Does knowledge of patient non-compliance change prescribing behavior in the real world? A claims-based analysis of patients with serious mental illness

Abstract

Jason Shafrin,1 Katalin Bognar,1 Katie Everson,1 Michelle Brauer,2 Darius N Lakdawalla,3 Felicia M Forma4 1Policy and Economics, Precision Health Economics, Los Angeles, CA, USA; 2Policy and Economics, Precision Health Economics, Boston, MA, USA; 3School of Pharmacy, Sol Price School of Public Policy, Leonard D. Schaeffer Center for Health Policy and Economics, University of Southern California, Los Angeles, CA, USA; 4Health Economics and Outcomes Management, Otsuka Pharmaceutical Development & Commercialization, Inc., Princeton, NJ, USA Background: New digital technologies offer providers the promise of more accurately tracking patients’ medication adherence. It is unclear, however, whether access to such information will affect provider treatment decisions in the real world.Methods: Using prescriber-reported information on patient non-compliance from health insurance claims data between 2008 and 2014, we examined whether prescribers’ knowledge of non-compliance was associated with different prescribing patterns for patients with serious mental illness (SMI). We examined patients who initiated an oral atypical antipsychotic, but were later objectively non-adherent to this treatment, defined as proportion of days covered (PDC) <0.8. We examined how a physician’s awareness of patient non-compliance (ICD-9 diagnosis code: V15.81) was correlated with the physician’s real-world treatment decisions for that patient. Treatment decisions studied included the share of patients who increased antipsychotic dose, augmented treatment, switched their antipsychotic, or used a long-acting injectable (LAI).Results: Among the 286,249 patients with SMI who initiated an antipsychotic and had PDC <0.8, 4,033 (1.4%) had documented non-compliance. When prescribers documented non-compliance, patients were more likely to be switched to another antipsychotic (32.8% vs 24.7%, P<0.001), have their dose increased (24.4% vs 22.1%, P=0.004), or receive an LAI (0.09% vs 0.04%, P=0.008), but were less likely to have augmented therapy with another antipsychotic (1.1% vs 1.3%, P=0.035) than patients without documented non-compliance.Conclusion: Among SMI patients with documented non-compliance, the frequency of dose, medication switches, and LAI use were higher and augmentation was lower compared to patients without documented non-compliance. Access to adherence information may help prescribers more rapidly switch ineffective medications as well as avoid unnecessary medication augmentation. Keywords: adherence, prescribing patterns, serious mental illness

Keywords

• adherence
• prescribing patterns
• serious mental illness