Single-subject clinical-outcome research: designs, data, effect sizes, and analyses

In the last 20 years, single-subject research designs have become important forms of aphasia-treatment research for assessing the effectiveness of treatment on a subject-by-subject (or patient-by-patient) basis. In that time, several important developments in the statistical literature centring on the reliability and validity of single-subject research have occurred. This work assesses the state of aphasia-treatment single-subject research in the context of that scholarship and proposes recommendations for future applications through a tutorial-like presentation. This paper details the analysis of published single subject results and proposes recommendations concerning future applications of single-subject designs. The work focuses on four domains: designs, data, effect sizes, and analyses. The findings indicate that aphasia-treatment single-subject studies, which are well designed for the most part, yield a short series of autocorrelated data manifesting generally large treatment effects. However, only one analysis satisfactorily controlled Type I and Type II errors under typical clinical-aphasiology applications. That procedure, ITSACORR, is easily accomplished and it expresses outcome in familiar terms. To facilitate understanding, the review promotes a hands-on understanding of the various analysis options through worked examples and clarifies the (in)appropriateness of each procedure for clinical applications. Although the focus of the work is treatment for aphasia, the central thesis has general application across disorder categories.