Ensuring balanced distribution of prognostic factors in treatment outcome research.

• DOI: 10.15288/jsas.1994.s12.70
• Publication Date: 2015-03-05T09:59:33Z
• Authors: R L Stout; P W Wirtz; J P Carbonari; F K Del Boca
• Keywords: Alcoholism; Analysis of Variance; 03 medical and health sciences; Outcome and Process Assessment, Health Care; Clinical Protocols; Patient Selection; Humans; Multicenter Studies as Topic; 0305 other medical science; Randomized Controlled Trials as Topic; 3. Good health

In comparative or matching research involving two more treatments, the equivalence of patient groups is critical importance. past, has either been imposed by balancing, assured statistically randomization. Matching and while useful in many contexts, nonetheless have important limitations, as does simple recent years, a new tool developed that represents compromise between balancing This method, urn randomization, gives clinical investigators options for improving credibility studies at relatively modest cost. Urn randomization systematically based favor balancing. It can be used with several covariates, both marginally jointly, producing optimal multivariate treatment large sample sizes. preserves primary basis assignment to less susceptible experimenter bias manipulation allocation process staff than Disadvantages include fact it difficult implement, violates probability model A number on addictions, including client-treatment trials, summary mechanics presented, guidelines its use are discussed.