Deprescribing Psychotropic Medications and Falls in Older Adults: A Setting-Stratified Systematic Review and Meta-Analysis.

OBJECTIVES

To evaluate the effect of deprescribing interventions that explicitly targeted psychotropic medications on fall outcomes in older adults across community, inpatient, long-term care, and postdischarge settings.

DESIGN

Systematic review and meta-analysis.

SETTING AND

PARTICIPANTS

Older adults (≥65 years).

METHODS

We searched MEDLINE, Embase, Cochrane CENTRAL, Web of Science, and CINAHL from inception to June 2025. We included randomized controlled trials (RCTs) and high-quality nonrandomized studies of interventions (NRSIs) that compared structured deprescribing or medication-optimization interventions with an explicit psychotropic component against usual care in adults aged ≥65 years. Data were pooled using random-effects models (Restricted Maximum Likelihood estimator), and heterogeneity was quantified using the I2 statistic. Subgroup analyses were performed based on intervention fidelity and clinical setting.

RESULTS

A total of 18 studies (11 RCTs and 7 NRSIs) were included. In the primary analysis of long-term community RCTs (n = 1425), psychotropic-targeting deprescribing interventions did not significantly reduce the odds of falls [odds ratio (OR), 0.91; 95% CI, 0.59-1.39]. However, subgroup analysis indicated a significant reduction in falls for high-fidelity interventions (OR, 0.61; 95% CI, 0.41-0.91), whereas low-fidelity interventions showed no effect. In the inpatient setting, a meta-analysis of 5 NRSIs (n = 5972) demonstrated a significant reduction in falls (OR, 0.43; 95% CI, 0.19-0.96; P = .03). No significant reduction in falls was observed in the long-term care or postdischarge transitional care settings.

CONCLUSIONS AND IMPLICATIONS: Based on low to very low certainty evidence, psychotropic-targeting deprescribing interventions may reduce falls in inpatient settings and in high-fidelity community programs, whereas effects were not clearly demonstrated in other settings. These findings support its potential role as a targeted strategy and underscore the need for high-fidelity protocols and adequately powered trials tailored to specific care settings.