• major depression;
  • pharmacotherapy;
  • treatment;
  • adherence;
  • patient characteristics


Objective: In spite of high levels of antidepressant nonadherence frequently observed among depressed samples, relatively little research has investigated psychosocial predictors of adherence. Fostering greater collaboration in depression treatment to increase adherence has been advocated, but this strategy has not been adequately studied. The aim of this study was to examine the interaction of provider collaboration and patient reactance in the prediction of antidepressant adherence during the acute treatment phase. Method: Fifty outpatients diagnosed with major depressive disorder beginning antidepressant treatment within psychiatry clinics of the VA San Diego Healthcare System comprised the study sample. Patients were administered questionnaires following their medication evaluations to measure predictor variables. Antidepressant adherence was assessed via brief telephone interviews 3, 6, 9, and 12 weeks after treatment initiation. The roles of provider collaboration, patient reactance, and their interaction in adherence were examined using multiple regression analyses. Results: The interaction between provider collaboration and patient reactance accounted for 18.3% of the variance in 3-week adherence (P<.01). Among more reactant patients, greater levels of collaboration predicted better adherence, whereas among patients lower in reactance less collaboration predicted better adherence. No relationships were observed beyond the initial 3 weeks of treatment. Conclusions: This study demonstrates that interpersonal process variables are important in influencing antidepressant adherence and challenges the advocacy of more collaboration in antidepressant treatment as a “blanket strategy.” Establishing a more collaborative set with reactant patients may ensure greater early treatment adherence, a critical period during which antidepressants have typically not yet taken effect. Depression and Anxiety, 2009. © 2008 Wiley-Liss, Inc.