- Email: [email protected]
Incidence and preventability of adverse drug events in nursing homes
Incidence and Preventability of Adverse Drug Events in Nursing Homes Jerry H. Gurwitz, MD, Terry S. Field, DSc, Jerry Avorn, MD, Danny McCormick, MD, MPH, Shailavi Jain, RPh, Marie Eckler, RN, MS, Marcia Benser, RN, MS, Amy C. Edmondson, PhD, David W. Bates, MD PURPOSE: Adverse drug events, especially those that may have been preventable, are among the most serious concerns about medication use in nursing homes. We studied the incidence and preventability of adverse drug events and potential adverse drug events in nursing homes. METHODS: We performed a cohort study of all long-term care residents of 18 community-based nursing homes in Massachusetts during a 12-month observation period. Potential drug-related incidents were detected by stimulated self-report by nursing home staff and by periodic review of the records of nursing home residents by trained nurse and pharmacist investigators. Each incident was classified by 2 independent physician-reviewers, using a structured implicit review process, by whether or not it constituted an adverse drug event or potential adverse drug event (those that may have caused harm, but did not because of chance or because they were detected), by the severity of the event (significant, serious, life-threatening, or fatal), and by whether it was preventable. Examples of significant events included nonurticarial rashes, falls without associated fracture, hemorrhage not requiring transfusion or hospitalization, and oversedation; examples of serious events included urticaria, falls with fracture, hemorrhage requiring transfusion or hospitalization, and delirium. RESULTS: During 28,839 nursing home resident-months of
observation in the 18 participating nursing homes, 546 adverse drug events (1.89 per 100 resident-months) and 188 potential adverse drug events (0.65 per 100 resident-months) were identified. Of the adverse drug events, 1 was fatal, 31 (6%) were life-threatening, 206 (38%) were serious, and 308 (56%) were significant. Overall, 51% of the adverse drug events were judged to be preventable, including 171 (72%) of the 238 fatal, lifethreatening, or serious events and 105 (34%) of the 308 significant events (P ⬍ 0.001). Errors resulting in preventable adverse drug events occurred most often at the stages of ordering and monitoring; errors in transcription, dispensing, and administration were less commonly identified. Psychoactive medications (antipsychotics, antidepressants, and sedatives/hypnotics) and anticoagulants were the most common medications associated with preventable adverse drug events. Neuropsychiatric events were the most common types of preventable adverse drug events. CONCLUSIONS: Adverse drug events are common and often preventable in nursing homes. More serious adverse drug events are more likely to be preventable. Prevention strategies should target the ordering and monitoring stages of pharmaceutical care. Am J Med. 2000;109:87–94. 䉷2000 by Excerpta Medica, Inc.
M
garding medication use in nursing homes. Although previous studies have evaluated the patterns and quality of medication prescribing in nursing homes (3–9), few studies have systematically examined the incidence of adverse drug events in nursing home residents (10). A retrospective review of “incident reports” that were submitted about adverse and unexpected events in an academically oriented, 700-bed, long-term care facility identified 50 reports of adverse drug reactions during a 1-year period (11). Skin rashes were the most frequently reported events, and antibiotics were the most commonly implicated type of medication. The limited number of reports of drug-related events in that study suggests that voluntary reporting systems in nursing homes led to very low reporting rates of a narrow spectrum of events. The preventability of adverse drug events in nursing homes is not known. To address this issue, we undertook an investigation in a convenience sample of 18 community-based nursing homes in Massachusetts to evaluate the incidence and preventability of adverse drug events in nursing home residents; to categorize adverse drug events by drug class, severity, and clinical effects; and to classify
ore than 1.5 million Americans receive care in nursing homes each year, and the number of beds committed to nursing home care in the U.S. exceeds the number of acute care beds (1). The average U.S. nursing home resident uses 6 different medications, and more than 20% use 10 or more different drugs (2). The occurrence of adverse drug events that may have been preventable is among the most serious concerns reFrom the Meyers Primary Care Institute, the Fallon Healthcare System, and the University of Massachusetts Medical School (JHG, TSF, DM, SJ, ME, MB), Worcester, Massachusetts; and Brigham and Women’s Hospital (JA, DWB) and Harvard University (JA, ACE, DWB), Boston, Massachusetts. Supported by research grant AG14472 from the National Institute on Aging, Bethesda, Maryland. The contents are solely the responsibility of the authors and do not necessarily reflect the official views of the National Institute on Aging. Requests for reprints should be addressed to Jerry H. Gurwitz, MD, Meyers Primary Care Institute, Fallon Healthcare System and University of Massachusetts Medical School, 100 Central Street, Worcester, Massachusetts 01608. Manuscript submitted September 14, 1999, and accepted in revised form April 3, 2000. 䉷2000 by Excerpta Medica, Inc. All rights reserved.
0002-9343/00/$–see front matter 87 PII S0002-9343(00)00451-4
Adverse Drug Events in Nursing Homes/Gurwitz et al
preventable events by the stage of the care process at which the error occurred.
METHODS Study Setting Eighteen community-based nursing homes located in central and eastern Massachusetts participated in the project. These facilities were recruited from among the 81 nursing homes with more than 50 beds in this geographic region that were served by a large pharmacy provider to long-term care facilities. At the time of this study, this pharmacy provider delivered services for 42,000 nursing home residents in Massachusetts, representing more than 80% of all long-term care beds in the state. The pharmacy provider assisted in the recruitment of the study nursing homes through invitational letters, telephone calls, and visits. The mean (⫾ SD) number of beds in participating facilities was 149 ⫾ 62 (range 72 to 333). Eleven of the study facilities were proprietary; the remainder were voluntary nonprofit nursing homes. All participating facilities were certified by both Medicare and Medicaid.
Study Sample The 18 study nursing homes were enrolled from March 14, 1997, through May 15, 1997. Subjects included all long-stay residents during the 12-month period following each facility’s enrollment. Residents admitted for short-term care (eg, subacute care or rehabilitation following hospitalization) were excluded. The study was approved by the institutional review board of the University of Massachusetts Medical School. Because of potential liability concerns, details of individual cases are not provided, and participating facilities are not identified.
Case-Finding Definitions and Classification of Events At 6-week intervals, systematic chart reviews were performed by 2 trained nurse investigators and 1 pharmacist investigator for each eligible nursing home resident. These investigators examined the nursing home records for possible drug-related incidents, such as new symptoms or events that might represent an adverse drug event, changes in medication regimens (including acute discontinuations or initiations of medications that might be used to treat a drug-induced effect), abnormal laboratory values, and all emergency room transfers and hospitalizations. In addition, the investigators encouraged nursing home staff to inform them of any possible drugrelated incidents, as well as incidents that may not have seemed to be a drug-related event (eg, new falls). Whenever possible, nursing home staff were interviewed to ascertain information about incidents beyond that available in the nursing home resident records. The nurse and pharmacist investigators had no direct contact with nursing home residents. 88
August 1, 2000
THE AMERICAN JOURNAL OF MEDICINE威
Volume 109
After an extensive training period, between-case investigator reliability for identifying relevant incidents in nursing home records was assessed through independent review of the same 10 medical records by each of the 3 case investigators. All 3 identified the same incident in 9 of the 10 nursing home records; for the remaining record, 2 of the case investigators identified the incident correctly. Bimonthly refresher sessions were held with the 3 case investigators throughout the study to reinforce consistent identification and record abstraction procedures. The primary outcome of the study was an adverse drug event, defined as an injury resulting from the use of a drug. Adverse drug events may have resulted from medication errors (ie, errors in prescribing, dispensing, administration, and monitoring) or from adverse drug reactions in which no error was involved. We also identified potential adverse drug events (“near misses”), defined as errors that had the capacity to cause injury but failed to do so, either by chance or because they were intercepted (eg, a nursing home resident who received an antibiotic despite a known allergy, but who did not have a reaction; or an excessively high international normalized ratio ([INR] ⱖ4) that occurred as the result of a drug interaction with warfarin when no effort was made to prevent the event by increasing the frequency of monitoring or reducing the warfarin dose). We did not include incidents in which an error was made but was judged to have minimal potential for injury. By definition, all potential adverse drug events were considered preventable. The definitions for adverse drug events and potential adverse drug events are consistent with definitions used previously (12–15). All possible drug-related incidents were evaluated independently by 2 physician-reviewers (2 internist-geriatricians [JHG, JA] and 2 general internists [DWB, DM]). The reviewers classified these incidents using a structured implicit review according to the following criteria: whether an adverse drug event or potential adverse drug event was present, the severity of the event, whether the event was preventable, and the effects on the resident. In determining whether an adverse drug event had occurred, the physician-reviewers considered the temporal relation between the drug exposure and the event, as well as whether the event reflected a known effect of the drug. Suboptimal prescribing choices (16 –18), such as the use of long-acting hypoglycemic agents, highly anticholinergic tricyclic antidepressants, or long elimination halflife benzodiazepines, were not considered unless an adverse drug event had occurred relating to that choice. Categories of severity were significant, serious, lifethreatening, and fatal (19). Examples of adverse drug events categorized as significant included a nonurticarial skin rash, a fall without associated fractures, hemorrhage not requiring transfusion or hospitalization, and oversedation. Examples of events categorized as serious included urticaria, falls with associated fractures, hemor-
Adverse Drug Events in Nursing Homes/Gurwitz et al
Table 1. Rates and Severity of Adverse Drug Events and Potential Adverse Drug Events among 2,916 Residents of 18 Nursing Homes Observed for a Mean of 9.9 Months
Type of Drug-Related Event
Number (Percent)
Rate per 100 Resident-Months (95% Confidence Interval)
Category of Severity Fatal
Life-Threatening
Serious
Significant
Number (Percent) Adverse drug events Preventable Nonpreventable Potential adverse drug events*
546 276 (50) 270 (50) 188
1.89 (1.74–2.05) 0.96 (0.85–1.07) 0.94 (0.83–1.05) 0.65 (0.56–0.75)
1 (0.2) 1 (0.4) 0 —
31 (6) 25 (9) 6 (2) 16 (9)
206 (38) 145 (52) 61 (23) 149 (79)
308 (56) 105 (38) 203 (75) 23 (12)
* All of these were considered preventable by definition.
rhage requiring transfusion or hospitalization but without hypotension, and delirium. Examples of lifethreatening events included hemorrhage with associated hypotension, hypoglycemic encephalopathy, and liver failure. Adverse drug events were considered to be preventable if they were due to an error and were preventable by any means then available. Categories of preventability included definitely preventable, probably preventable, probably not preventable, and definitely not preventable (20); results were collapsed into preventable and nonpreventable categories in the analyses. The effects of an adverse drug event on a nursing home resident were characterized by duration (ⱕ1 day of symptoms, ⬎1 day of symptoms) and by disability (nonpermanent, permanent). Disability was defined as a decrease in functional status from its previous level (21) and referred to a new deficit in activities of daily living as documented in the nursing home records (bathing, dressing, toileting, transferring, continence, or feeding). We also classified the stages of pharmaceutical care as ordering (essentially all by physicians), transcribing (performed by a secretary or nurse, depending on the nursing home unit and the time of day), dispensing (by the pharmacy vendor), administration (by nurses), and monitoring (by physicians, nurses, and the consultant pharmacist to the nursing home). Periodic review of medication regimens of residents of US long-term care facilities by consultant pharmacists has been federally mandated since 1974 (22). When there were disagreements that affected the classification of an incident by the 2 physician-reviewers (eg, 1 reviewer scored it as preventable, but the other did not), or about the presence of an event, its severity, or its preventability, the reviewers met and reached consensus. Consensus was reached in all instances where there was initial disagreement. For a random sample of 100 drugrelated incidents that were evaluated independently by 2 physician-reviewers, interrater reliabilities for initial judgments were calculated using the statistic. For judgments about the presence of an adverse drug event, the value was 0.80; for preventability, the value was 0.73; and for severity, the value was 0.62.
Prevalence of Medication Use To estimate the underlying prevalence of use of selected medication categories in the overall sample of nursing home residents, we collected drug use data from the medication administration records of a representative sample of 410 residents who were randomly selected from each facility in proportion to the number of residents identified as having experienced an adverse drug event in that facility. These residents had not experienced an adverse drug event at the time of data collection on medication use.
Analysis To determine crude rates of events, the numbers of adverse drug events or potential adverse drug events were divided by the total number of nursing home residentmonths, which was estimated by obtaining census data for all eligible subjects in each study nursing home at 6-week intervals throughout the course of the project; absences from the nursing home (eg, for hospitalization) were accounted for. Ninety-five percent confidence intervals (CI) were calculated for these estimates (23). Comparisons between categorical variables were performed using the chi-square test. A P value ⬍ 0.05 was considered significant. Analyses were performed using SAS, version 6 (SAS Institute, Inc., Cary, North Carolina).
RESULTS During the study period, 2916 nursing home residents (mean [⫾ SD] age of 84 ⫾ 9 years; 77% women) contributed 28,839 resident-months of observation in the 18 participating nursing homes. The nurse and pharmacist investigators identified 979 possible drug-related incidents during the study, of which 245 (25%) were not considered to be adverse drug events or potential adverse drug events by the physician-reviewers. Of the 734 remaining incidents, 546 (74%) were classified as adverse drug events and 188 (26%) were classified as potential adverse drug events (Table 1). Seventeen percent (n ⫽ 93) of the adverse drug events were identified by nursing August 1, 2000
THE AMERICAN JOURNAL OF MEDICINE威
Volume 109 89
Adverse Drug Events in Nursing Homes/Gurwitz et al
Table 2. Effects of Adverse Drug Events on Nursing Home Residents Duration of Symptoms Type of Drug-Related Event
ⱕ1 day
Type of Disability
⬎1 day
Nonpermanent
Permanent
Number (Percent) Adverse drug events (n ⫽ 546) Preventable (n ⫽ 276) Nonpreventable (n ⫽ 270)
92 (17) 33 (12) 59 (22)
416 (76) 216 (78) 200 (74)
staff reporting and the remainder through review of nursing home records. Only 4% (n ⫽ 7) of the potential adverse drug events were identified through nursing home staff reporting. Of the 546 adverse drug events, about half (n ⫽ 276) were considered preventable. (By definition, all 188 potential adverse drug events were considered preventable.) The overall rate of adverse drug events was 1.89 per 100 resident-months, with a rate of 0.96 preventable adverse drug events per 100 resident-months. The rate of potential adverse drug events was 0.65 per 100 residentmonths. One resident suffered a fatal adverse drug event (methotrexate-induced hepatotoxicity) during the study period (Table 1); this event was characterized as preventable. Of the 238 fatal, life-threatening, or serious adverse drug events, 72% (n ⫽ 171) were deemed preventable, compared with 34% (n ⫽ 105) of the 308 significant adverse drug events. Overall, more severe adverse drug events were more likely to be preventable (relative risk ⫽ 2.1, 95% CI: 1.8 to 2.5, P ⬍ 0.001). A greater percentage of potential adverse drug events was considered life-threatening or serious as compared with adverse drug events (88% vs 44%). Only a small minority (17%) of adverse drug events resulted in symptoms of ⱕ1 day in duration (Table 2). Thirty-eight events resulted in disability to the nursing home resident; in 8 cases the disability was permanent. Preventable adverse drug events were more likely to result in disability as compared with nonpreventable adverse drug events (relative risk ⫽ 2.4, 95% CI: 1.2 to 4.7, P ⬍ 0.01). The 546 adverse drug events were associated with a wide variety of different drug classes (Table 3). Antipsychotics were the most frequently implicated agents associated with these events, followed by antibiotics, antidepressants, and sedatives/hypnotics. Psychoactive medications (antipsychotics, antidepressants, and sedatives/hypnotics) and anticoagulants were the most common drug categories associated with preventable adverse drug events. Psychoactive drugs were commonly used by nursing home residents; more than 36% of residents used antidepressants, 24% used sedatives/hypnotics, and 17% used antipsychotic medications (Table 4). 90
August 1, 2000
THE AMERICAN JOURNAL OF MEDICINE威
Volume 109
30 (6) 24 (9) 6 (2)
8 (1) 3 (1) 5 (2)
Antibiotics were associated with fewer than 5% of all preventable adverse drug events, but they were responsible for 36% of nonpreventable adverse drug events (P ⬍ 0.0001). The vast majority of antibiotic-associated adverse drug events were rashes and confirmed Clostridium difficile diarrhea. The distribution of drug classes associated with potential adverse drug events was different from that of adverse drug events (Table 3). Almost 80% of all potential adverse drug events were associated with the use of warfarin, primarily the development of excessively high INR values (ⱖ4) due to errors in anticoagulation management (the mean INR in these patients was 6.1 [range 4.0 to 15.6]). Prescribers frequently made no modifications in the warfarin dosing scheme or the usual frequency of monitoring when drugs with well-established interactions with warfarin were prescribed. Consistently subtherapeutic INR levels (at least three consecutive values ⬍1.5 during several weeks of treatment) were also considered potential adverse drug events when the appropriate therapeutic range for the specified indication was 2 to 3. The mean INR in these patients was 1.2 (range 0.8 to 1.4). Neuropsychiatric events (eg, oversedation, confusion, hallucinations, delirium) were the most common types of both preventable and nonpreventable adverse drug events (Table 5). Falls were involved in 20% of preventable adverse drug events but only 4% of nonpreventable events. Similarly, bleeding was responsible for 14% of preventable adverse drug events compared with only 6% of nonpreventable adverse drug events. Dermatologic events were rarely deemed preventable. Essentially all infections were deemed nonpreventable, as this category almost entirely consisted of C. difficile diarrhea associated with previous antibiotic therapy. Gastrointestinal events, including nausea, vomiting, diarrhea, and abdominal pain, were involved in 12% of all adverse drug events. Among the 464 preventable events, errors occurred most commonly in the ordering (n ⫽ 315, 68%) and monitoring (n ⫽ 325, 70%) stages of care. Errors accounting for preventable events were rarely identified in the transcription (0.7%), dispensing (0.5%), or administration (3%) stages. Among the 315 ordering errors, wrong dose (eg, excessive dose for an elderly patient) was the most common (n ⫽ 200, 63%), followed by the pre-
Adverse Drug Events in Nursing Homes/Gurwitz et al
Table 3. Frequency of Adverse Drug Events and Potential Adverse Drug Events by Drug Class*
Drug Class
Adverse Drug Events (n ⫽ 546)
Preventable Adverse Drug Events (n ⫽ 276)
Nonpreventable Adverse Drug Events (n ⫽ 270)
Potential Adverse Drug Events (n ⫽ 188)
Number (Percent) Antipsychotics Antibiotics/antiinfectives Antidepressants Sedatives/hypnotics Anticoagulants Antiseizure Cardiovascular Hypoglycemics Nonopioid analgesics Opioids Antiparkinsonians Gastrointestinal Diuretics Antigout Muscle relaxants Alzheimer’s disease Nutrients/supplements Ophthalmics Antineoplastics Respiratory Osteoporosis Steroids Nonophthalmic topicals Antihistamines Antihyperlipidemics Miscellaneous
125 (23) 109 (20) 68 (13) 68 (13) 51 (9) 47 (9) 35 (6) 27 (5) 22 (4) 15 (3) 12 (2) 11 (2) 10 (2) 6 (1) 6 (1) 4 (0.7) 3 (0.5) 3 (0.5) 2 (0.4) 2 (0.4) 2 (0.4) 1 (0.2) 1 (0.2) 0 0 2 (0.4)
72 (26) 13 (5) 50 (18) 49 (18) 37 (13) 27 (10) 25 (9) 14 (5) 13 (5) 7 (3) 7 (3) 6 (2) 9 (3) 4 (1) 4 (1) 3 (1) 1 (0.4) 1 (0.4) 2 (0.7) 1 (0.4) 0 (0.0) 1 (0.4) 0 0 0 0
53 (20) 96 (36) 18 (7) 19 (7) 14 (5) 20 (7) 10 (4) 13 (5) 9 (3) 8 (3) 5 (2) 5 (2) 1 (0.4) 2 (0.7) 2 (0.7) 1 (0.4) 2 (0.7) 2 (1.0) 0 1 (0.4) 2 (0.7) 0 1 (0.4) 0 0 2 (0.7)
1 (0.5) 13 (7) 0 2 (1) 150 (80) 0 7 (4) 6 (3) 1 (0.5) 0 0 1 (0.5) 5 (3) 0 0 0 1 (0.5) 0 0 0 0 0 1 (0.5) 2 (1) 1 (0.5) 0
* Some events were associated with drugs in more than one category; thus, frequencies in each column sum to greater than the total number of events.
scription of a drug for which there was a well-established, clinically important interaction with another drug (eg, trimethoprim/sulfamethoxazole-warfarin interaction; n ⫽ 68, 22%). Wrong choice errors were the third most common type of ordering error (n ⫽ 28, 9%). Examples of such errors included the use of an antidepressant with anticholinergic properties for a resident with identifiable side effects to the agent, the use of a long-acting benzodiazepine for a resident with drug-related oversedation, and the use of an antipsychotic medication to treat hallucinations due to levodopa/carbidopa therapy in a patient with Parkinson’s disease. Monitoring errors generally referred to inadequate laboratory monitoring of drug therapies or to a delayed response, or failure to respond, to signs or symptoms of drug toxicity or laboratory evidence of toxicity. Among the 325 monitoring errors, failure to act on available information relating to clinical findings or laboratory results (n ⫽ 133, 41%) and inadequate monitoring (n ⫽ 83, 26%) were the most common errors. Examples of moni-
toring errors included an inadequate frequency of monitoring of warfarin leading to an elevated INR value or failure to respond to symptoms suggestive of digoxin toxicity (eg, nausea, vomiting, and anorexia).
DISCUSSION We found that adverse drug events occurred commonly among nursing home residents and that more than half were preventable. Serious, life-threatening, and fatal adverse drug events were more likely to be preventable than less severe events. Most errors associated with preventable events occurred at the ordering and monitoring stages. Although it is difficult to compare the findings of the present study on adverse drug events in nursing home residents with studies performed in other clinical settings involving different groups of patients (24), some comparisons are of interest. Bates et al (12) identified adverse August 1, 2000
THE AMERICAN JOURNAL OF MEDICINE威
Volume 109 91
Adverse Drug Events in Nursing Homes/Gurwitz et al
Table 4. Use of Selected Drug Classes among a Representative Sample of Nursing Home Residents (n ⫽ 410) Drug Class
Number (Percent)
Gastrointestinal Nutrients/supplements Nonopioid analgesics Cardiovascular Antidepressants Diuretics Sedatives/hypnotics Antipsychotics Antibiotics/antiinfectives Antiseizure Ophthalmics Anticoagulants Hypoglycemics Respiratory Steroids Opioids Antiparkinsonians Muscle relaxants Osteoporosis Antigout Antihistamines Antihyperlipidemics Alzheimer’s disease Antineoplastics Nonophthalmic topicals
308 (75) 280 (68) 258 (63) 201 (49) 149 (36) 143 (35) 97 (24) 71 (17) 57 (14) 53 (13) 49 (12) 47 (12) 42 (10) 32 (8) 31 (8) 28 (7) 28 (7) 14 (3) 12 (3) 11 (3) 10 (2) 8 (2) 5 (1) 5 (1) 2 (0.5)
drug events occurring during 4,031 adult admissions to two Boston tertiary care hospitals during a 6-month period. Of the 247 adverse drug events, 1% were fatal, 12% were life-threatening, 30% were serious, and 57% were significant. Twenty-eight percent were judged preventable. Of the life-threatening and serious adverse drug events, 42% were judged preventable, compared with 18% of significant adverse drug events. In the nursing home setting, more than half of all adverse drug events were preventable; similar to the situation in the hospital setting, more serious events were more likely to be judged preventable. In this study, psychoactive drugs (antipsychotics, antidepressants, and sedatives/hypnotics) and anticoagulants were the most common drug categories associated with preventable adverse drug events. Suboptimal prescribing of psychoactive medications in nursing homes has long been an issue of concern (25), leading to federal regulations about their use in this setting (26). Various interventions have been developed and tested to reduce the risks of psychoactive drug use in nursing homes (27). Academic detailing to educate physicians and nursing home staff about the use of psychoactive medications in geriatric patients has had some success in improving the quality of drug prescribing, specifically for antipsychotic medications and long-acting benzodiazepines (28,29). 92
August 1, 2000
Table 5. Frequency of Adverse Drug Events by Type*
THE AMERICAN JOURNAL OF MEDICINE威
Volume 109
Type
Adverse Drug Events (n ⫽ 546)
Preventable (n ⫽ 276)
Number (Percent) Neuropsychiatric Falls Gastrointestinal Dermatologic/allergic Hemorrhage Extrapyramidal symptoms/ tardive dyskinesia Infection Metabolic/endocrine Anorexia/weight loss Ataxia/difficulty with gait Cardiovascular Electrolyte/fluid balance abnormality Syncope/dizziness Functional decline† Respiratory Anticholinergic‡ Renal Hematologic Hepatic
150 (27) 67 (12) 65 (12) 59 (11) 57 (10) 52 (10)
83 (30) 55 (20) 30 (11) 7 (3) 40 (14) 19 (7)
34 (6) 27 (5) 20 (4) 18 (3) 15 (3) 9 (2)
1 (0.4) 14 (5) 14 (5) 9 (3) 10 (4) 5 (2)
8 (1) 7 (1) 3 (0.5) 3 (0.5) 3 (0.5) 2 (0.4) 1 (0.2)
5 (2) 6 (2) 3 (1) 2 (0.7) 1 (0.4) 2 (0.7) 1 (0.4)
* Adverse drug events could manifest as more than one type. † Adverse drug event manifested only as decline in activities of daily living without any other more specific type of event. Other types of events may have been associated with functional decline. ‡ Anticholinergic effects include dry mouth, dry eyes, urinary retention, and constipation.
Anticoagulants were also a common cause of adverse drug events and were the leading cause of potential adverse drug events. Concerns have been raised about the quality of anticoagulant use in nursing homes (30). A more systematic approach to decision-making about the use of warfarin for stroke prevention in the frail elderly is required, as is a more consistent approach to the management of therapy. More widespread use of specialized anticoagulation clinics to provide coordinated care may offer an option to improve the effectiveness and safety of warfarin therapy for these high-risk patients. The nursing homes that participated in this study were somewhat different from the average US nursing home. The average number of beds per nursing home among the study facilities was higher than the national average (149 vs 106), due to the exclusion of facilities with fewer than 50 beds. Nationally, 66% of nursing homes are proprietary, compared with 61% of the study nursing homes. Of all nursing homes in the United States, 70% are certified by both Medicare and Medicaid (1), compared with 100% of the participating study facilities. Nationally, 75% of nursing home residents are women and 83% of all residents are aged 75 years or older (31), compared with
Adverse Drug Events in Nursing Homes/Gurwitz et al
77% who were women and 89% who were 75 years of age or more in this study. If the findings of the present study apply to all US nursing homes, then about 24 adverse drug events and 8 potential adverse drug events should be identifiable each year in an average facility, many of which should be preventable. Thus, about 350,000 adverse drug events— more than half of which are preventable— occur each year in the 1.55 million residents of US nursing homes. There are almost 20,000 fatal or life-threatening adverse drug events per year, of which 80% are preventable. These estimates are likely to be conservative. To ascertain information on drug-related incidents, we relied mainly on information contained in nursing home records, which can be quite limited compared with other clinical settings. Without adequate documentation of relevant information in the medical record, possible drugrelated incidents could have been missed by the case investigators, and inadequate information would not have been available for the physician-reviewers to classify incidents as adverse drug events. There was no direct assessment of nursing home residents in this study, and such assessment might have allowed the identification of additional events. Furthermore, the nursing homes that participated in this project volunteered to do so, perhaps suggesting an increased sensitivity to the problem of adverse drug events compared with the average long-term care facility. How should the findings of this study be applied to improve the quality of care for residents of nursing homes? One approach is to identify and admonish health care providers who make errors that lead to adverse drug events, but that approach ignores failures in the design of systems of care that often contribute to the occurrence of medical errors, as well as the injuries that result from some of those errors (32–35). Enhanced surveillance and reporting systems for adverse drug events in nursing homes are required, and educational efforts about the optimal use of drug therapies in frail elderly patients are essential. However, as Leape et al (34) have concluded about the occurrence of serious medication errors in the hospital setting, preventive efforts that focus solely on the individual provider or that rely on inspection alone have limited effects: “analysis and the correction of underlying systems faults is much more likely to result in enduring changes and significant error reduction.” Ordering and monitoring errors in the nursing home setting may be particularly amenable to prevention strategies utilizing systems-based approaches. The benefits of this approach in the hospital setting through the use of computerized order entry have recently been reported (36); such a system could be designed to focus on ordering and monitoring issues in the nursing home. Successes in the hospital setting should pave the way for similar efforts in nursing
homes aimed at reducing drug-related injuries and disability and improving the quality of care provided to this vulnerable population (37).
ACKNOWLEDGMENT The authors thank the nursing homes participating in this study for their commitment and dedication to improving the quality of care provided to nursing home residents. They also thank Mary Ellen Stansky and Jackie Cernieux, MPH, for their assistance with technical aspects relating to the study and Bessie Petropoulos for assistance with manuscript preparation.
REFERENCES 1. Strahan GW. An Overview of Nursing Homes and Their Current Residents: Data from the 1995 National Nursing Homes Survey. Advance Data from Vital and Health Statistics, No. 280. Hyattsville, Md: National Center for Health Statistics, 1997. 2. Bernabei R, Gambassi G, Lapane K, et al. Characteristics of the SAGE database: a new resource for research on outcomes in long– term care. J Gerontol Series A, Biological Sciences, Medical Sciences. 1999;54:M25–M33. 3. Shorr RI, Fought RL, Ray WA. Changes in antipyschotic drug use in nursing homes during implementation of the OBRA-87 regulations. JAMA. 1994;271:358 –362. 4. Beers M, Avorn J, Soumerai SB, et al. Psychoactive medication use in intermediate-care facility residents. JAMA. 1988;260:3016 – 3020. 5. Pickering TD, Gurwitz JH, Zaleznick D, et al. The appropriateness of oral fluoroquinolone prescribing in the long-term care setting. J Am Geriatr Soc. 1994;42:28 –32. 6. Gurwitz JH, Noonan JP, Soumerai SB. Reducing the use of H2receptor antagonists in the long-term care setting. J Am Geriatr Soc. 1992;40:359 –364. 7. Ray WA, Federspiel CF, Schaffner W. A study of antipsychotic drug use in nursing homes: epidemiologic evidence suggesting misuse. Am J Public Health. 1980;70:485– 491. 8. Svarstad BL, Mount JK. Nursing home resources and tranquilizer use among the institutionalized elderly. J Am Geriatr Soc. 1991;39: 869 – 875. 9. Beers MH, Ouslander JG, Fingold SF, et al. Inappropriate medication prescribing in skilled-nursing facilities. Ann Intern Med. 1992; 117:684 – 689. 10. Gerety MB, Cornell JE, Plichta DT, Eimer M. Adverse events related to drugs and drug withdrawal in nursing home residents. J Am Geriatr Soc. 1993;41:1326 –1332. 11. Gurwitz JH, Sanchez-Cross MT, Eckler MA, Matulis J. The epidemiology of adverse and unexpected events in the long-term care setting. J Am Geriatr Soc. 1994;42:33–38. 12. 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. 13. 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. 14. Bates DW, Leape LL, Cullen DJ, et al. Effect of computerized physician order entry and a team intervention on prevention of serious medication errors. JAMA. 1998;280:1311–1216. 15. Leape LL, Cullen DJ, Clapp MD, et al. Pharmacist participation on physician rounds and adverse drug events in the intensive care unit. JAMA. 1999;282:267–270. 16. Beers MH, Ouslander JG, Rollingher I, et al. Explicit criteria for August 1, 2000
THE AMERICAN JOURNAL OF MEDICINE威
Volume 109 93
Adverse Drug Events in Nursing Homes/Gurwitz et al
17.
18. 19.
20. 21.
22.
23. 24.
25. 26. 27.
94
determining inappropriate medication use in nursing home residents. Arch Intern Med. 1991;151:1825–1832. Beers MH. Explicit criteria for determining potentially inappropriate medication use by the elderly. An update. Arch Intern Med. 1997;157:1531–1536. Gurwitz JH. Suboptimal medication use in the elderly. The tip of the iceberg. JAMA. 1994;272:316 –317. Folli HL, Poole RI, Benitz WE, Russo JC. Medication error prevention by clinical pharmacists in two children’s hospitals. Pediatrics. 1987;79:781–722. Dubois RW, Brook RH. Preventable deaths: who, how often, and why. Ann Intern Med. 1998;190:582–589. Katz S, Ford AB, Moskowitz RW, et al. Studies of illness in the aged: the index of ADL. A standardized measure of biological and psychosocial function. JAMA. 1963;185:914 –919. Department of Health, Education, and Welfare. Skilled nursing facility standards for certification and participation in Medicare and Medicaid programs. Fed Regist. 1974;39:2238. Rosner B. Fundamentals of Biostatistics, 3rd ed. Boston, Mass: PWSKent; 1990. Phillips DP, Christenfeld, N, Glynn LM. Increase in U.S. medication-error deaths between 1983 and 1993. Lancet. 1998;351:643– 644, 1983. Rango N. Nursing home care in the United States. N Engl J Med. 1982;307:883– 889. Elon R, Paulson LG. The impact of OBRA on medical practice within nursing facilities. J Am Geriatr Soc. 1992;40:958 –963. Ray WA, Taylor JA, Meador KG, et al. A randomized trial of a
August 1, 2000
THE AMERICAN JOURNAL OF MEDICINE威
Volume 109
28.
29.
30.
31.
32. 33. 34. 35. 36.
37.
consultation service to reduce falls in nursing homes. JAMA. 1997; 278:557–562. Avorn J, Soumerai SB, Everitt DE, et al. A randomized trial of a program to reduce the use of psychoactive drugs in nursing homes. N Engl J Med. 1992;327:168 –173. Ray WA, Taylor JA, Meador KG, et al. Reducing antipsychotic drug use in nursing homes. A controlled trial of provider education. Arch Intern Med. 1993;153:713–721. Gurwitz JH, Monette J, Rochon PA, et al. Atrial fibrillation and stroke prevention with warfarin in the long-term care setting. Arch Intern Med. 1997;157:978 –984. Dey AN. Characteristics of Elderly Nursing Home Residents: Data from the 1995 National Nursing Home Survey. Advance Data from Vital and Health Statistics, No. 289. Hyattsville, Md: National Center for Health Statistics; 1997. Berwick DM. Continuous improvement as an ideal in health care. N Engl J Med. 1989;320:53–56. Leape LL. Error in medicine. JAMA. 1994;272:1851–1857. Leape LL, Bates DW, Cullen DJ, et al. Systems analysis of adverse drug events. JAMA. 1995;274:35– 43. Horton R. The uses of error. Lancet. 1999;353:422– 423. Bates DW, Leape LL, Cullen DJ, et al. Effect of computerized physician order entry and a team intervention on prevention of serious medication errors. JAMA. 1998;280:1311–1316. Kohn LT, Corrigan JM, Donaldson MS, eds. To Err is Human: Building a Safer Health System. Washington, DC: National Academy Press; 1999.
observation in the 18 participating nursing homes, 546 adverse drug events (1.89 per 100 resident-months) and 188 potential adverse drug events (0.65 per 100 resident-months) were identified. Of the adverse drug events, 1 was fatal, 31 (6%) were life-threatening, 206 (38%) were serious, and 308 (56%) were significant. Overall, 51% of the adverse drug events were judged to be preventable, including 171 (72%) of the 238 fatal, lifethreatening, or serious events and 105 (34%) of the 308 significant events (P ⬍ 0.001). Errors resulting in preventable adverse drug events occurred most often at the stages of ordering and monitoring; errors in transcription, dispensing, and administration were less commonly identified. Psychoactive medications (antipsychotics, antidepressants, and sedatives/hypnotics) and anticoagulants were the most common medications associated with preventable adverse drug events. Neuropsychiatric events were the most common types of preventable adverse drug events. CONCLUSIONS: Adverse drug events are common and often preventable in nursing homes. More serious adverse drug events are more likely to be preventable. Prevention strategies should target the ordering and monitoring stages of pharmaceutical care. Am J Med. 2000;109:87–94. 䉷2000 by Excerpta Medica, Inc.
M
garding medication use in nursing homes. Although previous studies have evaluated the patterns and quality of medication prescribing in nursing homes (3–9), few studies have systematically examined the incidence of adverse drug events in nursing home residents (10). A retrospective review of “incident reports” that were submitted about adverse and unexpected events in an academically oriented, 700-bed, long-term care facility identified 50 reports of adverse drug reactions during a 1-year period (11). Skin rashes were the most frequently reported events, and antibiotics were the most commonly implicated type of medication. The limited number of reports of drug-related events in that study suggests that voluntary reporting systems in nursing homes led to very low reporting rates of a narrow spectrum of events. The preventability of adverse drug events in nursing homes is not known. To address this issue, we undertook an investigation in a convenience sample of 18 community-based nursing homes in Massachusetts to evaluate the incidence and preventability of adverse drug events in nursing home residents; to categorize adverse drug events by drug class, severity, and clinical effects; and to classify
ore than 1.5 million Americans receive care in nursing homes each year, and the number of beds committed to nursing home care in the U.S. exceeds the number of acute care beds (1). The average U.S. nursing home resident uses 6 different medications, and more than 20% use 10 or more different drugs (2). The occurrence of adverse drug events that may have been preventable is among the most serious concerns reFrom the Meyers Primary Care Institute, the Fallon Healthcare System, and the University of Massachusetts Medical School (JHG, TSF, DM, SJ, ME, MB), Worcester, Massachusetts; and Brigham and Women’s Hospital (JA, DWB) and Harvard University (JA, ACE, DWB), Boston, Massachusetts. Supported by research grant AG14472 from the National Institute on Aging, Bethesda, Maryland. The contents are solely the responsibility of the authors and do not necessarily reflect the official views of the National Institute on Aging. Requests for reprints should be addressed to Jerry H. Gurwitz, MD, Meyers Primary Care Institute, Fallon Healthcare System and University of Massachusetts Medical School, 100 Central Street, Worcester, Massachusetts 01608. Manuscript submitted September 14, 1999, and accepted in revised form April 3, 2000. 䉷2000 by Excerpta Medica, Inc. All rights reserved.
0002-9343/00/$–see front matter 87 PII S0002-9343(00)00451-4
Adverse Drug Events in Nursing Homes/Gurwitz et al
preventable events by the stage of the care process at which the error occurred.
METHODS Study Setting Eighteen community-based nursing homes located in central and eastern Massachusetts participated in the project. These facilities were recruited from among the 81 nursing homes with more than 50 beds in this geographic region that were served by a large pharmacy provider to long-term care facilities. At the time of this study, this pharmacy provider delivered services for 42,000 nursing home residents in Massachusetts, representing more than 80% of all long-term care beds in the state. The pharmacy provider assisted in the recruitment of the study nursing homes through invitational letters, telephone calls, and visits. The mean (⫾ SD) number of beds in participating facilities was 149 ⫾ 62 (range 72 to 333). Eleven of the study facilities were proprietary; the remainder were voluntary nonprofit nursing homes. All participating facilities were certified by both Medicare and Medicaid.
Study Sample The 18 study nursing homes were enrolled from March 14, 1997, through May 15, 1997. Subjects included all long-stay residents during the 12-month period following each facility’s enrollment. Residents admitted for short-term care (eg, subacute care or rehabilitation following hospitalization) were excluded. The study was approved by the institutional review board of the University of Massachusetts Medical School. Because of potential liability concerns, details of individual cases are not provided, and participating facilities are not identified.
Case-Finding Definitions and Classification of Events At 6-week intervals, systematic chart reviews were performed by 2 trained nurse investigators and 1 pharmacist investigator for each eligible nursing home resident. These investigators examined the nursing home records for possible drug-related incidents, such as new symptoms or events that might represent an adverse drug event, changes in medication regimens (including acute discontinuations or initiations of medications that might be used to treat a drug-induced effect), abnormal laboratory values, and all emergency room transfers and hospitalizations. In addition, the investigators encouraged nursing home staff to inform them of any possible drugrelated incidents, as well as incidents that may not have seemed to be a drug-related event (eg, new falls). Whenever possible, nursing home staff were interviewed to ascertain information about incidents beyond that available in the nursing home resident records. The nurse and pharmacist investigators had no direct contact with nursing home residents. 88
August 1, 2000
THE AMERICAN JOURNAL OF MEDICINE威
Volume 109
After an extensive training period, between-case investigator reliability for identifying relevant incidents in nursing home records was assessed through independent review of the same 10 medical records by each of the 3 case investigators. All 3 identified the same incident in 9 of the 10 nursing home records; for the remaining record, 2 of the case investigators identified the incident correctly. Bimonthly refresher sessions were held with the 3 case investigators throughout the study to reinforce consistent identification and record abstraction procedures. The primary outcome of the study was an adverse drug event, defined as an injury resulting from the use of a drug. Adverse drug events may have resulted from medication errors (ie, errors in prescribing, dispensing, administration, and monitoring) or from adverse drug reactions in which no error was involved. We also identified potential adverse drug events (“near misses”), defined as errors that had the capacity to cause injury but failed to do so, either by chance or because they were intercepted (eg, a nursing home resident who received an antibiotic despite a known allergy, but who did not have a reaction; or an excessively high international normalized ratio ([INR] ⱖ4) that occurred as the result of a drug interaction with warfarin when no effort was made to prevent the event by increasing the frequency of monitoring or reducing the warfarin dose). We did not include incidents in which an error was made but was judged to have minimal potential for injury. By definition, all potential adverse drug events were considered preventable. The definitions for adverse drug events and potential adverse drug events are consistent with definitions used previously (12–15). All possible drug-related incidents were evaluated independently by 2 physician-reviewers (2 internist-geriatricians [JHG, JA] and 2 general internists [DWB, DM]). The reviewers classified these incidents using a structured implicit review according to the following criteria: whether an adverse drug event or potential adverse drug event was present, the severity of the event, whether the event was preventable, and the effects on the resident. In determining whether an adverse drug event had occurred, the physician-reviewers considered the temporal relation between the drug exposure and the event, as well as whether the event reflected a known effect of the drug. Suboptimal prescribing choices (16 –18), such as the use of long-acting hypoglycemic agents, highly anticholinergic tricyclic antidepressants, or long elimination halflife benzodiazepines, were not considered unless an adverse drug event had occurred relating to that choice. Categories of severity were significant, serious, lifethreatening, and fatal (19). Examples of adverse drug events categorized as significant included a nonurticarial skin rash, a fall without associated fractures, hemorrhage not requiring transfusion or hospitalization, and oversedation. Examples of events categorized as serious included urticaria, falls with associated fractures, hemor-
Adverse Drug Events in Nursing Homes/Gurwitz et al
Table 1. Rates and Severity of Adverse Drug Events and Potential Adverse Drug Events among 2,916 Residents of 18 Nursing Homes Observed for a Mean of 9.9 Months
Type of Drug-Related Event
Number (Percent)
Rate per 100 Resident-Months (95% Confidence Interval)
Category of Severity Fatal
Life-Threatening
Serious
Significant
Number (Percent) Adverse drug events Preventable Nonpreventable Potential adverse drug events*
546 276 (50) 270 (50) 188
1.89 (1.74–2.05) 0.96 (0.85–1.07) 0.94 (0.83–1.05) 0.65 (0.56–0.75)
1 (0.2) 1 (0.4) 0 —
31 (6) 25 (9) 6 (2) 16 (9)
206 (38) 145 (52) 61 (23) 149 (79)
308 (56) 105 (38) 203 (75) 23 (12)
* All of these were considered preventable by definition.
rhage requiring transfusion or hospitalization but without hypotension, and delirium. Examples of lifethreatening events included hemorrhage with associated hypotension, hypoglycemic encephalopathy, and liver failure. Adverse drug events were considered to be preventable if they were due to an error and were preventable by any means then available. Categories of preventability included definitely preventable, probably preventable, probably not preventable, and definitely not preventable (20); results were collapsed into preventable and nonpreventable categories in the analyses. The effects of an adverse drug event on a nursing home resident were characterized by duration (ⱕ1 day of symptoms, ⬎1 day of symptoms) and by disability (nonpermanent, permanent). Disability was defined as a decrease in functional status from its previous level (21) and referred to a new deficit in activities of daily living as documented in the nursing home records (bathing, dressing, toileting, transferring, continence, or feeding). We also classified the stages of pharmaceutical care as ordering (essentially all by physicians), transcribing (performed by a secretary or nurse, depending on the nursing home unit and the time of day), dispensing (by the pharmacy vendor), administration (by nurses), and monitoring (by physicians, nurses, and the consultant pharmacist to the nursing home). Periodic review of medication regimens of residents of US long-term care facilities by consultant pharmacists has been federally mandated since 1974 (22). When there were disagreements that affected the classification of an incident by the 2 physician-reviewers (eg, 1 reviewer scored it as preventable, but the other did not), or about the presence of an event, its severity, or its preventability, the reviewers met and reached consensus. Consensus was reached in all instances where there was initial disagreement. For a random sample of 100 drugrelated incidents that were evaluated independently by 2 physician-reviewers, interrater reliabilities for initial judgments were calculated using the statistic. For judgments about the presence of an adverse drug event, the value was 0.80; for preventability, the value was 0.73; and for severity, the value was 0.62.
Prevalence of Medication Use To estimate the underlying prevalence of use of selected medication categories in the overall sample of nursing home residents, we collected drug use data from the medication administration records of a representative sample of 410 residents who were randomly selected from each facility in proportion to the number of residents identified as having experienced an adverse drug event in that facility. These residents had not experienced an adverse drug event at the time of data collection on medication use.
Analysis To determine crude rates of events, the numbers of adverse drug events or potential adverse drug events were divided by the total number of nursing home residentmonths, which was estimated by obtaining census data for all eligible subjects in each study nursing home at 6-week intervals throughout the course of the project; absences from the nursing home (eg, for hospitalization) were accounted for. Ninety-five percent confidence intervals (CI) were calculated for these estimates (23). Comparisons between categorical variables were performed using the chi-square test. A P value ⬍ 0.05 was considered significant. Analyses were performed using SAS, version 6 (SAS Institute, Inc., Cary, North Carolina).
RESULTS During the study period, 2916 nursing home residents (mean [⫾ SD] age of 84 ⫾ 9 years; 77% women) contributed 28,839 resident-months of observation in the 18 participating nursing homes. The nurse and pharmacist investigators identified 979 possible drug-related incidents during the study, of which 245 (25%) were not considered to be adverse drug events or potential adverse drug events by the physician-reviewers. Of the 734 remaining incidents, 546 (74%) were classified as adverse drug events and 188 (26%) were classified as potential adverse drug events (Table 1). Seventeen percent (n ⫽ 93) of the adverse drug events were identified by nursing August 1, 2000
THE AMERICAN JOURNAL OF MEDICINE威
Volume 109 89
Adverse Drug Events in Nursing Homes/Gurwitz et al
Table 2. Effects of Adverse Drug Events on Nursing Home Residents Duration of Symptoms Type of Drug-Related Event
ⱕ1 day
Type of Disability
⬎1 day
Nonpermanent
Permanent
Number (Percent) Adverse drug events (n ⫽ 546) Preventable (n ⫽ 276) Nonpreventable (n ⫽ 270)
92 (17) 33 (12) 59 (22)
416 (76) 216 (78) 200 (74)
staff reporting and the remainder through review of nursing home records. Only 4% (n ⫽ 7) of the potential adverse drug events were identified through nursing home staff reporting. Of the 546 adverse drug events, about half (n ⫽ 276) were considered preventable. (By definition, all 188 potential adverse drug events were considered preventable.) The overall rate of adverse drug events was 1.89 per 100 resident-months, with a rate of 0.96 preventable adverse drug events per 100 resident-months. The rate of potential adverse drug events was 0.65 per 100 residentmonths. One resident suffered a fatal adverse drug event (methotrexate-induced hepatotoxicity) during the study period (Table 1); this event was characterized as preventable. Of the 238 fatal, life-threatening, or serious adverse drug events, 72% (n ⫽ 171) were deemed preventable, compared with 34% (n ⫽ 105) of the 308 significant adverse drug events. Overall, more severe adverse drug events were more likely to be preventable (relative risk ⫽ 2.1, 95% CI: 1.8 to 2.5, P ⬍ 0.001). A greater percentage of potential adverse drug events was considered life-threatening or serious as compared with adverse drug events (88% vs 44%). Only a small minority (17%) of adverse drug events resulted in symptoms of ⱕ1 day in duration (Table 2). Thirty-eight events resulted in disability to the nursing home resident; in 8 cases the disability was permanent. Preventable adverse drug events were more likely to result in disability as compared with nonpreventable adverse drug events (relative risk ⫽ 2.4, 95% CI: 1.2 to 4.7, P ⬍ 0.01). The 546 adverse drug events were associated with a wide variety of different drug classes (Table 3). Antipsychotics were the most frequently implicated agents associated with these events, followed by antibiotics, antidepressants, and sedatives/hypnotics. Psychoactive medications (antipsychotics, antidepressants, and sedatives/hypnotics) and anticoagulants were the most common drug categories associated with preventable adverse drug events. Psychoactive drugs were commonly used by nursing home residents; more than 36% of residents used antidepressants, 24% used sedatives/hypnotics, and 17% used antipsychotic medications (Table 4). 90
August 1, 2000
THE AMERICAN JOURNAL OF MEDICINE威
Volume 109
30 (6) 24 (9) 6 (2)
8 (1) 3 (1) 5 (2)
Antibiotics were associated with fewer than 5% of all preventable adverse drug events, but they were responsible for 36% of nonpreventable adverse drug events (P ⬍ 0.0001). The vast majority of antibiotic-associated adverse drug events were rashes and confirmed Clostridium difficile diarrhea. The distribution of drug classes associated with potential adverse drug events was different from that of adverse drug events (Table 3). Almost 80% of all potential adverse drug events were associated with the use of warfarin, primarily the development of excessively high INR values (ⱖ4) due to errors in anticoagulation management (the mean INR in these patients was 6.1 [range 4.0 to 15.6]). Prescribers frequently made no modifications in the warfarin dosing scheme or the usual frequency of monitoring when drugs with well-established interactions with warfarin were prescribed. Consistently subtherapeutic INR levels (at least three consecutive values ⬍1.5 during several weeks of treatment) were also considered potential adverse drug events when the appropriate therapeutic range for the specified indication was 2 to 3. The mean INR in these patients was 1.2 (range 0.8 to 1.4). Neuropsychiatric events (eg, oversedation, confusion, hallucinations, delirium) were the most common types of both preventable and nonpreventable adverse drug events (Table 5). Falls were involved in 20% of preventable adverse drug events but only 4% of nonpreventable events. Similarly, bleeding was responsible for 14% of preventable adverse drug events compared with only 6% of nonpreventable adverse drug events. Dermatologic events were rarely deemed preventable. Essentially all infections were deemed nonpreventable, as this category almost entirely consisted of C. difficile diarrhea associated with previous antibiotic therapy. Gastrointestinal events, including nausea, vomiting, diarrhea, and abdominal pain, were involved in 12% of all adverse drug events. Among the 464 preventable events, errors occurred most commonly in the ordering (n ⫽ 315, 68%) and monitoring (n ⫽ 325, 70%) stages of care. Errors accounting for preventable events were rarely identified in the transcription (0.7%), dispensing (0.5%), or administration (3%) stages. Among the 315 ordering errors, wrong dose (eg, excessive dose for an elderly patient) was the most common (n ⫽ 200, 63%), followed by the pre-
Adverse Drug Events in Nursing Homes/Gurwitz et al
Table 3. Frequency of Adverse Drug Events and Potential Adverse Drug Events by Drug Class*
Drug Class
Adverse Drug Events (n ⫽ 546)
Preventable Adverse Drug Events (n ⫽ 276)
Nonpreventable Adverse Drug Events (n ⫽ 270)
Potential Adverse Drug Events (n ⫽ 188)
Number (Percent) Antipsychotics Antibiotics/antiinfectives Antidepressants Sedatives/hypnotics Anticoagulants Antiseizure Cardiovascular Hypoglycemics Nonopioid analgesics Opioids Antiparkinsonians Gastrointestinal Diuretics Antigout Muscle relaxants Alzheimer’s disease Nutrients/supplements Ophthalmics Antineoplastics Respiratory Osteoporosis Steroids Nonophthalmic topicals Antihistamines Antihyperlipidemics Miscellaneous
125 (23) 109 (20) 68 (13) 68 (13) 51 (9) 47 (9) 35 (6) 27 (5) 22 (4) 15 (3) 12 (2) 11 (2) 10 (2) 6 (1) 6 (1) 4 (0.7) 3 (0.5) 3 (0.5) 2 (0.4) 2 (0.4) 2 (0.4) 1 (0.2) 1 (0.2) 0 0 2 (0.4)
72 (26) 13 (5) 50 (18) 49 (18) 37 (13) 27 (10) 25 (9) 14 (5) 13 (5) 7 (3) 7 (3) 6 (2) 9 (3) 4 (1) 4 (1) 3 (1) 1 (0.4) 1 (0.4) 2 (0.7) 1 (0.4) 0 (0.0) 1 (0.4) 0 0 0 0
53 (20) 96 (36) 18 (7) 19 (7) 14 (5) 20 (7) 10 (4) 13 (5) 9 (3) 8 (3) 5 (2) 5 (2) 1 (0.4) 2 (0.7) 2 (0.7) 1 (0.4) 2 (0.7) 2 (1.0) 0 1 (0.4) 2 (0.7) 0 1 (0.4) 0 0 2 (0.7)
1 (0.5) 13 (7) 0 2 (1) 150 (80) 0 7 (4) 6 (3) 1 (0.5) 0 0 1 (0.5) 5 (3) 0 0 0 1 (0.5) 0 0 0 0 0 1 (0.5) 2 (1) 1 (0.5) 0
* Some events were associated with drugs in more than one category; thus, frequencies in each column sum to greater than the total number of events.
scription of a drug for which there was a well-established, clinically important interaction with another drug (eg, trimethoprim/sulfamethoxazole-warfarin interaction; n ⫽ 68, 22%). Wrong choice errors were the third most common type of ordering error (n ⫽ 28, 9%). Examples of such errors included the use of an antidepressant with anticholinergic properties for a resident with identifiable side effects to the agent, the use of a long-acting benzodiazepine for a resident with drug-related oversedation, and the use of an antipsychotic medication to treat hallucinations due to levodopa/carbidopa therapy in a patient with Parkinson’s disease. Monitoring errors generally referred to inadequate laboratory monitoring of drug therapies or to a delayed response, or failure to respond, to signs or symptoms of drug toxicity or laboratory evidence of toxicity. Among the 325 monitoring errors, failure to act on available information relating to clinical findings or laboratory results (n ⫽ 133, 41%) and inadequate monitoring (n ⫽ 83, 26%) were the most common errors. Examples of moni-
toring errors included an inadequate frequency of monitoring of warfarin leading to an elevated INR value or failure to respond to symptoms suggestive of digoxin toxicity (eg, nausea, vomiting, and anorexia).
DISCUSSION We found that adverse drug events occurred commonly among nursing home residents and that more than half were preventable. Serious, life-threatening, and fatal adverse drug events were more likely to be preventable than less severe events. Most errors associated with preventable events occurred at the ordering and monitoring stages. Although it is difficult to compare the findings of the present study on adverse drug events in nursing home residents with studies performed in other clinical settings involving different groups of patients (24), some comparisons are of interest. Bates et al (12) identified adverse August 1, 2000
THE AMERICAN JOURNAL OF MEDICINE威
Volume 109 91
Adverse Drug Events in Nursing Homes/Gurwitz et al
Table 4. Use of Selected Drug Classes among a Representative Sample of Nursing Home Residents (n ⫽ 410) Drug Class
Number (Percent)
Gastrointestinal Nutrients/supplements Nonopioid analgesics Cardiovascular Antidepressants Diuretics Sedatives/hypnotics Antipsychotics Antibiotics/antiinfectives Antiseizure Ophthalmics Anticoagulants Hypoglycemics Respiratory Steroids Opioids Antiparkinsonians Muscle relaxants Osteoporosis Antigout Antihistamines Antihyperlipidemics Alzheimer’s disease Antineoplastics Nonophthalmic topicals
308 (75) 280 (68) 258 (63) 201 (49) 149 (36) 143 (35) 97 (24) 71 (17) 57 (14) 53 (13) 49 (12) 47 (12) 42 (10) 32 (8) 31 (8) 28 (7) 28 (7) 14 (3) 12 (3) 11 (3) 10 (2) 8 (2) 5 (1) 5 (1) 2 (0.5)
drug events occurring during 4,031 adult admissions to two Boston tertiary care hospitals during a 6-month period. Of the 247 adverse drug events, 1% were fatal, 12% were life-threatening, 30% were serious, and 57% were significant. Twenty-eight percent were judged preventable. Of the life-threatening and serious adverse drug events, 42% were judged preventable, compared with 18% of significant adverse drug events. In the nursing home setting, more than half of all adverse drug events were preventable; similar to the situation in the hospital setting, more serious events were more likely to be judged preventable. In this study, psychoactive drugs (antipsychotics, antidepressants, and sedatives/hypnotics) and anticoagulants were the most common drug categories associated with preventable adverse drug events. Suboptimal prescribing of psychoactive medications in nursing homes has long been an issue of concern (25), leading to federal regulations about their use in this setting (26). Various interventions have been developed and tested to reduce the risks of psychoactive drug use in nursing homes (27). Academic detailing to educate physicians and nursing home staff about the use of psychoactive medications in geriatric patients has had some success in improving the quality of drug prescribing, specifically for antipsychotic medications and long-acting benzodiazepines (28,29). 92
August 1, 2000
Table 5. Frequency of Adverse Drug Events by Type*
THE AMERICAN JOURNAL OF MEDICINE威
Volume 109
Type
Adverse Drug Events (n ⫽ 546)
Preventable (n ⫽ 276)
Number (Percent) Neuropsychiatric Falls Gastrointestinal Dermatologic/allergic Hemorrhage Extrapyramidal symptoms/ tardive dyskinesia Infection Metabolic/endocrine Anorexia/weight loss Ataxia/difficulty with gait Cardiovascular Electrolyte/fluid balance abnormality Syncope/dizziness Functional decline† Respiratory Anticholinergic‡ Renal Hematologic Hepatic
150 (27) 67 (12) 65 (12) 59 (11) 57 (10) 52 (10)
83 (30) 55 (20) 30 (11) 7 (3) 40 (14) 19 (7)
34 (6) 27 (5) 20 (4) 18 (3) 15 (3) 9 (2)
1 (0.4) 14 (5) 14 (5) 9 (3) 10 (4) 5 (2)
8 (1) 7 (1) 3 (0.5) 3 (0.5) 3 (0.5) 2 (0.4) 1 (0.2)
5 (2) 6 (2) 3 (1) 2 (0.7) 1 (0.4) 2 (0.7) 1 (0.4)
* Adverse drug events could manifest as more than one type. † Adverse drug event manifested only as decline in activities of daily living without any other more specific type of event. Other types of events may have been associated with functional decline. ‡ Anticholinergic effects include dry mouth, dry eyes, urinary retention, and constipation.
Anticoagulants were also a common cause of adverse drug events and were the leading cause of potential adverse drug events. Concerns have been raised about the quality of anticoagulant use in nursing homes (30). A more systematic approach to decision-making about the use of warfarin for stroke prevention in the frail elderly is required, as is a more consistent approach to the management of therapy. More widespread use of specialized anticoagulation clinics to provide coordinated care may offer an option to improve the effectiveness and safety of warfarin therapy for these high-risk patients. The nursing homes that participated in this study were somewhat different from the average US nursing home. The average number of beds per nursing home among the study facilities was higher than the national average (149 vs 106), due to the exclusion of facilities with fewer than 50 beds. Nationally, 66% of nursing homes are proprietary, compared with 61% of the study nursing homes. Of all nursing homes in the United States, 70% are certified by both Medicare and Medicaid (1), compared with 100% of the participating study facilities. Nationally, 75% of nursing home residents are women and 83% of all residents are aged 75 years or older (31), compared with
Adverse Drug Events in Nursing Homes/Gurwitz et al
77% who were women and 89% who were 75 years of age or more in this study. If the findings of the present study apply to all US nursing homes, then about 24 adverse drug events and 8 potential adverse drug events should be identifiable each year in an average facility, many of which should be preventable. Thus, about 350,000 adverse drug events— more than half of which are preventable— occur each year in the 1.55 million residents of US nursing homes. There are almost 20,000 fatal or life-threatening adverse drug events per year, of which 80% are preventable. These estimates are likely to be conservative. To ascertain information on drug-related incidents, we relied mainly on information contained in nursing home records, which can be quite limited compared with other clinical settings. Without adequate documentation of relevant information in the medical record, possible drugrelated incidents could have been missed by the case investigators, and inadequate information would not have been available for the physician-reviewers to classify incidents as adverse drug events. There was no direct assessment of nursing home residents in this study, and such assessment might have allowed the identification of additional events. Furthermore, the nursing homes that participated in this project volunteered to do so, perhaps suggesting an increased sensitivity to the problem of adverse drug events compared with the average long-term care facility. How should the findings of this study be applied to improve the quality of care for residents of nursing homes? One approach is to identify and admonish health care providers who make errors that lead to adverse drug events, but that approach ignores failures in the design of systems of care that often contribute to the occurrence of medical errors, as well as the injuries that result from some of those errors (32–35). Enhanced surveillance and reporting systems for adverse drug events in nursing homes are required, and educational efforts about the optimal use of drug therapies in frail elderly patients are essential. However, as Leape et al (34) have concluded about the occurrence of serious medication errors in the hospital setting, preventive efforts that focus solely on the individual provider or that rely on inspection alone have limited effects: “analysis and the correction of underlying systems faults is much more likely to result in enduring changes and significant error reduction.” Ordering and monitoring errors in the nursing home setting may be particularly amenable to prevention strategies utilizing systems-based approaches. The benefits of this approach in the hospital setting through the use of computerized order entry have recently been reported (36); such a system could be designed to focus on ordering and monitoring issues in the nursing home. Successes in the hospital setting should pave the way for similar efforts in nursing
homes aimed at reducing drug-related injuries and disability and improving the quality of care provided to this vulnerable population (37).
ACKNOWLEDGMENT The authors thank the nursing homes participating in this study for their commitment and dedication to improving the quality of care provided to nursing home residents. They also thank Mary Ellen Stansky and Jackie Cernieux, MPH, for their assistance with technical aspects relating to the study and Bessie Petropoulos for assistance with manuscript preparation.
REFERENCES 1. Strahan GW. An Overview of Nursing Homes and Their Current Residents: Data from the 1995 National Nursing Homes Survey. Advance Data from Vital and Health Statistics, No. 280. Hyattsville, Md: National Center for Health Statistics, 1997. 2. Bernabei R, Gambassi G, Lapane K, et al. Characteristics of the SAGE database: a new resource for research on outcomes in long– term care. J Gerontol Series A, Biological Sciences, Medical Sciences. 1999;54:M25–M33. 3. Shorr RI, Fought RL, Ray WA. Changes in antipyschotic drug use in nursing homes during implementation of the OBRA-87 regulations. JAMA. 1994;271:358 –362. 4. Beers M, Avorn J, Soumerai SB, et al. Psychoactive medication use in intermediate-care facility residents. JAMA. 1988;260:3016 – 3020. 5. Pickering TD, Gurwitz JH, Zaleznick D, et al. The appropriateness of oral fluoroquinolone prescribing in the long-term care setting. J Am Geriatr Soc. 1994;42:28 –32. 6. Gurwitz JH, Noonan JP, Soumerai SB. Reducing the use of H2receptor antagonists in the long-term care setting. J Am Geriatr Soc. 1992;40:359 –364. 7. Ray WA, Federspiel CF, Schaffner W. A study of antipsychotic drug use in nursing homes: epidemiologic evidence suggesting misuse. Am J Public Health. 1980;70:485– 491. 8. Svarstad BL, Mount JK. Nursing home resources and tranquilizer use among the institutionalized elderly. J Am Geriatr Soc. 1991;39: 869 – 875. 9. Beers MH, Ouslander JG, Fingold SF, et al. Inappropriate medication prescribing in skilled-nursing facilities. Ann Intern Med. 1992; 117:684 – 689. 10. Gerety MB, Cornell JE, Plichta DT, Eimer M. Adverse events related to drugs and drug withdrawal in nursing home residents. J Am Geriatr Soc. 1993;41:1326 –1332. 11. Gurwitz JH, Sanchez-Cross MT, Eckler MA, Matulis J. The epidemiology of adverse and unexpected events in the long-term care setting. J Am Geriatr Soc. 1994;42:33–38. 12. 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. 13. 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. 14. Bates DW, Leape LL, Cullen DJ, et al. Effect of computerized physician order entry and a team intervention on prevention of serious medication errors. JAMA. 1998;280:1311–1216. 15. Leape LL, Cullen DJ, Clapp MD, et al. Pharmacist participation on physician rounds and adverse drug events in the intensive care unit. JAMA. 1999;282:267–270. 16. Beers MH, Ouslander JG, Rollingher I, et al. Explicit criteria for August 1, 2000
THE AMERICAN JOURNAL OF MEDICINE威
Volume 109 93
Adverse Drug Events in Nursing Homes/Gurwitz et al
17.
18. 19.
20. 21.
22.
23. 24.
25. 26. 27.
94
determining inappropriate medication use in nursing home residents. Arch Intern Med. 1991;151:1825–1832. Beers MH. Explicit criteria for determining potentially inappropriate medication use by the elderly. An update. Arch Intern Med. 1997;157:1531–1536. Gurwitz JH. Suboptimal medication use in the elderly. The tip of the iceberg. JAMA. 1994;272:316 –317. Folli HL, Poole RI, Benitz WE, Russo JC. Medication error prevention by clinical pharmacists in two children’s hospitals. Pediatrics. 1987;79:781–722. Dubois RW, Brook RH. Preventable deaths: who, how often, and why. Ann Intern Med. 1998;190:582–589. Katz S, Ford AB, Moskowitz RW, et al. Studies of illness in the aged: the index of ADL. A standardized measure of biological and psychosocial function. JAMA. 1963;185:914 –919. Department of Health, Education, and Welfare. Skilled nursing facility standards for certification and participation in Medicare and Medicaid programs. Fed Regist. 1974;39:2238. Rosner B. Fundamentals of Biostatistics, 3rd ed. Boston, Mass: PWSKent; 1990. Phillips DP, Christenfeld, N, Glynn LM. Increase in U.S. medication-error deaths between 1983 and 1993. Lancet. 1998;351:643– 644, 1983. Rango N. Nursing home care in the United States. N Engl J Med. 1982;307:883– 889. Elon R, Paulson LG. The impact of OBRA on medical practice within nursing facilities. J Am Geriatr Soc. 1992;40:958 –963. Ray WA, Taylor JA, Meador KG, et al. A randomized trial of a
August 1, 2000
THE AMERICAN JOURNAL OF MEDICINE威
Volume 109
28.
29.
30.
31.
32. 33. 34. 35. 36.
37.
consultation service to reduce falls in nursing homes. JAMA. 1997; 278:557–562. Avorn J, Soumerai SB, Everitt DE, et al. A randomized trial of a program to reduce the use of psychoactive drugs in nursing homes. N Engl J Med. 1992;327:168 –173. Ray WA, Taylor JA, Meador KG, et al. Reducing antipsychotic drug use in nursing homes. A controlled trial of provider education. Arch Intern Med. 1993;153:713–721. Gurwitz JH, Monette J, Rochon PA, et al. Atrial fibrillation and stroke prevention with warfarin in the long-term care setting. Arch Intern Med. 1997;157:978 –984. Dey AN. Characteristics of Elderly Nursing Home Residents: Data from the 1995 National Nursing Home Survey. Advance Data from Vital and Health Statistics, No. 289. Hyattsville, Md: National Center for Health Statistics; 1997. Berwick DM. Continuous improvement as an ideal in health care. N Engl J Med. 1989;320:53–56. Leape LL. Error in medicine. JAMA. 1994;272:1851–1857. Leape LL, Bates DW, Cullen DJ, et al. Systems analysis of adverse drug events. JAMA. 1995;274:35– 43. Horton R. The uses of error. Lancet. 1999;353:422– 423. Bates DW, Leape LL, Cullen DJ, et al. Effect of computerized physician order entry and a team intervention on prevention of serious medication errors. JAMA. 1998;280:1311–1316. Kohn LT, Corrigan JM, Donaldson MS, eds. To Err is Human: Building a Safer Health System. Washington, DC: National Academy Press; 1999.