V S Westerberg1, W R Miller, J S Tonigan. 1. Center on Alcoholism, Substance Abuse, and Addictions, University of New Mexico, Albuquerque 87106, USA.
Abstract
OBJECTIVE: There has been continuing concern that clients who accept randomization into a controlled trial may not be representative of those in noncontrolled trials or ordinary treatment situations. However, it is not possible to test the impact of randomization through a randomized trial. Two parallel studies conducted at the same treatment facility provided an opportunity for a quasi-experimental study to evaluate whether participation in a controlled trial itself affects treatment outcome. METHOD: Two concomitant samples of clients were enrolled during overlapping recruitment periods: one (n = 226) into a randomized clinical trial (RCT) and the other (n = 122) offered treatment as usual (non-RCT). Both samples were given extensive baseline and follow-up assessment. RESULTS: Baseline assessment indicated demographic similarity of the two samples, with somewhat higher problem severity in the non-RCT sample, consistent with the RCT selection criteria. Client retention in treatment was somewhat comparable, and follow-up rates exceeded 90% in both studies. Overall outcomes did not differ for the RCT and non-RCT samples. CONCLUSIONS: It appeared that clients enrolled into an RCT did not differ from those receiving ordinary treatment. Retention was similarly high in both studies, clients completed a comparable number of outpatient sessions, and the number of informal treatment sessions attended during the 6 months of follow-up was comparable. There are some aspects of this study that limit the ability to draw firmer conclusions but despite some pretreatment differences, participation in the RCT did not itself exert an apparent effect on aggregate treatment outcomes.
RCT Entities:
OBJECTIVE: There has been continuing concern that clients who accept randomization into a controlled trial may not be representative of those in noncontrolled trials or ordinary treatment situations. However, it is not possible to test the impact of randomization through a randomized trial. Two parallel studies conducted at the same treatment facility provided an opportunity for a quasi-experimental study to evaluate whether participation in a controlled trial itself affects treatment outcome. METHOD: Two concomitant samples of clients were enrolled during overlapping recruitment periods: one (n = 226) into a randomized clinical trial (RCT) and the other (n = 122) offered treatment as usual (non-RCT). Both samples were given extensive baseline and follow-up assessment. RESULTS: Baseline assessment indicated demographic similarity of the two samples, with somewhat higher problem severity in the non-RCT sample, consistent with the RCT selection criteria. Client retention in treatment was somewhat comparable, and follow-up rates exceeded 90% in both studies. Overall outcomes did not differ for the RCT and non-RCT samples. CONCLUSIONS: It appeared that clients enrolled into an RCT did not differ from those receiving ordinary treatment. Retention was similarly high in both studies, clients completed a comparable number of outpatient sessions, and the number of informal treatment sessions attended during the 6 months of follow-up was comparable. There are some aspects of this study that limit the ability to draw firmer conclusions but despite some pretreatment differences, participation in the RCT did not itself exert an apparent effect on aggregate treatment outcomes.
Authors: David Wendler; Raynard Kington; Jennifer Madans; Gretchen Van Wye; Heidi Christ-Schmidt; Laura A Pratt; Otis W Brawley; Cary P Gross; Ezekiel Emanuel Journal: PLoS Med Date: 2005-12-06 Impact factor: 11.069