1044 — Comparison of Strategies to Improve Antipsychotic Monitoring and Management for Schizophrenia

Owen RR (Center for Mental Healthcare and Outcomes Research (CMHOR), Central Arkansas Veterans Health Care System), Smith JL (CMHOR, Central Arkansas Veterans HCS), Hudson TJ (CMHOR, Central Arkansas Veterans HCS), Curran GM (CMHOR, Central Arkansas Veterans HCS)

Objectives:
To develop and test implementation strategies to increase adherence to schizophrenia guideline recommendations for use of moderate antipsychotic doses and routine monitoring for metabolic side effects.

Methods:
Using a multiple case study design, the project tested two influence strategies - a team-based quality improvement (QI) approach versus a clinical opinion leader (OL) strategy. Other intervention components included performance monitoring with monthly feedback, information system tools, and educational materials. Three VAMCs from each of two VISNs were selected to participate on the basis of current performance on study indicators. Within each network, sites were randomly assigned to team-based QI, OL, or “usual care” study conditions. On-site needs assessments were conducted at all sites. At intervention sites, investigators facilitated selection of and trained QI team members or opinion leaders. The research team provided ongoing external facilitation during the study’s 6-month implementation phase. Performance indicators included the proportion of patients with schizophrenia prescribed excessive antipsychotic doses; and the proportion of these patients who had body weight, blood glucose, and/or lipid profile assessed when a new antipsychotic medication was started.

Results:
Five facilities completed the study as planned; one Team QI site did not continue after on-site needs assessment. The Team QI site increased the proportion of patients with recommended weight monitoring from 70% (baseline) to 93% (follow-up), with dramatically increased glucose and lipid monitoring (53% to 80% and 29% to 67%, respectively). At one OL site, monitoring rates increased substantially (weight: 80% to 88%; glucose: 48% to 63%; and lipids: 30% to 57%), while rates at the other OL site decreased. Usual care sites did not substantially increase monitoring rates during the study. Intervention strategies did not affect the proportion of patients receiving excessive antipsychotic doses. Formative evaluation findings suggest that designating a provider or clinic to ensure that monitoring occurred, and a computer routine to identify patients due for monitoring, were effective QI components.

Implications:
Two intervention sites achieved dramatic improvements in recommended metabolic side effect monitoring. Although neither influence strategy was clearly superior to the other, promising intervention elements were identified.

Impacts:
Findings from this study are informing further efforts to improve antipsychotic side effect monitoring and management.