The relationship between study characteristics and the prevalence of medication-related hospitalizations: a literature review and novel analysis

Abstract

Background: Studies on medication-related hospitalizations differ in study setting, studied population, outcome, and method of data collection. Thus, extrapolations based on a meta-analysis of unselected studies may be biased.

Objective: To explore the influence of study characteristics on the prevalence of medication-related hospitalizations.

Methods: After a structured literature search, the retrieved studies were categorized based on the following aspects: (i) study setting (e.g. all hospital admissions vs only acute hospital admissions); (ii) study population (e.g. an entire hospital, study ward(s), selected population and/or age group); (iii) outcome of medication-related problem (e.g. adverse drug reaction [ADR] vs adverse drug event [ADE]); (iv) method of data collection (e.g. medical chart review, spontaneous reporting or database research); and (v) continent in which the study took place (only for studies looking at all acute admissions). We then examined the relationship between these factors and reported prevalence of medication-related hospital admissions.

Results: Ninety-five studies were analysed, with a range of reported prevalence of medication-related hospitalizations from 0.1% to 54%. Higher prevalences were found in the studies examining all hospital admissions than in the studies examining only acute hospital admissions. In addition, higher prevalences were found in the elderly population than in children. As would be expected, higher prevalences were also found in studies examining ADEs than in studies examining only ADRs. With respect to the method of data collection, medical chart screening resulted in higher prevalences of medication-related hospitalizations than database methods or spontaneous reporting. Combined studies in Europe show lower prevalences of medication-related hospital admissions than in other continents included in the study.

Discussion: The reported prevalence of medication-related hospital admissions varies as a function of the setting (all admissions or only acute admissions), studied population (entire hospital, specific wards, selected population and age group), outcome (ADR/ADE), the method of data collection and the continent in which the study is performed.

Conclusion: Extrapolation using national hospital admission data and the prevalence identified by pooling international studies should be carried out with great caution.