Use of sequential medical trials in rehabilitation research.

Abstract

Although randomized, clinical trial designs represent the pinnacle of research excellence, they are often difficult to implement in natural or clinical settings. An alternative approach to analysis of data collected during rehabilitation research is the sequential medical trial design. In a sequential medical trial, subjects are serially recruited and the results are continuously analyzed. The use of preconstructed sequential charts remove the need for clinicians to perform complicated statistical analyses and allow an immediate visual indication of the direction the results are taking. Benefits of sequential medical trials are as follows: as soon as enough data are collected to show treatment preference, the trial is stopped; a trial that is showing no beneficial or even harmful effects can be quickly terminated; fewer subjects are needed (without loss of statistical power); features such as randomization, blinding, and cross-over designs can be included; the design suits the serial nature of rehabilitation practice. Limitations to sequential medical trials are that in simple designs extraneous factors cannot be controlled for and that multiple dependent variables are difficult to assess. This article discusses the theory behind the sequential medical trial design and outlines how to plan a sequential medical trial, recruit subjects, construct and use a simple sequential chart, and analyze and interpret the results. Examples of sequential medical trials used in rehabilitation research are presented. It is hoped that this article will increase the exposure of this useful design and result in greater use of sequential medical trials by rehabilitation professionals.