American Journal of Epidemiology Advance Access originally published online on January 6, 2009
American Journal of Epidemiology 2009 169(6):653-656; doi:10.1093/aje/kwn386
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Estimating the Impact of the Discontinuation of Medical Interventions on Health Outcomes
Correspondence to Dr. Noel S. Weiss, Department of Epidemiology, University of Washington, Box 357236, Seattle, WA 98195 (e-mail: nweiss{at}u.washington.edu).
Received for publication September 22, 2008. Accepted for publication November 7, 2008.
The same study designs are used to document the magnitude of the influence of currently used and of previously used treatments on health outcomes. However, because randomized trials rarely are utilized to study cessation of treatment, most of what we know about the experience of former users of a given treatment comes from cohort and case-control studies. These studies generally compare former users with never users or current users, but occasionally different groups of former users are compared according to how long ago the treatment was stopped. The experience of whole populations following relatively rapid and widespread discontinuation of a previously common treatment sometimes can serve as a secondary source of data on the consequences of discontinuation of therapy.
bias (epidemiology); confounding factors (epidemiology); drug interactions; outcome assessment (health care); pharmaceutical preparations; treatment outcome