Int J Clin Pharm. 2026 Jun 4. doi: 10.1007/s11096-026-02167-3. Online ahead of print.
ABSTRACT
INTRODUCTION: Medication-related hospital admission is a major contributor to preventable patient harm and healthcare burden worldwide. The relationship between age and medication-related hospital admission has not been comprehensively synthesised.
AIM: To evaluate the association between age and medication-related admission in adults and to provide research recommendations.
METHOD: MEDLINE (Ovid), EMBASE, Scopus, Web of Science, PsycINFO, CINAHL, Cochrane Library, and Global Health databases were searched from January 2000 to March 2025. Studies of adults reporting medication-related admission, providing data on participant age for an exposed and comparator group, published in English were included. Studies restricted to children were excluded. Selection through dual, independent title/abstract screening and full-text assessment using predefined eligibility criteria. Disagreements resolved by third reviewer or consensus. Study quality assessed using Joanna Briggs Institute tool for systematic reviews addressing questions of prevalence. Findings were reported descriptively and two meta-analyses were performed. Risk ratios of experiencing medication-related admission by age group (16-64 years-old, 65+years-old) were pooled. Mean ages of those experiencing medication-related admission compared to another cause of admission were pooled. Prediction intervals and sub-group analyses were used to explain heterogeneity.
RESULTS: Fifty studies, recruiting 210,514 participants, were included, with the majority judged as good quality. Thirteen studies exclusively recruited an older cohort, 30 studies were undertaken in Europe. Age data were variably described. The pooled risk ratio 1.64, (95% CI 1.15-2.35, K = 12, I2 = 96%, n = 56,101) suggested adults ≥ 65 years had a higher risk of experiencing medication-related admission than younger adults (16-64 years). Studies undertaken in Africa had a higher effect size. The pooled mean difference (2.86 years, 95% CI – 1.63-7.36; p = 0.21, K = 10, I2 = 93%, n = 12,440) indicated no statistically significant difference in age between groups. Risk of bias judgement was a statistically significant contributor to heterogeneity. Prediction intervals for both meta-analyses were wide, suggesting that findings of future studies could differ significantly. The certainty of evidence, GRADE approach, was very low.
CONCLUSION: Age is an unreliable predictor for medication-related admission without accounting for confounding and context. Future research should consistently define age categories and better explore other risk factors for medication-related admission.
PMID:42240819 | DOI:10.1007/s11096-026-02167-3
