How practitioners (and others) can make scientifically viable contributions to clinical-outcome research using the single-case time-series design

Abstract

Although clinicians typically possess considerable interest in research, especially about which interventions do and do not work, all too often they dismiss the notion that they themselves can make viable scientific contributions to the outcome literature. This derives from an unfortunate assumption that the only true experiment is a between-groups experiment. There is another form of true experiment that is perfectly compatible with real-world clinical practice: the single-case time-series design. Intensive and systematic tracking of one or a few patients over time can yield viable inferences about efficacy, effectiveness, and, under some circumstances, mechanism of change. This paper describes how clinicians working with hypnosis can carry out such research. The rationale and essential features of time-series studies are outlined; each design is illustrated with actual studies from the hypnosis literature; and new methods of statistical analysis, well within the statistical competence of practitioners, are described.

Pediatric Emotional Dysregulation and Behavioral Disruptiveness Treated With Hypnosis:A Time-Series Design

“…Said differently, the symptom levels improved during the first treatment phase, reoccurred during the interruption of treatment phase, and once again demonstrated improvement during the second treatment phase (see Figure 1). According to Borckardt and Nash (2002), "when the results follow this pattern, one can make a strong argument that the intervention caused improvement" (p. 186).…”

“…If the omnibus F test is significant, one then proceeds to examine the two remaining tests: the t test for intercept and the t test for slope. This allows one to determine whether the problem being treated has changed significantly from the first to the second phase (i.e., the intercepts of the two phases are different) and/or if the trend (of getting better or worse) has changed from the first to the second phase (i.e., the slopes of the two phases are different (Borckardt & Nash, 2002). A pertinent point to keep in mind is that, although ITSACORR can maintain an acceptable level of Type I error with as few as five scores per phase, this is not an optimal length and, according to Crosbie (1993), should be considered the absolute minimum.…”

Single-Case Design in Psychophysiological Research. Part I: Context, Structure, and Techniques

“…They can serve to communicate immediately usable clinical data through the scientific literature while helping build a body of referenceable work that demonstrates the potential and constraints of techniques. (Skinner, 1933;Watson & Rayner, 1920), although they have seen a number of developments and refinements over the past two decades that made single-case interventions even more powerful, relevant, and accessible to clinicians in practice (Borckardt & Nash, 2002). Conceptual and analytical refinements have resulted in a consensus that well-constructed single-case design studies can produce powerful statements about both efficacy and effectiveness (Nock, Michel, & Photos, 2007).…”

Efficacy of Hypnotherapy in the Treatment of Eating Disorders

“…An excellent guide to single-subject research using time series analysis appears in Borckardt and Nash (2002).…”