A randomized clinical trial conducted in a large Massachusetts health system showed that two EHR-based interventions increased deprescribing of medications considered potentially inappropriate for older adults. Long-term use of these drugs is linked to higher risks for falls and hospitalization. Despite guideline recommendations to limit them, reducing doses or discontinuing these medications remains uncommon in routine care due to factors such as clinician time constraints and a lack of effective tools to support medication changes. Results of the study were published in JAMA.
The study included 201 primary care physicians and 1,146 patients aged 65 years and older who had been prescribed benzodiazepines, nonbenzodiazepine sedative hypnotics, or multiple anticholinergic medications. The mean age of the patient population was 73.6 years, and most patients were women. Follow-up averaged nearly 290 days.
No serious adverse events were reported through the health system’s reporting pathways. Death rates were similar across groups, with no indication of harm associated with the deprescribing prompts.
Behavioral Prompts Drive Measurable Shifts in Prescribing
Physicians were assigned to a usual care group (no intervention) or to one of the two behavioral science–informed EHR strategy groups (precommitment or boostering). In the precommitment group, clinicians received an initial-visit notification prompting them to begin a conversation about medication risks, followed by a second notification at the next visit encouraging deprescribing. In the boostering group, physicians received an in-visit notification and a follow-up (“booster”) reminder delivered through the EHR inbox 4 weeks later.
Across the study period, 32.5% of patients had at least one medication deprescribed. Rates were highest in the precommitment group at 36.8%, followed by 34.3% in the boostering group and 26.8% in the usual care group. Compared with usual care, deprescribing was 40% more likely in the precommitment group and 26% more likely in the boostering group.
Low-Cost Interventions and Implications for Health System Efficiency
The findings highlight the potential operational value of EHR-embedded behavioral strategies in addressing medication safety concerns among older adults. Prior deprescribing efforts have often depended on resource-intensive approaches such as pharmacist-led reviews or specialist referrals. In contrast, the interventions tested in this trial required no additional staffing and were integrated directly into existing workflows.
The results suggest that targeted EHR notifications may offer a scalable method for reducing use of medications linked to falls, hospitalizations, and other adverse outcomes among older adults. Although the interventions did not significantly reduce cumulative medication doses across drug classes, the increase in deprescribing events indicates that behavioral design principles can influence clinician decision-making in routine practice. For health systems seeking to improve medication safety while minimizing operational burden, these findings provide evidence supporting the use of structured, behavior-informed EHR tools.