Systematic review of the incidence and characteristics of preventable adverse drug events in ambulatory care.

OBJECTIVE: To estimate the incidence and describe characteristics of preventable adverse drug events (pADEs) in ambulatory care. DATA SOURCES: Studies were searched in PubMed (1966-March 2007), International Pharmaceutical Abstracts (1970-December 2006), the Cochrane database of systematic reviews (1993-March 2007), EMBASE (1980-February 2007), and Web of Science (1945-March 2007). Key words included medication error, adverse drug reaction, iatrogenic disease, outpatient, ambulatory care, primary health care, general practice, patient admission, hospitalization, observational study, retrospective studies, health services research, and follow-up studies. Additional articles were found in the reference sections of retrieved articles. STUDY SELECTION AND DATA EXTRACTION: Peer-reviewed articles assessing pADEs in ambulatory care, with detailed descriptions/frequency distributions of (1) ADE/pADE incidence, (2) clinical outcomes, (3) associated drug groups, and/or (4) underlying medication errors were included. Study country, year and design, sample size, follow-up time, ADE/pADE identification method, proportion of ADEs/pADEs and ADEs/pADEs requiring hospital admission, and frequency distribution of adverse outcome, associated drug groups, or medication errors were extracted. DATA SYNTHESIS: Twenty-nine studies met inclusion criteria: 14 were ambulatory-based and 15 were hospital-based. Seven studies enrolled only elderly patients. The median ADE incidence was 14.9 (range 4.0-91.3) per 1000 person-months, and the pADE incidence was 5.6 per 1000 person-months (1.1-10.1). The median ADE preventability rate was 21% (11-38%). The median incidence of ADEs requiring hospital admission was 0.45 (0.10-13.1) per 1000 person-months, and the median incidence of pADEs requiring hospital admission was 4.5 per 1000 person-months. Cardiovascular drugs, analgesics, and hypoglycemic agents together accounted for 86.5% of pADEs, and 77.2% of pADEs resulted in symptoms of the central nervous system, electrolyte/renal system, and gastrointestinal tract. Medication errors resulting in pADEs occurred in the prescribing and monitoring stages. The most frequent drug therapy problem and error of commission reported in ambulatory-based studies on pADEs was the use of inappropriate drugs (42.7%; 40.4-45%). For pADEs requiring hospital admission, the most frequent drug therapy problem and error of omission reported was inadequate monitoring (45.4%; range 22.2-69.8%). Failure to prescribe prophylaxis to patients taking nonsteroidal antiinflammatory drugs or antiplatelet drugs frequently caused gastrointestinal toxicity, whereas lack of monitoring of diuretic, hypoglycemic, and anticoagulant use caused over- or under-diuresis, hyper- or hypoglycemia, and bleeding.
CONCLUSIONS: ADEs in ambulatory care are common, with many being preventable and many resulting in hospitalization. Quality improvement programs should target errors in prescribing and monitoring, especially for patients using cardiovascular, analgesic, and hypoglycemic agents.