Persistence, Switching, and Discontinuation Rates Among Patients Receiving Sertraline, Paroxetine, and Citalopram

Abstract

To compare persistence, switching, and discontinuation rates among patients taking brand-name selective serotonin reuptake inhibitors (SSRIs). Retrospective cohort study. Data Source. Protocare Sciences managed care database. A total of 14,933 patients with depression, posttraumatic stress disorder, or social anxiety disorder whose prescriptions for brand-name SSRIs were filled from January 1, 1999-June 30, 2002. A total of 5598 patients received sertraline, 4131 citalopram, and 5204 paroxetine. Adherence status was categorized as persistence, switching (from one SSRI to another SSRI), and discontinuation. Persistence was determined based on the number of days' supply of the prescribed drug, with a minimum of 15 days to refill. Survival analyses using life-table survival curves and Cox proportional hazard models were conducted. Age, sex, and copayment were covariates in the Cox proportional models. Sensitivity analysis with a longer time to refill was performed to determine whether the results were sensitive to the algorithm for determining adherence. Compared with patients receiving sertraline and citalopram, those receiving paroxetine had lower rates of persistence (23.79% vs 25.96% for sertraline [p = 0.0093] and 26.56% for citalopram [p = 0.0022]) and higher rates of switching (3.55% vs 3.32% for sertraline [p = 0.5076] and 2.78% for citalopram [p = 0.0359]) and discontinuation (72.66% vs 70.72% for sertraline [p = 0.0258] and 70.66% for citalopram [p = 0.0334]). Survival curves showed that persistence rates with sertraline and citalopram were significantly greater than with paroxetine (p<0.05, log-rank and Wilcoxon tests). Age was an independent predictor of persistence; male sex and copayment were not. The comparisons across SSRIs were robust in the sensitivity analysis that varied the time to refill allowed. Differences in the persistence rates were noted among patients receiving three brand-name SSRIs, with patients receiving paroxetine having lower persistence than those receiving sertraline and citalopram. Observed differences in persistence across SSRIs were not sensitive to model specifications. Prescription of SSRIs that demonstrate better adherence would benefit both the patient and the health care system.