Medication reviews in hospitalised patients for reduced hospital readmission and mortality. Systematic review, meta-analysis and meta-regression of RCTs.

Efforts to reduce preventable medication-related harm through medication reviews have increased, but interventions often yield null-results regarding clinical outcomes. We conducted a systematic literature search in four data bases and summarised the available evidence from randomised controlled trials (RCTs) comparing medication reviews and usual care in hospitalised patients regarding hospital readmissions and all-cause mortality by random-effects meta-analyses. Effect size differences by methodological study differences were of special interest. The meta-analysis of all 24 trials on hospital readmissions, including 12,539 participants, showed a statistically significant 8 % decrease in hospital readmissions (risk ratio (RR) [95 % confidence interval]: (0.92 [0.88-0.97], p = 0.002). The number of patient contacts was the most prominent effect modifier in meta-regression (p = 0.003) and the effect of medication reviews was approximately twice as strong (15 %) in 11 trials with 2 or more patient contacts (0.85 [0.78-0.92], p < 0.001). No statistically significant reduction in all-cause mortality was observed in a meta-analysis of all 22 trials with data for this outcome (0.95 [0.86-1.04], p = 0.24), including 12,350 participants. The method of mortality assessment was identified as an effect modifier by meta-regression (p = 0.01). A meta-analysis of 10 trials with complete mortality ascertainment via registries or primary care data showed a significantly 19 % reduced mortality (0.81 [0.70-0.94], p < 0.01). In conclusion, medication reviews reduce the risk of hospital readmission and might also reduce all-cause mortality. Comprehensive mortality assessment was essential for successful trials. Clinical guidelines should recommend medication reviews with multiple patient contacts, involving pharmacists, either for repeated medication reviews or to improve adherence.