Adherence to psychiatric medications: Comparing patients with schizophrenia, bipolar disorder and major depression.

Abstract

Findings of three articles reporting results in 1372 stabilized outpatients taking 2454 medications in Spain, Argentina, and Venezuela were combined. Prevalence of good adherence was not obviously different across diagnoses: 69.5% (N=212) for schizophrenia, 66.3% (N=142) for bipolar disorder, and 69.8% (N=521) for depression. Besides the focus on stabilized outpatients, other study biases included use of a research sample; limited to oral medications, ignoring long-acting injectable antipsychotics; and lack of data on active substance abuse, clinical severity, and insight. Logistic regression models explored predictors of good vs. poor adherence. The six self-reported variables studied were pharmacophobia, pharmacophilia, high psychological reactance, high internal health locus of control (LOC), high doctor LOC, and skepticism concerning specific medications. ORs were significant in 56% (47/84) of the statistical tests vs. 24% (23/98) of ORs significant in case of 7 demographic/clinical variables (p=0.001). At least 2/3 of the ORs for pharmacophobia, pharmacophilia and skepticism were significantly associated with adherence in cases and controls, indicating their independence from diagnoses. In need of replication, three other self-reported measures had differential effects on adherence across diagnoses. High psychological reactance was associated with decreased adherence to antidepressant medications in general, or for patients with mood disorders. High internal LOC as associated with poor adherence may reflect the distrust patients with schizophrenia or severe bipolar disorder have of other people. High doctor LOC was significantly associated with increased adherence only in patients with bipolar disorder, but was significant for all medications, mood stabilizers and antipsychotics, indicating the relevance of the patient-psychiatrist relationship in these patients. (Neuropsychopharmacol Hung 2021; 23(4): 363-373).