
events might depend on the severity of the ADR or litigation concerns.
Spontaneous reporting of ADRs and errors is also biased toward causal associations
between drugs and events that are known and well established.[25, 27]
For example, ADR reporting rates will increase independently of drug-prescribing
rates because the reaction becomes more widely appreciated.
Similarly, ADR tracer programs can only identify preventable ADEs that are treated
with specific drugs or antidotes. Since approximately two thirds of the preventable
ADEs analyzed in this study were identified by a tracer program, the high-priority
areas could have been systematically influenced, and spontaneous reports might
better mirror the true heterogeneity of preventable ADEs. However, when only spontaneous
preventable ADEs were considered, only hypoglycemia treated with 50% dextrose
injection was excluded from the list of high-priority areas previously identified.
Drug-induced hypoglycemia, although not routinely reported by health care practitioners,
has been described as a common problem when assessed by expert audit.
Thus, the tracer drug program and spontaneous reports may complement each other
to some extent.
It is important to note that ADR databases usually include only errors of commission,
such as overdoses or the choice of a contraindicated drug.
Omission errors (lack of necessary drug therapy or undertreatment) are not well
represented, even though they have been found to contribute substantially to preventable
ADEs. Studies have only recently begun including omission errors
in their definition of ADEs, and examples of therapeutic failure and lack of access
to drug therapy are comparably rare. Our review identified only three preventable
ADE studies in hospital inpatients that specifically included omission errors
(i.e., errors with a clear definition in their inclusion criteria or example cases).[4,
22, 29] The only omission cases that were relevant to our purposes included
uncontrolled pain resulting from undertreatment with opioids, therapeutic failure
associated with the selection of ineffective antimicrobials, and gastric bleeding
in patients taking NSAIDs with lack of preventive comedication.
None of these adverse consequences of omission errors were identified in our analyses,
probably because they were not reported and because the applied preventability
criteria would not capture "lack of drug therapy" or "underdoses" as a preventable
ADE. The inclusion of omission errors would provide a more comprehensive and balanced
assessment of drug therapy quality and should be addressed with updated reporting
and assessment criteria.
The data for these analyses were representative of only one particular hospital,
and it could be argued that differences might be observed between academic versus
community and tertiary versus primary care hospitals. However, the comparison
of our results with published literature supports the identified high-priority
areas as well described and very common across institutions. These high-priority
areas are not representative of all possible preventable ADEs, but they are valuable
areas in which to initiate systematic quality improvement. These areas need to
be supplemented with reports and analyses of omission errors and other new medication
errors as the definition of evidencebased guidelines evolves.
The novel systematic method used in this analysis to define and aggregate reports
was effective in identifying high-priority preventable ADEs. Aggregation by preventability
code or type of adverse outcome added new aspects to the analysis by drug class.
Even though various drugs can cause allergies, the error of not considering a
documented history of allergy makes this preventable ADE a unique problem. Each
high-priority preventable ADE presents an indication (or contraindication) and
a specific inappropriate pattern of care that is associated with an adverse patient
The risk of the identified medications has been well documented by previous studies.
However, this risk seems to expand from the drug product to its use (i.e., these
medications were also most often associated with inappropriate care and preventable
ADEs). Moreover, most of these inappropriate patterns of care or medication errors
were repeated despite known evidence-based standards or guidelines (as opposed
to random slips or lapses). For example, published literature emphasizes drug-
drug interactions with drugs that have additive sedative effects and suggests
dosage reductions, but these adjustments were repeatedly not performed for various
reasons. These reasons need to be explored in root-cause analyses to identify
ways to improve the current drug-use system. A review of the reported hypoglycemic
events suggested that insulin-dependent patients tend to miss meals because they
are scheduled for diagnostic procedures, yet they received normal insulin dosages
in expectation of normal food intake (e.g., extensive waiting times for x-rays
or other procedures). Frequent causes of the identified high-priority preventable
ADEs were lack of monitoring, insufficient consideration of drug dosing and drug-drug
interactions, and insufficient documentation, access, and flow of patient and
By combining elements of drug classes, ADEs, and error types, we identified common
patterns of inappropriate care. Only a few high-priority areas could describe
the majority of ADR reports assessed as preventable. Further efforts to enhance
the information on the nature of preventable ADEs should focus on two areas: training
of clinicians to improve the number and comprehensiveness of preventable ADE reports,
and development and implementation of more advanced classification systems to
identify common patterns and root causes of preventable ADEs.
Analysis of a hospital ADR database identified circumstances associated with clinically
significant preventable ADEs.
Presented at a symposium at the ASHP Midyear Clinical Meeting, New Orleans,
LA, December 4, 2001.
Kohn LT, Corrigan JM, Donaldson MD, eds. To err is human: building a safer
health system. Washington, DC: National Academy Press; 1999.
Winterstein AG, Sauer BC, Hepler CD et al. Preventable drug-related hospital
admissions. Ann Pharmacother. 2002; 36: 1238-48.
Revisions to Joint Commission standards in support of patient safety and
medical/ health care error reduction. Oakbrook Terrace, IL: Joint Commission
on Accreditation of Healthcare Organizations; 2001.
Bates DW, Cullen DJ, Laird N et al. Incidence of adverse drug events and
potential adverse drug events: implications for prevention. JAMA. 1995;
Hallas J, Gram LF, Grodum E et al. Drug related admissions to medical wards:
a population based survey. Br J Clin Pharmacol. 1992; 33:61-8.
Hatton RC. Criteria-based adverse drug reaction reporting and the use of
a relational database. Top Hosp Pharm Manage. 1992; 12:40-8.
Johnston PE, Morrow JD, Brennan TA. Use of a database computer program to
identify trends in reporting of adverse drug reactions. Am J Hosp Pharm.
Faich GA. Adverse drug reaction monitoring. N Engl J Med. 1986; 314:1589-92.
Accreditation manual for hospitals. Oak-brook Terrace, IL: Joint Commission
on Accreditation of Hospitals; 1978.
Schumock GT, Thornton JP. Focusing on the preventability of adverse drug
reactions. Hosp Pharm. 1992; 27:538.
Suh DC, Woodall BS, Shin SK et al. Clinical and economic impact of adverse
drug reactions in hospitalized patients. Ann Pharmacother. 2000; 34:1373-9.
Classen DC, Pestotnik SL, Evans C et al. Adverse drug events in hospitalized
patients: excess length of stay, extra costs, and attributable mortality.
JAMA. 1997; 277:301-6.
Seeger JD, Kong SX, Schumock GT. Characteristics associated with ability
to prevent adverse drug reactions in hospitalized patients. Pharmacotherapy.
Pearson TF, Pittman DG, Longley JM et al. Factors associated with preventable
adverse drug reactions. Am J Hosp Pharm. 1994; 51:2268-72.
Dalton-Bunnow MF, Halvachs FJ. Computer-assisted use of tracer antidote
drugs to increase detection of adverse drug reactions: a retrospective and
concurrent trial. Hosp Pharm. 1993; 28:746-9,752-5.
Koch KE. Use of standardized screening procedures to identify adverse drug
reactions. Am J Hosp Pharm. 1990; 47:1314-20.
Food and Drug Administration. The MedWatch online voluntary reporting form
(3500). www.accessdata.fda.gov/ scripts/medwatch/ (accessed 2002 18 Jun).
Jones JK. Adverse drug reactions in the community health setting: approaches
to recognizing, counseling, and reporting. Fam Community Health. 1982;
Leape LL, Kabcenell AI, Gandhi TK et al. Reducing adverse drug events: lessons
from a breakthrough series collaborative. Jt Comm J Qual Improv. 2000;
Hepler CD, Strand LM. Opportunities and responsibilities in pharmaceutical
care. Am J Hosp Pharm. 1990; 47:533-43.
AHFS drug information 2001. McEvoy GK, ed. Bethesda, MD: American Society
of Health-System Pharmacists; 2001.
Thomas EJ, Studdert DM, Burstin HR et al. Incidence and types of adverse
events and negligent care in Utah and Colorado. Med Care. 2000; 38:261-71.
Gray SL, Sager M, Lestico MR et al. Adverse drug events in hospitalized
elderly. J Gerontol A Biol Sci Med Sci. 1998; 53: M59-63.
Classen DC, Pestotnik SL, Evans RS et al. Computerized surveillance of adverse
drug events in hospital patients. JAMA. 1991; 266:2847-51.
Bero LA, Lipton HL, Bird JA. Characterization of geriatric drug-related
hospital readmissions. Med Care. 1991; 29:989-1003.
Leape LL. Error in medicine. JAMA. 1994; 272:1851-7.
Cullen DJ, Bates DW, Leape LL. Prevention of adverse drug events: a decade
of progress in patient safety. J Clin Anesth. 2000; 12:600-14.
Fischer KF, Lees JA, Newman JH. Hypoglycemia in hospitalized patients: causes
and outcomes. N Engl J Med. 1986; 315:1245-50.
Kaushal R, Bates DW, Landrigan C et al. Medication errors and adverse drug
events in pediatric inpatients. JAMA. 2001; 285:2114-20.
Supported by a grant from the American Society of Health-System Pharmacists Research
and Education Foundation, Bethesda, MD, and the University of Florida, Research
and Graduate Programs, Research Opportunity Fund, Gainesville, FL.
Dr. Winterstein at the Department of Pharmacy Health Care Administration, College
of Pharmacy, University of Florida, P.O. Box 100496, Suite P111, 1600 S.W. Archer
Road, Gainesville, FL 32610 (firstname.lastname@example.org).
Almut G. Winterstein, Ph.D., is Clinical Assistant Professor, Department
of Pharmacy Health Care Administration, College of Pharmacy, University of Florida
(UF), Gainesville. Randy C. Hatton, Pharm.D., Bcps, is Co-Director, Shands
at the University of Florida (SUF), Gainesville, and Clinical Professor, Department
of Pharmacy Practice, College of Pharmacy, UF. Ricardo Gonzalez-Rothi, M.D.,
is Professor, Department of Medicine, College of Medicine, UF. Thomas E. Johns,
Pharm.D., Bcps, is Manager, Clinical Practice Operations, Department of Pharmacy,
SUF, and Clinical Assistant Professor, Department of Pharmacy Practice, College
of Pharmacy, UF. Richard Segal, Ph.D., is Professor and Chair, Department
of Pharmacy Health Care Administration, College of Pharmacy, UF.