Use your browser's back button to choose another title or click here for a New Search.

How to Get the Report

 Open PDF


Bookmark and Share

Statistical Methodology for a SMART Design in the Development of Adaptive Treatment Strategies.

State College, PA : The Methodology Center, Pennsylvania State University, Technical Report #07-82, 2007, 30p.

Alena I. Scott, PhD (University of Michigan), Janet A. Levy, PhD (Center for the Clinical Trials Network, NIDA), Susan A. Murphy, PhD (University of Michigan).

The treatment of drug addiction presents a challenge to clinicians for many reasons including heterogeneity in response to treatment, the chronic nature of the disease, and the high probability of relapse after a response to treatment. As a result, clinicians must make decisions regarding the sequencing of treatments over time. Sequences of treatments which are guided by a patient’s responses to prior treatments have been termed "adaptive treatment strategies." Specialized experimental designs (SMART or Sequential Multiple Assignment Randomized Trials) have been proposed as a way to develop such sequences of treatments.

Using an example of an adaptive treatment strategy for the treatment of prescription opioid dependence (modeled after protocols CTN-0001/2), the authors demonstrate how such a strategy would be refined through a SMART trial. They specify four research hypotheses that the SMART design might answer; two concerning the efficacy of individual treatments within a sequence, and two concerning the efficacy of the sequences themselves. For each hypothesis, the authors present a test statistic and sample size formula. Two of the sample size formulae are newly developed; one for supporting hypotheses about prespecified sequences of treatments beginning with different initial treatments, and another supporting the estimation of the best treatment strategy among those tested. The authors present the results of a set of simulations to evaluate the robustness of the newly developed sample size formulae in the presence of violations of their assumptions. Both formulae performed well in the presence of mild to moderate violations to their assumptions. In conclusion, the authors make recommendations for future methodological research and highlight the promise of SMART trials to untangle the factors affecting relapse and treatment withdrawal in drug addiction treatment. (Report, PDF, English, 2007)

Keywords: Adaptive treatment strategies | Buprenorphine | Buprenorphine/Naloxone | CTN platform/ancillary study | Opioid dependence | Opioid detoxification | Pharmacological therapy | Research design | Statistical analysis | Suboxone | Subutex

Document No: 187

Submitted by the CTN Dissemination Librarians, 11/5/2007.

Levy, Janet A. search
Murphy, Susan A. search mail
Scott, Alena I. search mail
NIDA-CTN-0001 search www
NIDA-CTN-0002 search www


dark blue line
Supported by a grant from the National Institute on Drug Abuse to the University of Washington Alcohol and Drug Abuse Institute.
The materials on this site have neither been created nor reviewed by NIDA.
Updated 4/2009 --
dark blue line