Clinical case studies: their place in evidence-based practice

Abstract

The study, documentation and communication of case studies are central to the practice of psychotherapy. However, case studies lie low in the hierarchy of evidence generally used by researchers, health providers and purchasers. This paper examines why this should be. Ways are proposed by which they can be improved for scientific purposes and become more than exercises in persuasion.

Keywords:

• Psychotherapy case studies
• strengths
• weaknesses
• research
• evidence-based practice

Notes

The hierarchy is not absolute but there is substantial overlap between different systems. In the UK, the Cochrane database publishes and commissions reviews of the research literature. Their hierarchy can be found at www.cochrane.org. Also in the UK, the National Institute for Clinical Excellence (NICE) examines the literature and recommends optimal treatments (www.nice.org.uk). The criteria used by the American Psychological Association's Task Force on Promotion and Dissemination of Psychological Procedures in the mid-1990s to identify empirically supported therapies were based on robust, independent RCTs demonstrating significant advantage over alternative bona fide treatments or placebo (Chambless, D. L., & Hollon, S. D. (1998). Journal of Consulting & Clinical Psychology, 66, 7 – 18).

Idiographic measures focus on the single individual and make no comparison with others e.g. repertory grids and Q-sorts. Nomothetic measures compare individuals with others and may be norm referenced e.g. IQ tests.

I like the stories told by Irving Yalom. Early in his career, he published a fascinating account of a long-term individual therapy in which he and the patient wrote lengthy accounts of each session and swapped the accumulated record every six months. Intriguingly though they had much experience in common (the therapy hours), their account was very different. He valued his elegant interpretations. She barely heard them and instead prized small personal acts of recognition and instances of his human fallibility (Yalom, I. D., & Elkin, G. (1974). Every Day Gets a Little Closer. Basic Books: New York. Recently his writing has become less discursive and more distilled. The vivid stories read like parables. For an example, see Yalom, I. D. (1999). Momma and the meaning of life. London: Piartkus.

In developing theory and practice, Dora was one of Freud's formative cases. For an example of the different take that authors may have on the published data, see the feminist, literary and philosophical compilation: Bernheimer, C., & Kahane, C. (Eds.) (1985). In Dora's Case. London: Virago Press.

Clinical work begins with the attempt to understand how the world appears from the patient's perspective. Research and audit need to hear the patient's voice and views, not because they are necessarily right, but because they are vital to the dialogue and illuminate the experience of the person for whom the work of therapy is being done. The range of methods for self-report is wide. Patients may write their own account of sessions, either as part of therapy or for publication. Publication accounts may be co-authored by patient and therapist. Post-therapy, some patients may choose to describe their experience in detail. Consuming Psychotherapy by Ann France (1988). London, Free Association Books is a fascinating example, which depicts three therapies over eight years. Q-sorts are a well-established research method to tap into individual perception. Interpersonal process recall is a systematic way to review the significance of interactions within sessions (Elliott, R. and D.A. Shapiro (1988). “Brief structured recall: a more efficent method for studying significant therapy moments.” British Journal of Medical Psychology 61: 141-153). Finally, it has been shown that making outcome is improved by making the therapist aware of deterioration in the expected trajectory in change during therapy as determined by patient ratings of their symptom severity (Lambert, M., N. B. Hansen, et al. (2001). “Patient-focused research: using patient outcome data to enhance treatment effects.” Journal of Consulting & Clinical Psychology 69: 159-172).

For a discussion of how discrete choice experiments might influence the choice of healthcare, see Ryan, M. (2004). British Medical Journal, 328, 360 – 361.

Effect size is a statistical measure of the strength of the underlying relationship in which you are interested and is the basis of power analysis, which is used to determine the necessary sample size for robust results (Cohen, J. (1988). Statistical Power Analysis for the Behavioural Sciences, Erlbaum, New Jersey: Hillsdale.