1159 — Effectiveness of an Implementation Strategy to Improve Metabolic Side Effect Monitoring for Patients Prescribed Antipsychotics

Owen RR, Center for Mental Healthcare & Outcomes Research; Li C, Center for Mental Healthcare & Outcomes Research; UAMS College of Pharmacy; Viverito KM, Center for Mental Healthcare & Outcomes Research; Smith JL, Mental Health QUERI; Austen MA, Center for Mental Healthcare & Outcomes Research; Marchant KE, Center for Mental Healthcare & Outcomes Research; Brace CD, Center for Mental Healthcare & Outcomes Research; Drummond KL, Center for Mental Healthcare & Outcomes Research;

Objectives:
Although practice guidelines recommend monitoring for metabolic side effects of antipsychotics, including weight gain, diabetes, and dyslipidemia, recent research has found gaps in monitoring rates. This study tested the effectiveness of an evidence-based quality improvement (EBQI) plus facilitation strategy for improving metabolic monitoring and management for patients prescribed antipsychotics.

Methods:
This cluster-randomized clinical trial was conducted at 12 systematically selected VA facilities, which were randomized to intervention or control conditions. The intervention involved researchers facilitating development of tailored local implementation strategies by clinical stakeholders to improve metabolic monitoring and management, with ongoing external facilitation for 6 months. A VistA software routine was offered to sites to identify patients due for monitoring in real time. Summative evaluation examined rates of monitoring (weight, glucose, LDL) around the time of a new antipsychotic prescription (baseline) and between 31 and 120 days after the new prescription (follow-up) during three 6-month study phases: pre-implementation, implementation, and sustainability.

Results:
At all sites, between 4,205 and 4,779 patients were identified as having a new antipsychotic prescription requiring baseline or follow-up metabolic monitoring in the pre-implementation or implementation periods. Monitoring rates were low overall (e.g., the baseline monitoring rate for glucose across sites was 26.3% during the pre-implementation period, and 16.4% for LDL), with generally higher monitoring rates at follow-up as compared to baseline timepoints, and mostly lower monitoring rates during the implementation period as compared to pre-implementation. At intervention sites, rates of baseline monitoring for the three metabolic parameters decreased less than monitoring for the comparable parameters at control sites; preliminary regression analyses found no significant intervention condition X time interactions. Further analyses will control for differences in patient characteristics and include time series analysis to examine changes in monthly monitoring rates during pre-implementation, implementation, and sustainability phases.

Implications:
Metabolic monitoring continues to be under-performed at VA facilities, and preliminary analyses suggest that an EBQI/facilitation strategy does not substantially improve monitoring rates. Possible barriers to implementation included competing demands for managers and clinicians, QI participant turnover, and information technology delays.

Impacts:
More research is needed to overcome implementation barriers to improve monitoring for antipsychotics' metabolic effects, patient safety and long-term outcomes.