Advertisement

The incidence of adverse drug events in two large academic long-term care facilities

      Purpose

      To assess the incidence of and risk factors for adverse drug events in the long-term care setting.

      Methods

      We performed a cohort study of all long-stay residents of two academic long-term care facilities over a period of up to 9 months during 2000 to 2001. We assessed the number of adverse drug events, the severity of events (classified as less serious, serious, life threatening, or fatal), and whether the events were preventable. A case-control study was nested within the prospective study to identify resident-level risk factors for the occurrence of adverse drug events.

      Results

      There were 815 adverse drug events, of which 42% were judged preventable. The overall rate of adverse drug events was 9.8 per 100 resident-months, with a rate of 4.1 preventable adverse drug events per 100 resident-months. Errors associated with preventable events occurred most often at the stages of ordering and monitoring. Residents taking medications in several drug categories were at increased risk of a preventable adverse event. In multivariate analyses, the adjusted odds ratio was 3.4 (95% confidence interval [CI]: 2.0 to 5.9) for those taking antipsychotic agents, 2.8 (95% CI: 1.6 to 4.7) for those taking anticoagulants, 2.2 (95% CI: 1.2 to 4.0) for those taking diuretics, and 2.0 (95% CI: 1.1 to 3.7) for those taking antiepileptics.

      Conclusion

      Our findings reinforce the need for a special focus on the ordering and monitoring stages of pharmaceutical care for preventing adverse drug events in the long-term care setting. Patients taking antipsychotic agents, anticoagulants, diuretics, and antiepileptics are at increased risk.

      Keywords

      There are more than 1.6 million residents of long-term care facilities in the United States. Medication use is high in these settings, and it is widely acknowledged that these patients are at special risk of adverse drug events. Although few studies have systematically examined the incidence of adverse drug events among residents of long-term care facilities, one study, conducted in a sample of community-based nursing homes, found that adverse drug events occur commonly, and that at least half may be preventable.
      • Gurwitz J.H.
      • Field T.S.
      • Avorn J.
      • et al.
      Incidence and preventability of adverse drug events in nursing homes.
      If findings from that study are extrapolated to all U.S. nursing homes, then approximately 350 000 adverse drug events occur annually among this patient population, including 20 000 fatal or life-threatening events.
      The purpose of the present study was to determine the rate of adverse drug events in two large academic long-term care facilities, as well as to identify risk factors for such events in this population.

      Methods

      Study setting

      This study was conducted in two large academic long-term care facilities in Connecticut and Ontario, Canada. The two facilities have a total of 1229 beds. Patients residing in areas of the facilities related to short-term care (eg, subacute care, hospital-level care, or rehabilitation) were not included.
      The study was conducted over an 8-month period in one facility (December 2000 through July 2001) and a 9-month period in the other facility (December 2000 through August 2001). The study was approved by the institutional review boards of the University of Massachusetts Medical School and of the participating facilities.

      Case-finding definitions and classification of events

      The study was limited to drug-related incidents occurring in the long-term care setting. Events were detected through review of medical records in monthly segments performed by trained pharmacist-investigators for each eligible resident of a long-term care facility. These investigators searched 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 event), abnormal laboratory values, and all emergency department transfers and hospitalizations. Medical records were also targeted for review based on information derived from selected computer-generated signals, including abnormal serum drug levels, abnormal laboratory results, and the use of medications considered to be antidotes for adverse drug effects (Appendix). Administrative incident reports generated within each facility were also reviewed for any indication of a drug-related incident.

      Outcome measures

      The primary outcome was an adverse drug event, defined as an injury resulting from the use of a drug. This definition is consistent with definitions used in previous studies.
      • Gurwitz J.H.
      • Field T.S.
      • Avorn J.
      • et al.
      Incidence and preventability of adverse drug events in nursing homes.
      • Gurwitz J.H.
      • Field T.S.
      • Harrold L.R.
      • et al.
      Incidence and preventability of adverse drug events among older persons in the ambulatory setting.
      • Bates D.W.
      • Cullen D.J.
      • Laird N.
      • et al.
      Incidence of adverse drug events and potential adverse drug events implications for prevention. ADE Prevention Study Group.
      • Leape L.L.
      • Bates D.W.
      • Cullen D.J.
      • et al.
      Systems analysis of adverse drug events.
      • Leape L.L.
      • Cullen D.J.
      • Clapp M.D.
      • et al.
      Pharmacist participation on physician rounds and adverse drug events in the intensive care unit.
      Adverse drug events may have resulted from medication errors (eg, errors in ordering, dispensing, administration, monitoring) or from adverse drug reactions in which there was no error.
      After an extensive training period, the between-pharmacist-investigator reliability for identifying relevant incidents in medical records was assessed through independent review of the same 10 medical records by each of the 2 investigators. Each of the 2 pharmacist-investigators identified the same incident in the 10 medical records; 1 investigator identified an additional incident in one record that had not been prespecified as an incident warranting review.
      The possible drug-related incidents were presented by a pharmacist-investigator to pairs of physician-reviewers (JHG, JJ, PR, LRH) who independently classified incidents using structured implicit review according to the following criteria: whether an adverse drug event was present, the severity of the event, whether the event was preventable, and the effects of the event on the patient. 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. This structured implicit review process has been used in prior studies of adverse drug events in various clinical settings.
      • Gurwitz J.H.
      • Field T.S.
      • Avorn J.
      • et al.
      Incidence and preventability of adverse drug events in nursing homes.
      • Gurwitz J.H.
      • Field T.S.
      • Harrold L.R.
      • et al.
      Incidence and preventability of adverse drug events among older persons in the ambulatory setting.
      • Bates D.W.
      • Cullen D.J.
      • Laird N.
      • et al.
      Incidence of adverse drug events and potential adverse drug events implications for prevention. ADE Prevention Study Group.
      • Leape L.L.
      • Cullen D.J.
      • Clapp M.D.
      • et al.
      Pharmacist participation on physician rounds and adverse drug events in the intensive care unit.
      • Kaushal R.
      • Bates D.W.
      • Landrigan C.
      • et al.
      Medications errors and adverse drug events in pediatric in-patients.
      • Bates D.W.
      • Leape L.L.
      • Cullen D.J.
      • et al.
      Effect of computerized physician order entry and a team intervention on prevention on serious medication errors.
      • Bates D.W.
      • Spell N.
      • Cullen J.J.
      • et al.
      The costs of adverse drug events in hospitalized patients.
      The severity of adverse drug events was categorized as less serious, serious, life threatening, or fatal. Events categorized as less serious included a nonurticarial skin rash, falls without associated fracture, hemorrhage not requiring transfusion or hospitalization, and oversedation. Examples of events categorized as serious included urticaria, falls with associated fracture, hemorrhage requiring transfusion or hospitalization but without hypotension, and delirium. Examples of life-threatening events included hemorrhage with associated hypotension, hypoglycemic encephalopathy, and acute renal failure. Adverse drug events were considered to be preventable if they were judged to be due to an error and were preventable by any means available.
      • Bates D.W.
      • Cullen D.J.
      • Laird N.
      • et al.
      Incidence of adverse drug events and potential adverse drug events implications for prevention. ADE Prevention Study Group.
      “Preventability” was categorized as preventable, probably preventable, probably not preventable, or definitely not preventable; results were collapsed into the categories of preventable (preventable and probably preventable) and nonpreventable (probably not preventable and definitely not preventable) in the analysis. The effects of adverse drug events on patients were categorized as abnormal laboratory results without signs and symptoms, symptoms lasting <1 day, symptoms lasting ≥1 day, nonpermanent disability, permanent disability, and death. Physician-reviewers characterized an event as causing permanent disability based on evidence that a drug-induced injury had caused physical disability or deficits in functioning.
      • Freedman V.A.
      • Martin L.G.
      • Schoeni R.F.
      Recent trends in disability and functioning among older adults in the United States a systematic review.
      The stages of pharmaceutical care during which an error leading to a preventable adverse drug event occurred were ordering, dispensing, administration, and monitoring. Monitoring stage errors included inadequate laboratory monitoring of drug therapies, or a delayed response or failure to respond to signs or symptoms or laboratory evidence of drug toxicity. For a single adverse drug event, it was possible to identify errors at more than one stage of pharmaceutical care or to identify more than one error within a single stage of care.
      When the physician-reviewers disagreed on the classification of an incident regarding the presence of an adverse drug event, its severity, or its preventability, they met and reached consensus; consensus was reached in all instances where there was initial disagreement. We compared the initial assessments of the physician-reviewers and calculated inter-rater reliability using the κ statistic, with κ = 0.65 for judgments regarding the presence of an adverse drug event, 0.49 for preventability, and 0.40 for severity. A κ score of 0.4 to 0.6 reflects “moderate agreement,” 0.6 to 0.8 reflects “substantial” agreement, and 0.8 to 1.0 is considered “almost perfect.”
      • Sackett D.L.
      • Haynes R.B.
      • Guyatt G.H.
      • Tugwell P.

      Case-control study

      For each resident with an adverse drug event, we focused on their first event during the study period and randomly selected a control among residents who were present in the nursing home on the day of that event. All residents who had not yet experienced an adverse drug event at the time of the event were eligible to serve as controls. Information on potential risk factors for cases and controls was collected through medical record review using standardized forms. Data included age, sex, comorbid conditions, and medication use as of the date of the event. Burden of illness was assessed using the Charlson Comorbidity Index.
      • Charlson M.E.
      • Pompei P.
      • Ales K.L.
      • MacKenzie C.R.
      A new method of classifying prognostic comorbidity in longitudinal studies development and validation.
      Information on medication use at the time of the event included the number of regularly scheduled medications and use of any drug within selected drug categories.
      We analyzed preventable adverse drug events separately. For this portion of the study, cases included all subjects who experienced a preventable event. Potential risk factor information was collected as of the date of the first preventable adverse drug event for cases and controls.

      Statistical analysis

      To determine crude rates of events, the number of adverse drug events was divided by the total number of long-term care resident-months, which was estimated (with 95% confidence intervals
      • Rosner B.
      by obtaining census data for all eligible subjects in each facility at monthly intervals throughout the course of the project; we accounted for absences from the long-term care facilities (eg, for hospitalization). Comparisons between categorical variables were performed using the chi-squared test. A P value of <0.05 was considered significant. Analyses were performed using SAS, version 8.0 (SAS Institute Inc, Cary, North Carolina).
      For the case-control studies, separate multivariate models were constructed using all adverse drug events and preventable adverse drug events as the outcome in stepwise conditional logistic regression.
      Age, sex, comorbidity, and number of regularly scheduled medications were forced into the models. Drug categories were retained in the models if they were found to have a P value ≤0.05.

      Results

      In the two study sites combined, 1247 long-term care residents (mean [± SD] age, 86 ± 8 years) yielded 8336.4 resident-months of observation time. A comparison of this sample with the overall U.S. population of nursing home residents demonstrated similar sex characteristics,
      National Nursing Home Survey (NNHS)
      but the study sample was somewhat older. In the United States in 1999, 72% of nursing home residents were women and 78% of all residents were aged 75 years or older, compared with 72% who were women and 92% who were 75 years of age or older in the present study.

      Rates of adverse drug events

      The clinical pharmacist-investigators identified 1042 possible drug-related incidents, of which 815 (78%) were judged to represent adverse drug events by the physician-reviewers. Of the 815 adverse drug events, 78% were identified from periodic medical record reviews, 20% were identified through computerized signals, and 2% were identified from incident reports.
      The overall rate of adverse drug events was 9.8 per 100 resident-months (95% confidence interval [CI]: 9.2 to 10.4 per 100 resident-months), with a rate of 4.1 preventable adverse drug events per 100 resident-months (95% CI: 3.6 to 4.5 per 100 resident-months). Of the 815 adverse drug events, 42% (n = 338) were judged preventable. Of the 225 serious, life-threatening, or fatal adverse drug events, 61% (n = 137) were deemed preventable, compared with 34% (n = 201) of the 590 less serious adverse drug events (Table 1). Overall, more severe adverse drug events were significantly more likely to be considered preventable (relative risk = 1.8; 95% CI: 1.5 to 2.1).
      Table 1Severity and effects of adverse drug events
      Total (n = 815)Preventable (n = 338)
      Number (%)
      Category of severity
       Fatal4 (<1)3 (1)
       Life threatening33 (4)24 (7)
       Serious188 (23)110 (32)
       Less serious590 (72)201 (60)
      Effects of adverse drug events
       Only laboratory abnormality68 (8)27 (8)
       ≤1 day of symptoms148 (18)46 (14)
       >1 day of symptoms590 (72)259 (76)
       Nonpermanent disability3 (<1)2 (1)
       Permanent disability2 (<1)1 (<1)
       Death4 (<1)3 (1)

      Effects of adverse drug events

      Most adverse drug events (>70%) resulted in symptoms lasting more than 1 day (Table 1). Six events resulted in permanent disability (n = 2) or death (n = 4). Events resulting in permanent disability included functional decline without subsequent improvement due to lithium toxicity and tardive dyskinesia. Fatal events included intracranial hemorrhage, aspiration pneumonia related to oversedation, digoxin toxicity, and a drug-associated metabolic abnormality (hypercalcemia).

      Types of adverse drug events

      Neuropsychiatric events (eg, oversedation, confusion, hallucinations, delirium) comprised the most common type of preventable and the second most common type of nonpreventable event (Table 2). The most frequently identified types of preventable adverse drug events also included gastrointestinal (eg, abdominal pain, diarrhea, constipation, impaction), hemorrhagic (bleeding events), renal/electrolytes (eg, azotemia, dehydration, hyperkalemia, hypokalemia, renal failure), and metabolic/endocrine (hypoglycemic events, thyroid abnormalities) events.
      Table 2Frequency of types of adverse drug events
      Adverse drug events could manifest as more than one type.
      Adverse Drug EventTotal (n = 815)Preventable (n = 338)
      Number (%)
      Neuropsychiatric
      Neuropsychiatric events include oversedation, confusion, hallucinations, and delirium.
      199 (24)97 (29)
      Hemorrhagic159 (20)53 (16)
      Gastrointestinal140 (17)55 (16)
      Renal/electrolytes80 (10)40 (12)
      Metabolic/endocrine64 (8)35 (10)
      Dermatologic36 (4)4 (1)
      Cardiovascular36 (4)15 (4)
      Extrapyramidal signs/symptoms30 (4)7 (2)
      Fall without injury21 (3)11 (3)
      Fall with injury21 (3)17 (5)
      Infection19 (2)1 (<1)
      Syncope/dizziness16 (2)8 (2)
      Anticholinergic
      Anticholinergic effects include dry mouth, dry eyes, urinary retention, and constipation.
      9 (1)3 (1)
      Ataxia/difficulty with gait9 (1)5 (2)
      Hematologic8 (1)3 (1)
      Respiratory6 (1)4 (1)
      Anorexia3 (<1)2 (<1)
      Functional decline
      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.
      3 (<1)2 (<1)
      Hepatic1 (<1)1 (<1)
      * Adverse drug events could manifest as more than one type.
      Neuropsychiatric events include oversedation, confusion, hallucinations, and delirium.
      Anticholinergic effects include dry mouth, dry eyes, urinary retention, and constipation.
      § 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.
      We assessed the specific medication types that were most frequently associated with adverse drug events (Table 3). Among preventable adverse drug events, warfarin, atypical antipsychotic agents, loop diuretics, intermediate-acting benzodiazepines, opioids, and angiotensin-converting enzyme inhibitors were most commonly involved.
      Table 3Frequency of adverse drug events by drug type
      Drugs in more than one category were associated with some events. Frequencies in each column sum to greater than the total number of events.
      Drug ClassTotal (n = 815)Preventable (n = 338)
      Number (%)
      Warfarin121 (15)42 (12)
      Atypical antipsychotic agents
      Olanzapine, risperidone, quetiapine, clozapine.
      92 (11)42 (12)
      Loop diuretics69 (8)33 (10)
      Opioids51 (6)26 (8)
      Antiplatelets46 (6)23 (7)
      ACE inhibitors45 (6)27 (8)
      Antidepressants (non-SSRI, nontricyclic)
      Trazodone, venlafaxine, mirtazapine, nefazodone, bupropion.
      43 (5)25 (7)
      Laxatives43 (5)16 (5)
      Benzodiazepines (intermediate acting)
      Lorazepam, temazepam, oxazepam.
      39 (5)30 (9)
      Insulins37 (5)18 (5)
      Quinolones27 (3)2 (1)
      Clindamycin23 (3)2 (1)
      Valproic acid22 (3)11 (3)
      SSRIs21 (3)10 (3)
      Potassium-sparing diuretics20 (2)9 (3)
      COX-2 inhibitors19 (2)11 (3)
      Beta-blockers18 (2)7 (2)
      Trimethoprim-sulfamethoxazole17 (2)6 (2)
      Typical antipsychotic agents
      Haloperidol, chlorpromazine, perphenazine, methotrimeprazine, fluphenazine.
      17 (2)10 (3)
      Phenytoin15 (2)10 (3)
      Thyroid15 (2)9 (3)
      Nitrates14 (2)3 (1)
      Tricyclic antidepressants13 (2)6 (2)
      Penicillins13 (2)1 (0)
      Cephalosporins13 (2)3 (1)
      Thiazide diuretics13 (2)8 (2)
      Benzodiazepines (long-acting)
      Diazepam, clonazepam.
      12 (1)5 (1)
      Heparin11 (1)7 (2)
      Dopamine agonists11 (1)4 (1)
      Bladder medications
      Oxybutynin, tolterodine.
      11 (1)5 (1)
      Gabapentin10 (1)5 (1)
      Digoxin10 (1)8 (2)
      Benzodiazepines (short-acting)
      Triazolam.
      10 (1)8 (2)
      Antihistamines8 (<1)5 (1)
      Glyburide8 (<1)5 (1)
      NSAIDs (traditional)7 (<1)5 (1)
      Donepezil6 (<1)0 
      Erythromycin6 (<1)0 
      Calcium channel blockers6 (<1)3 (1)
      Potassium chloride6 (<1)3 (1)
      Lithium5 (<1)3 (1)
      Prednisone5 (<1)1 (0)
      Proton pump inhibitors4 (<1)2 (1)
      Metformin4 (<1)3 (1)
      Vaccine4 (<1)0 
      Carbamazepine3 (<1)1 (0)
      Metronidazole3 (<1)0 
      Nitrofurantoin3 (<1)0 
      Cholestyramine3 (<1)3 (1)
      Calcium supplements3 (<1)3 (1)
      ACE = angiotensin-converting enzyme; COX-2 = cyclooxygenase-2; NSAID = nonsteroidal anti-inflammatory drug; SSRI = selective serotonin reuptake inhibitor.
      * Drugs in more than one category were associated with some events. Frequencies in each column sum to greater than the total number of events.
      Olanzapine, risperidone, quetiapine, clozapine.
      Trazodone, venlafaxine, mirtazapine, nefazodone, bupropion.
      § Lorazepam, temazepam, oxazepam.
      Haloperidol, chlorpromazine, perphenazine, methotrimeprazine, fluphenazine.
      Diazepam, clonazepam.
      # Oxybutynin, tolterodine.
      ** Triazolam.

      Risk factors for adverse drug events

      The characteristics of residents at the time of their first adverse drug event and of controls are summarized in Table 4. Age, sex, and the Charlson Comorbidity Index score were not associated significantly with adverse drug events in multivariate analyses. However, the number of regularly scheduled medications was associated with adverse drug events; compared with residents taking 1 to 5 medications, the odds ratio was 1.4 (95% CI: 0.9 to 2.0) for residents taking 6 to 8 medications, 1.7 (95% CI: 1.1 to 2.6) for those taking 9 to 11 medications, and 2.1 (95% CI: 1.3 to 3.5) for those taking ≥12 medications. Residents taking drugs from several drug categories were at higher risk; the odds ratio was 3.1 (95% CI: 1.7 to 5.6) among those taking anticoagulants, 2.4 (95% CI: 1.7 to 3.5) among those taking antipsychotic agents, 1.9 (95% CI: 1.3 to 2.8) among those taking anti-infectives, and 1.6 (95% CI: 1.1 to 2.5) among those taking antiepileptic agents.
      Table 4Characteristics of residents who had adverse drug events and of controls
      CharacteristicAll EventsPreventable Events
      Cases (n = 476)Controls (n = 476)P ValueCases (n = 247)Controls (n = 247)P Value
      Number (%)Number (%)
      Age (years)<0.010.14
       <7539 (8)31 (7)17 (7)17 (7)
       75–84147 (31)113 (24)71 (29)51 (21)
       85–94243 (51)255 (54)129 (52)137 (55)
       ≥9547 (10)77 (16)30 (12)42 (17)
      Female sex335 (70)368 (77)0.01169 (68)190 (77)0.03
      Charlson Comorbidity<0.010.02
      Index Score
       018 (4)22 (5)6 (2)11 (4)
       1–2240 (50)302 (63)121 (49)150 (61)
       3–4150 (32)113 (24)86 (35)61 (25)
       ≥568 (14)39 (8)34 (14)25 (10)
      Number of regularly scheduled medications<0.01<0.01
       0–569 (15)147 (31)24 (10)63 (26)
       5–8133 (28)158 (33)72 (29)86 (35)
       9–11131 (28)103 (22)71 (29)56 (23)
       ≥12143 (30)68 (14)80 (32)42 (17)
      Current medications
       Alzheimer disease186 (39)157 (33)0.05100 (40)85 (34)0.16
       Analgesic (nonopioid)206 (43)160 (34)<0.01116 (47)82 (33)<0.01
       Anticoagulant229 (43)161 (34)<0.01128 (52)84 (34)<0.01
       Antidepressant304 (64)262 (55)<0.01167 (68)141 (57)0.02
       Antihistamine202 (42)132 (28)<0.01116 (47)71 (29)<0.01
       Antihyperlipidemic200 (42)120 (25)<0.01114 (46)66 (27)<0.01
       Antibiotic/anti-infective241 (51)140 (29)<0.01134 (54)82 (33)<0.01
       Anti-gout51 (11)35 (7)0.0732 (13)20 (8)0.08
       Antineoplastic113 (24)51 (11)<0.0157 (23)27 (11)<0.01
       Antiparkinsonian100 (21)42 (9)<0.0153 (21)21 (9)<0.01
       Antiplatelet166 (35)123 (26)<0.0179 (32)58 (23)0.03
       Antipsychotic132 (28)69 (15)<0.0169 (28)30 (12)<0.01
       Antiepileptic99 (21)59 (12)<0.0152 (21)28 (11)<0.01
       Cardiovascular168 (35)140 (29)0.0587 (35)74 (30)0.21
       Digoxin89 (19)68 (14)0.0748 (19)35 (14)0.12
       Diuretic113 (24)88 (18)0.0564 (26)39 (16)<0.01
       Gastrointestinal164 (34)208 (44)<0.0179 (21)115 (47)<0.01
       Hematologic32 (7)24 (5)0.2721 (10)12 (5)0.10
       Hormone36 (8)30 (6)0.4421 (9)23 (9)0.75
       Hypoglycemic96 (20)60 (13)<0.0153 (21)29 (12)<0.01
       Immunomodulator8 (2)5 (1)0.404 (2)3 (1)0.70
       Muscle relaxant8 (2)7 (1)0.792 (1)6 (2)0.15
       Nutrient/supplement101 (21)147 (31)<0.0141 (17)72 (29)<0.01
       Ophthalmic47 (10)70 (15)0.0221 (9)47 (19)<0.01
       Opioid34 (7)33 (7)0.9016 (6)17 (7)0.86
       Osteoporosis21 (4)20 (4)0.8711 (4)11 (4)1.0
       Respiratory30 (6)27 (6)0.6812 (5)19 (8)0.19
       Sedative/hypnotic22 (5)18 (4)0.5215 (6)13 (5)0.70
       Steroid7 (1)7 (1)1.004 (2)7 (3)0.36
       Thyroid25 (5)33 (7)0.2812 (5)15 (6)0.55
       Topical44 (9)72 (15)<0.0117 (7)32 (13)0.02
      In multivariate analyses of potential risk factors associated with preventable adverse drug events, age, sex, the Charlson Comorbidity Index score, and the number of regularly scheduled medications were not associated significantly with the occurrence of an event. However, residents taking medications in several drug categories were at higher risk, including those taking antipsychotic agents (adjusted odds ratio [OR] = 3.4; 95% CI: 2.0 to 5.9), anticoagulants (adjusted OR =2.8; 95% CI: 1.6 to 4.7), diuretics (adjusted OR = 2.2; 95% CI: 1.2 to 4.0), and antiepileptics (adjusted OR = 2.0; 95% CI: 1.1 to 3.7).

      Errors associated with preventable adverse drug events

      Among the 338 preventable adverse drug events, errors occurred most commonly at the ordering (n = 198 [59%]) and monitoring (n = 271 [80%]) stages of pharmaceutical care. Errors accounting for preventable adverse drug events were less commonly identified at the dispensing (n = 16 [5%]) and administration (n = 43 [13%]) stages. A total of 154 (46%) of the preventable adverse drug events were associated with an error at 2 stages of pharmaceutical care, whereas 18 (5%) were associated with an error at 3 stages of pharmaceutical care.
      Among the 198 prescribing errors, the most common were wrong dose (n = 96 [48%]), wrong drug choice (n = 76 [38%]), and known drug interaction (n = 23 [12%]). 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 271 monitoring errors, the most common were inadequate monitoring (n = 197 [73%]) and failure to act on monitoring (n = 180 [66%]).

      Discussion

      We found that adverse drug events occurred frequently and were often preventable in the two academic long-term care facilities that participated in this study. Serious, life-threatening, and fatal adverse drug events were more likely to be preventable than were less severe events. Most errors associated with preventable events occurred at the prescribing and monitoring stages of pharmaceutical care.
      Our study offers the opportunity to compare observed adverse drug event rates with those described in a previous study involving smaller community-based nursing homes.
      • Gurwitz J.H.
      • Field T.S.
      • Avorn J.
      • et al.
      Incidence and preventability of adverse drug events in nursing homes.
      In that study, we observed a rate of 1.9 events per 100 resident-months, and 51% of adverse drug events were judged preventable. In the present study, we observed an event rate that was more than 5 times greater and a rate of preventable events that was more than 4 times greater. In both studies, more severe events were more likely to be judged preventable. However, in the present study, a smaller percentage of the events fell into the more severe categories; 28% of events were deemed fatal, life threatening, or serious, compared with 44% in the former study.
      We believe that the main factor accounting for the higher rate of adverse drug events detected in this study was an enhanced approach to the identification of drug-related incidents. The clinical pharmacist-investigators had a continuous presence within the participating facilities throughout the study, providing increased access to medical record information relevant to possible drug-related incidents. In addition to periodic review of all medical records for all residents, computer-generated signals were used to assist in targeting medical records for more intensive review. Such a level of access to relevant information was not possible in our previous study. This enhanced approach may have allowed the identification of a substantially larger number of events than might have been expected, many of which were of lower severity.
      If our findings are applied to all U.S. nursing homes, then approximately 120 adverse drug events should be identifiable each year in an average facility (bed size of 105), many of which should be preventable. About 1.9 million adverse drug events—more than 40% of which are preventable—may occur each year among the 1.6 million residents of U.S. nursing homes. There may be more than 86,000 fatal or life-threatening adverse drug events per year, of which 70% may be preventable. Although these estimates are higher than previously reported, they are still likely to be conservative. Despite enhanced efforts to identify drug-related incidents, we relied solely on information from medical records. There was no direct assessment of long-term care facility residents; such assessments may have allowed for the identification of additional events.
      How should the findings of this study be applied to improve the quality of care for residents of long-term care facilities? Enhanced surveillance and reporting systems for adverse drug events in the long-term care setting are required, and intensified educational efforts concerning the optimal use of drug therapies in frail elderly patients are essential. Current efforts relating to identifying and analyzing adverse events in the long-term care setting are inadequate and rarely lead to sustainable changes that result in improvements in medication safety. Computerized order entry with clinical decision support may hold the greatest promise for reducing medication error rates in the long-term care setting. The benefits of this approach for reducing medication errors in other clinical settings have been established.
      • Kuperman G.J.
      • Gibson R.F.
      Computer physician order entry benefits, costs, and issues.
      • Kaushal R.
      • Shojania K.G.
      • Bates D.W.
      Effects of computerized physician order entry and clinical decisions support systems on medication safety.
      However, few long-term care facilities have implemented such systems owing to cost, complexity, and logistical challenges, as well as uncertainty about how effective these systems actually are for reducing drug-related injuries once implemented.
      Improving patient safety in the long-term care setting will require substantial financial resources. According to Berwick, “Until we find ways to make errors and injuries routinely visible in local health care settings, a national will to improve safety will be hard to translate into local intent.”
      • Berwick D.M.
      Errors today and errors tomorrow.
      Our research findings serve to shine a light on an often overlooked patient population and emphasize the need to develop and test innovative strategies for preventing adverse drug events in the long-term care setting.

      Acknowledgment

      We thank Mary Ellen Stansky and Jackie Cernieux, MPH, for their assistance with technical aspects of this study, and Bessie Petropoulos for assistance with manuscript preparation.

      Appendix

      Computer-generated signals: abnormal serum drug levels, abnormal laboratory results, and drug triggers suggestive of adverse drug events.
      Drug levels
       Serum quinidine >5 μg/mL
       Serum valproate >120 μg/mL
       Serum theophylline >20 μg/mL
       Serum procainamide >12 μg/mL
       Serum phenobarbital >45 μg/mL
       Serum phenytoin results >20 μg/mL
       Serum digoxin >2.0 ng/mL
       Serum carbamazepine >13.0 μg/mL
      Laboratory results (including drug-laboratory combinations)
       Serum alkaline phosphatase >350 U/L
       Serum bilirubin >4.0 mg/dL
       Serum potassium <3.0 mEq/L
       Serum potassium >5.6 mEq/L
       Serum aspartate aminotransferase >84 U/L
       Serum alanine aminotransferase >80 U/L
       Serum urea nitrogen >50 mg/dL
       International normalized ratio >5
       Platelet count <50,000/μL
       Serum creatinine >2.5 mg/dL
       L-thyroxine and thyroid-stimulating hormone <0.3 μIU/mL
      Clostridium difficile testing
      Drug triggers
       Diphenhydramine
       Prednisone and diphenhydramine
       Phytonadione (vitamin K)
       Naloxone
       Sodium polystyrene sulfonate
       Protamine sulfate
       Glucagon
       Hydroxyzine
       Hydroxyzine and prednisone
       Oral vancomycin
       Metronidazole
       Nystatin
       Glucocorticoid and hypoglycemic agent

      References

        • Gurwitz J.H.
        • Field T.S.
        • Avorn J.
        • et al.
        Incidence and preventability of adverse drug events in nursing homes.
        Am J Med. 2000; 109: 87-94
        • Gurwitz J.H.
        • Field T.S.
        • Harrold L.R.
        • et al.
        Incidence and preventability of adverse drug events among older persons in the ambulatory setting.
        JAMA. 2003; 289: 1107-1116
        • Bates D.W.
        • Cullen D.J.
        • Laird N.
        • et al.
        Incidence of adverse drug events and potential adverse drug events implications for prevention. ADE Prevention Study Group.
        JAMA. 1995; 274: 29-34
        • Leape L.L.
        • Bates D.W.
        • Cullen D.J.
        • et al.
        Systems analysis of adverse drug events.
        JAMA. 1995; 274: 35-43
        • Leape L.L.
        • Cullen D.J.
        • Clapp M.D.
        • et al.
        Pharmacist participation on physician rounds and adverse drug events in the intensive care unit.
        JAMA. 1999; 282: 267-270
        • Kaushal R.
        • Bates D.W.
        • Landrigan C.
        • et al.
        Medications errors and adverse drug events in pediatric in-patients.
        JAMA. 2001; 285: 2114-2120
        • Bates D.W.
        • Leape L.L.
        • Cullen D.J.
        • et al.
        Effect of computerized physician order entry and a team intervention on prevention on serious medication errors.
        JAMA. 1998; 280: 1311-1316
        • Bates D.W.
        • Spell N.
        • Cullen J.J.
        • et al.
        The costs of adverse drug events in hospitalized patients.
        JAMA. 1997; 277: 307-311
        • Freedman V.A.
        • Martin L.G.
        • Schoeni R.F.
        Recent trends in disability and functioning among older adults in the United States.
        JAMA. 2002; 288: 3137-3146
        • Sackett D.L.
        • Haynes R.B.
        • Guyatt G.H.
        • Tugwell P.
        Clinical Epidemiology. 2nd ed. Little Brown & Co, Boston, Massachusetts1991
        • Charlson M.E.
        • Pompei P.
        • Ales K.L.
        • MacKenzie C.R.
        A new method of classifying prognostic comorbidity in longitudinal studies.
        J Chronic Dis. 1987; 40: 373-383
        • Rosner B.
        Fundamentals of Biostatistics. 3rd ed. PWS-Kent, Boston, Massachusetts1990
      1. Stata Statistical Software. Release 6.0. Stata Corp, College Station, Texas:1999
        • National Nursing Home Survey (NNHS)
        2003 (Available at:http://www.cdc.gov/nchs/about/major/nnhsd/nnhsd.htm.Accessed July 18)
        • Kuperman G.J.
        • Gibson R.F.
        Computer physician order entry.
        Ann Intern Med. 2003; 139: 31-39
        • Kaushal R.
        • Shojania K.G.
        • Bates D.W.
        Effects of computerized physician order entry and clinical decisions support systems on medication safety.
        Arch Intern Med. 2003; 163: 1409-1416
        • Berwick D.M.
        Errors today and errors tomorrow.
        N Engl J Med. 2003; 348: 2570-2572