Antipsychotic Treatment Communication Patterns: Predictors and Follow-up Risk

Discussing treatment compliance is challenging when patients have serious mental illness.  Little is known about compliance-related discussions between healthcare providers and patients.  The objectives of this analysis included identification of characteristics associated with treatment compliance discussions, and assessment of the impact of discussions on acute behavioral (BH) treatment in a US insured population.
Adult (ages 18-65), fully-insured commercial health plan patients with diagnoses of major depressive disorder (MDD), bipolar disorder (BD) or schizophrenia (SZ), an antipsychotic medication claim and acute BH treatment (inpatient/emergency department; index date) between 01JAN2014 and 31DEC2015 were identified from administrative claims data.  Outpatient BH visits approximately 12-months post-index event were abstracted. Medication-related compliance discussions were identified using pre-specified exact terms and unspecified related terms; descriptive and multivariable analyses were conducted.      
A convenience sample of 90 medical charts (MDD = 26, BD = 47, SZ = 17; mean age = 41, 62% female) were abstracted. Of the 680 BH visits, 62% included any treatment-related compliance discussions, with 21% using specific compliance terms.  Logistic regression showed lowest odds of a compliance discussion among SZ patients (OR=0.562, p=0.337) and highest among MDD patients (OR=1.514, p=0.534).  A Cox-proportional hazards model, assessing the relationship between a compliance discussion and subsequent acute BH event, found patients with ≥2 compliance discussions had lower risk of an acute event (HR= 0.239; p=0.023; 95% confidence interval  0.070 – 0.818).
Treatment compliance discussions were associated with reduced risk of an acute BH event; however discussions appear to occur less often among patients with severe disease.