The Potential of Clinical Data-Mining in Evidence-Based Practice Knowledge-Building

Although the research use of available administrative and practice data is not new, Clinical Data-Mining (CDM) is a more systematic practice-based research approach whereby social work and allied health practitioner-researchers “mine” their own qualitative and quantitative data to describe the people they serve, the interventions they provide and the outcomes they achieve. In addition to practitioners, a growing number of practice-oriented social work doctoral students have employed CDM as a dissertation research strategy by itself or in combination with other more conventional research designs and data-collection strategies.

Because CDM outcome studies are essentially retrospective, frequently relying on non-standardized outcome data and non-randomly assigned client groups, they rarely rival randomized controlled experiments in promoting causal inferences. As a result, CDM studies are usually excluded from meta-analytic studies or systematic reviews. Nonetheless, CDM has an important role to play in the future development of empirical research knowledge of “best practices” and in the generation of a fully “evidence-informed” practice.

This roundtable will involve two senior social work research academics who have employed and facilitated CDM with practitioners and PhD students over the course of their careers and three recent social work CDM/PhD recipients—one from the City University of New York and the others from the University of Hong Kong.

The Roundtable will begin with an overview of CDM as a practitioner-research and a PhD dissertation research strategy. Discussion of three CDM/PhD dissertation exemplars will follow. The first exemplar represents the first explicitly labeled a CDM dissertation.  Conceived as a feasibility study of CDM as a dissertation methodology, it tested the fidelity and effectiveness of Intensive Family Preservation Services in a single  IFPS setting in Vermont.

The second, a national “best dissertation” award-winning, mixed-method study, employed CDM by converting available qualitative administrative information into original quantitative research data in order to assess program fidelity of a multi-disciplinary consultation in child protective services in New York City.  In so doing, the study generated “organizational best practices” for implementing multi-disciplinary programs.

The final exemplar dissertation employed CDM to study the effectiveness of Group Cognitive Behavioral Therapy (GCBT) with Hong Kong Chinese adults suffering from anxiety and depression. Using standardized clinical assessment measures that the host agency routinely employed, this study measured the effect size of GCBT and identified personality characteristics and other contextual factors that moderated the effects of GCBT. On discovering that Group CBT was not consistently helpful, the researcher initiated an RCT employing a mindfulness intervention also addressing anxiety and depression. The results of the two intervention studies were then compared.

Using these and other CDM/PhD exemplars, the final presenter will discuss the potential uses as well as the limitations of CDM in contributing to an evidence-base for social work interventions at the individual and organizational levels. The strengths and limitations of CDM as a dissertation research strategy will be discussed as well by the recent PhD presenters.