European Journal of Clinical Pharmacology

, Volume 68, Issue 1, pp 65–71

New users of antidepressant medications: first episode duration and predictors of discontinuation

Pharmacoepidemiology and Prescription



Clinical guidelines recommend that antidepressant treatment should be continued for at least 6 months. We examined persistence in patients initiated on an antidepressant medication and factors associated with treatment cessation.


We conducted a cohort study of antidepressant initiators in the Australian veteran population. Patients included were those who had filled at least one prescription for an antidepressant between 2005 and 2008 but had no antidepressants in the preceding 12 months. Kaplan-Meier was used to assess the treatment duration and Cox regression estimated the hazard ratio (HR) of ceasing antidepressant therapy.


There were 91,183 antidepressant users of which more than 50% were prevalent users. Among new users (n = 28,585), 50% discontinued within 6 months, and 61% within 12 months. The median duration for antidepressant treatment was 175 days [95% confidence interval (CI), 168–180]. Patients who received psychological or psychiatrist care around the time of initiation had higher persistence (HR: 0.86, CI: 0.77–0.95 and HR: 0.67, CI: 0.61–0.75 respectively), those with psychiatric disorders and dementia were also more persistent (HR: 0.82, CI: 0.77–0.88 and HR: 0.77, CI: 0.72–0.83 respectively). In contrast, patients with cancer and multi-morbidities had a higher risk of discontinuation (HR: 1.10, CI: 1.03–1.18 and HR: 1.16, CI: 1.05–1.28 respectively).


This analysis indicates persistence with antidepressants in the Australian veteran patients is inconsistent with clinical guidelines, with the majority of initiating patients provided with insufficient duration. Enhanced management of patients with characteristics associated with shorter persistence, including cancer and multi-morbidities, may improve antidepressant use.


Persistence Determinants Antidepressant Mental health 

Copyright information

© Springer-Verlag 2011

Authors and Affiliations

  1. 1.Department of Population MedicineHarvard Medical School and the Harvard Pilgrim Health Care InstituteBostonUSA
  2. 2.Quality Use of Medicines and Pharmacy Research Centre, Sansom InstituteUniversity of South AustraliaAdelaideAustralia

Personalised recommendations