Hospitalizations due to preventable adverse reactions-a systematic review.

Purpose: The study aimed to measure the percentage of preventable adverse drug reactions that lead to the hospitalization (PADR) and to explore the heterogeneity in its estimation through subgroup analysis of study characteristics.

Methods: Two investigators independently searched in electronic databases and related bibliography for prospective studies involving PADR. We excluded studies investigating medication errors and spontaneous and retrospective reporting. The primary outcome was PADR percentage. To explore the heterogeneity, we performed subgroup analysis based on study region, wards, age groups, adverse drug reaction (ADR) definitions, preventability assessment, ADR identification methods, study duration and sample size. We explored fatal PADR and causative drugs as a secondary outcome. We used the generic inverse variance method with random effect model to compute meta-analytic summary.

Results: Of the 68 full-text articles assessed, we included 22 studies. The mean PADR percentage was 45.11 % (95 % CI = 33.06-57.15; I  = 99 %). Studies including elderly (63.31 %) and all age groups (49.03 %) showed higher percentages than paediatric population (16.40 %). Studies examining all hospital populations showed higher percentages than specific wards. We observed high percentages in studies using Edwards and Aronson as an ADR definition and Hallas et al. as a preventability assessment tool. After age group adjustment, ADR detection methods did not show significant difference. The fatal PADR percentage was 1.58 % (95 % CI = -0.60 to 3.76; I  = 47 %). Paediatric and elderly studies showed a different causative drug pattern.

Conclusion: Variation in PADR across the studies can be explained by difference in study populations and data collection methods. Extrapolation of preventable reactions should be carried out considering all these factors with caution.