Single Subject Experimental Design: One Way Forward?

Abstract

In a recent Forum paper in this Journal, Bithell (1994) argued that single subject experimental designs are a cause for concern for two main reasons: the single case method may lack scientific credibility; and there are problematic issues in design, data collection and analysis. Bithell concluded that group designs must be carried out to demonstrate the general effectiveness of a treatment for patients with a particular problem; single case designs alone or in series will not do. Clinical practice should be based on sound evidence. As a profession, we still do not know the type, amount and/or the duration of therapeutic strategies that should be offered to our patients to optimise outcome. While randomised controlled trials may be able to determine which treatments should be offered to individuals who meet the specific inclusion criteria of the trial, they also have their own limitations; in particular the difficulty in obtaining a homogeneous patient group (other practical problems include the difficulties in recruiting sufficient numbers of patients and possible resource implications). Effective physiotherapists choose treatment strategies tailored to suit the needs of a patient with specific problems, not a typical treatment for a typical patient. Single case designs may be the best way of achieving ‘tailored’ treatment; however, if a positive therapeutic finding can be demonstrated for a particular individual, strong claims cannot be made regarding the therapy until the effects of the particular intervention have been replicated with a number of individuals.