Transforming the Medication Regimen Review Process of High-Risk Drugs Using a Patient-Centered Telemedicine-Based Approach to Prevent Adverse Drug Events in the Nursing Home

Introduction. The National Action Plan for Adverse Drug Event Prevention identified the nearly 16,000 NHs as a clinical setting where adverse drug event (ADE) prevention strategies for high-risk drug classes are lacking. A recent Office of Inspector General (OIG) report estimates that 37% of all harmful adverse events are related to drugs and two-thirds are preventable. The National Action Plan calls for focus on 3 high-risk drug classes, anticoagulants, diabetic agents, and opioids, which account for most preventable ADEs. Our prior work showed that drugs associated acute kidney injury (AKI) occur at a rate of 4.1 events per 1000 resident-days in the NH, underscoring the need to also include AKI in ADE prevention strategies. Current strategies are failing to improve medication safety in NH residents because: 1) medication regimen reviews (MRR) are usually conducted retrospectively, thus missing residents whose stay is < 30 days; 2) pharmacists are usually not involved in MRR on admission to the NH; 3) there has not been consensus regarding high-risk drug classes that require vigilant monitoring for ADE prevention; and 4) the MRR process is not patient-centered as most NH pharmacists never interact directly with the residents.

The aims of this project were to: (1) Evaluate the effect of pharmacist-led MRRs using patient-centered telemedicine for residents receiving high-risk drugs commonly associated with ADEs, and (2) Evaluate patient reported outcomes and the perception of healthcare professionals for pharmacist performance of enhanced services.

Methods. A clinical decision support system was developed for use in the NH to monitor prescription of high-risk drugs. Patient-centered MRRs were conducted using telemedicine with the NH resident when a high-risk drug was prescribed. Consultant pharmacists provided structured feedback and recommendations to the resident’s attending physician following the interaction with the NH resident. A cluster randomized controlled trial, using step-wedged randomization on the NH level was conducted on 4 NHs within the health system. At intervention completion, a blinded review of medical records for ADEs was conducted. Surveys were administered to patients, nurses and physicians pre and post intervention regarding their interactions with consulting pharmacists.

Results. During the 3 month run-in period, there were no significant differences between alert rates of intervention and control sites (57.18 vs 75.95 per 10000 resident days; adjusted incident rate ratio=AIRR=1.03; p=0.8257). During the intervention period, there were 365 alerts, 9 ADEs, and 9 possible ADEs over 63,831 resident days in intervention; and 386 alerts, 31 ADEs, 31 possible ADEs over 50,823 resident days in control. There were 76% less ADEs in intervention than controls (1.41 vs 6.10 per 10000 resident days; AIRR=0.24; p=0.0015). NH residents, nurses and physicians in the intervention sites rated pharmacists more positively pre-to-post intervention.

Conclusions. The product of this research is a generalizable EMR-agnostic MRR model including decision support rules and structured communication tools to optimally execute the consultant pharmacist’s role in ADE prevention in the NH.