OBJECTIVES: To describe the variation in medication reconciliation proportions among pharmacies and evaluate the association between pharmacy medication reconciliation proportions and patient-level polypharmacy. METHODS: Using a nationwide Japanese claims database, we identified patients aged ≥75 years who visited a pharmacy in 2019. We compared patients' polypharmacy statuses between pharmacies with high medication reconciliation proportions (upper 10% pharmacy-level medication reconciliation proportions adjusted for patient characteristics) and those with low proportions (lower 90%). The primary outcome was patient-level polypharmacy, defined as regularly receiving prescriptions for ≥6 oral medications. Secondary outcomes included the number of prescribed medications, hyper-polypharmacy (≥10 oral medications), and the use of potentially inappropriate medications (PIMs). We constructed multivariate linear regression models adjusted for patient characteristics, pharmacy characteristics, and medical institution fixed effects. KEY FINDINGS: In 2019, 1 250 707 patients aged ≥75 years received prescriptions for ≥1 medication at 5817 pharmacies. Among them, 383 461 (30.7%) and 88 080 (7.0%) patients regularly received prescriptions for ≥6 and ≥10 medications, respectively, and 419 163 (33.5%) patients were prescribed PIMs. The mean number of prescribed medications was 4.0 (standard deviation, 3.4). After confounder adjustment, there was no significant difference in the polypharmacy prevalence between pharmacies with low and high medication reconciliation proportions (adjusted difference, -0.48 percentage points; 95% confidence interval, -1.09 to 0.14 percentage points; P = .13). No significant differences were observed in the secondary outcomes. CONCLUSIONS: Polypharmacy prevalence did not differ according to pharmacy-level medication reconciliation proportions, suggesting that these proportions may not be appropriate as a quality indicator for reducing polypharmacy among older adults.
Kajino et al. (Mon,) studied this question.