An overview of longitudinal data analysis methodologies applied to the dose -response relationship in psychotherapy outcome research

There is a growing consensus that psychotherapy research is having minimal impact on actual clinical practice. There is also a growing trend by researchers to remedy this by producing research that is both practical and applicable to clinical practice. Clinical significance methodology, dose-response research, and phase models of patient change are three such areas where practical applications of research findings to clinical practice are possible. To date, there have been a wide variety of approaches applied to understanding the dose-response relationship in psychotherapy, including: (1) general linear modeling, (2) logistic regression, (3) actuarial analysis using life tables, and (4) survival analysis. The current study provides an overview of each of these methods and then covers comprehensive applications of these methods using a large data set (N = 6331) to identify the most useful manner of assessing the dose-response relationship. Within these applications, clinical significance methodology has been used to address the meaningfulness of patient change. The same statistical methods listed above have also been applied to the present data set to identify differences in differential symptom improvement along the lines of a phase model of patient change. The present study has identified non-parametric statistical approaches, in conjunction with clinical significance, as the methods of choice for assessing change over time when considering the dose-response relationship and phase models of change. Survival analysis, in particular, has stood out as the most useful and correct method of assessing change over time within the questions addressed in this work.