The aim of this study was to examine the association between duloxetine adherence/persistence and hospital utilization. In a managed care claims database, 8521 patients with a major depressive disorder diagnosis were initiated on duloxetine in 2006. Patients had no active duloxetine prescription for 6 months before initiation and had continuous enrollment for 12 months preinitiation and postinitiation. Adherence was defined as medication possession ratio of 0.8 or more, and persistence was defined as the duration of therapy without exceeding a 30-day gap. Logistic regression and negative binominal regression were conducted. Overall, 55.8% of patients were adherent and the average duration of duloxetine therapy was 118.4 days within 6 months after initiation. Adherent patients had significantly lower rates of hospitalization (19.7 vs. 23.4%, P<0.0001) and emergency room visits (30.6 vs. 36.9%, P<0.0001) than nonadherent patients. Hospitalization and emergency room visits were significantly reduced with treatment persistence (P<0.0001). After adjustment for demographics, comorbidities, and prior hospitalization, adherence was associated with reduced hospitalization (odds ratio=0.86) and emergency room visits (odds ratio=0.80). Patients on duloxetine of more than 90 days, compared with less than 31 days, were 16% less likely to be hospitalized and 22% less likely to have emergency room visits. Duloxetine adherence and persistence appear to be associated with reduced hospital utilization in the 1-year follow-up period.
aLilly Research Laboratories, Eli Lilly and Company
bDepartment of Psychiatry, Indiana University School of Medicine, Indianapolis, Indiana
cUniversity of Pittsburgh, Graduate School of Public Health, Pittsburgh, Pennsylvania, USA
Correspondence to Dr Xianchen Liu, MD, PhD, c/o Kory Schuh, Eli Lilly and Company, Drop Code 4135, Lilly Corporate Center, Indianapolis, IN 46285, USA Tel: +1 317 276 6627; fax: +1 317 651 1726; e-mail: firstname.lastname@example.org
These data were presented at the International Society for Pharmacoeconomics and Outcomes Research; Atlanta, Georgia, USA; 15–19 May 2010.
Received August 27, 2010
Accepted December 14, 2010