Drug-related hospital admissions

Abstract

Objective: To review and summarize studies reporting rates of drug-related hospital admissions.

Data sources: Manual and computerized literature searches using MEDLINE, Index Medicus, and International Pharmaceutical Abstracts as databases (key words: drug, drug-related, or iatrogenic; admission, hospital admission, or hospitalization; and ADR or adverse drug reaction). References from retrieved articles were searched to locate further studies.

Study selection: Included were English-language studies of humans admitted to the hospital because of medications. Problems investigated were admissions prompted by adverse drug reactions (ADRs) when drugs were used by the patient and admissions resulting from a patient's noncompliant or unintentionally inappropriate drug use. Excluded were cases involving drug abuse, alcoholism, suicide attempts, intoxication, or inadequate prescribing.

Data synthesis: Between 1966 and 1989, ADR rates from 49 hospitals or groups of hospitals in a variety of international settings were published in 36 articles. Samples sizes ranged from 41 to 11,891 patients, with a median of 714 (interquartile range [IQR] 275-1245) and a mean of 1412 (SD 2233). The prevalence of reported admissions resulting from ADRs ranged from 0.2 to 21.7 percent; the median was 4.9 percent (IQR 2.9-6.7 percent) and the mean was 5.5 percent (SD 4.1 percent). The weighted meta-analytic estimate was 5.1 percent (95 percent confidence interval 4.4-5.8). Of those ADR admissions, 71.5 percent were side effects, 16.8 percent excessive effects, 11.3 percent hypersensitivity reactions, and 0.4 percent idiosyncratic; 3.7 percent of patients admitted for ADRs died. Eleven reports indicated that 22.7 percent of ADR hospitalizations were induced by noncompliance.

Conclusions: Drug-induced hospitalizations account for approximately five percent of all admissions. Results apply only to people from highly developed industrialized countries. Economic analyses have not been performed. Future research should include the Third World and nonindustrialized nations as well as specific cultural groups.