The effect of objectively collected medication adherence information on bipolar I and major depressive disorder treatment decisions: A randomized case vignette study of psychiatric clinicians

Abstract

• In the absence of credible medication adherence data, prescribers were more likely to confound medication nonadherence and nonresponse. • In the presence of credible adherence data, prescribers were more likely to prescribe LAI formulations. This effect was much stronger for those treating BD than those treating MDD. • The difference in prescribing decisions between to those exposed to credible information and those relying solely upon patient and collateral self-report illustrate the challenges associated with assessing adherence when psychiatric symptoms are not well controlled. Adherence to antipsychotic medications is critical for treatment and relapse prevention in serious mental illness. Accurate assessment of medication adherence can be difficult, clinicians frequently overestimate patient adherence, and new accurate methods assessing medication adherence are needed. This study estimated the influence of antipsychotic medication adherence information on clinician treatment decisions among patients with bipolar I (BD) and major depressive (MDD) disorders. This study was a cross-sectional, random assignment survey of psychiatric clinicians who prescribe medication for the treatment of BD and/or MDD. Clinicians ( N = 180) were recruited using national association lists and were presented with either BD or MDD case vignettes that described symptoms, level of functioning, and self-reported medication adherence. Cases were randomly assigned to vignettes that presented subjective (self-report) or objective adherence (from refill rates and MEMS caps data) information. Respondents were asked to recommend adjustments in medication regimen. A total of 180 clinicians participated. In the BD arm ( N = 90), participants were primarily female (55.5%), and specialized in psychiatry (97.8%). In the MDD arm ( N = 90), participants were primarily female (64.4%) and specialized in psychiatry (98.9%). The BD arm was more likely to modify antipsychotic treatment as opposed to mood stabilizers and the MDD arm was more likely to modify antidepressant treatment as opposed to antipsychotic prescriptions. Among vignettes indicating non-adherence, cases that reviewed objective adherence data were more likely to adopt medication adherence enhancing interventions including long-acting injectable antipsychotics. As expected, the rate of LAI choice overall was greater for cases reviewing BD compared to those reviewing MDD vignettes. This study utilized a convenience sample and synthetic vignettes. The survey response arrays do not cover the full array of treatment options available to all clinicians. These data suggest the presence of objective adherence data may help clinicians distinguish between medication nonresponse and medication nonadherence as illustrated by the relationship between the presence of objective adherence information and selection of treatment modifications.