My Papers

AMCP Nexus Poster: Adherence Patterns for Oral Atypical Antipsychotics in Patients Diagnosed with Schizophrenia

If you are attending AMCP Nexus in October, I am a co-author on a poster presentation titled “Adherence Patterns for Oral Atypical Antipsychotics in Patients Diagnosed with Schizophrenia” co-authored with Joanna MacEwan, Felicia Forma, Ainslie Hatch and Darius Lakdawalla.  The presentation abstract is below and here.



Many payers rely on claims-based metrics—such as those from Healthcare Effectiveness Data and Information Set—that measure adherence over a lengthy 1-year period. For patients with schizophrenia, however, missing 25% of the prescribed oral antipsychotic dose over just a 2-week period results in a clinically meaningful risk of psychotic exacerbation.



This study describes how schizophrenia patients’ adherence to atypical antipsychotics varies over a range of time periods using a group-based trajectory model (GBTM).



Using the Truven Health MarketScan Commercial Claims and Medicaid Databases, as well as Humana’s Medicare database (2007-2013), we identified patients with a schizophrenia diagnosis initiating oral atypical antipsychotics. We required continuous enrollment for 12 months after the initial prescription and no atypical antipsychotic prescriptions in the 6 months prior to index. We applied GBTM using a third-order polynomial to fit adherence trends and multinomial logistic regression to model a patient’s probability of belonging to each adherence group. We controlled for patient demographics, comorbidities, and prior substance abuse.



The 39,746 patients with schizophrenia meeting our inclusion criteria were divided into 6 adherence groups, as this grouping produced the best model fit. The 6 adherence trajectory groups were best described as: adherent (10%), discontinuation after 3 months (15%) and 6 months (24%), stop-start after 6 months (10%) and 3 months (30%), and discontinuation after one month (11%). Compared

to patients in the adherent group, patients displaying a stop-start pattern after 3 months were more likely to have a history of drug abuse (OR: 1.56, 95% CI 1.34-1.78) and alcohol abuse (OR: 1.75, 95% CI: 1.55-1.95), have a higher Charlson comorbidity index (OR: 1.24, 95% CI: 1.16-1.42), and less likely to be 35-54 years of age (OR 0.48, CI 0.41- 0.55).



Diagnosed schizophrenia patients’ adherence to antipsychotic therapy exhibits three broad patterns: adherent, discontinuing, and stop-start, but the timing of changes in adherence varied within the discontinuation and stop-start groups. Trajectory modeling could be used to identify patients likely to benefit from a variety of new adherence interventions including digital pill-boxes, ingestible sensors, provider messaging, and mobile pill reminders.



Otsuka America Pharmaceutical.

Leave a Reply

Your email address will not be published. Required fields are marked *