Search | Search by Center | Search by Source | Keywords in Title
Lee ML, Rubenstein LV, Yano EM. Design Strategies and Statistical Approaches to Evaluating Healthcare Provider Behavior Change. Paper presented at: VA HSR&D National Meeting; 2002 Feb 1; Washington, DC.
Objective(s): Health services research interventions that aim to change provider behavior tend to require relatively complex research designs and sampling methods to effectively gauge their impact. These frequently complex interventions act upon health care providers and their care environments. To assure that investigators adequately prepare and plan for these types of trials, this workshop will describe key research designs (e.g., repeated measures, group randomization) and sampling techniques (e.g., hierarchical sampling) needed to draw valid conclusions, as well as the feasibility and cost issues associated with their conduct. Particular in this regard are key issues of power, sample size and randomization that are critical to the development of an adequate research design. The purpose of this workshop, therefore, is to: a. Review various techniques for the development of power and sample size in the context of longitudinal and complex survey schemes for evaluating provider behavior change studies. b. Examine the strengths and weaknesses of competing designs in order to help attendees evaluate the appropriate scheme for their particular research goals. Activities: A multidisciplinary panel of methodologists (biostatistician, epidemiologist and physician health services researcher) will present the basic ideas of longitudinal and complex intervention evaluation designs and sampling methods, providing relatively direct solutions, and describing available software with illustrations on how these designs are employed and how the statistical issues are dealt with. The strengths and limitations of potential approaches will also be described, enabling participants to intelligently apply these techniques to research problems at their home institutions. Target audience: Health services researchers may not be completely familiar with the methodologic and practical issues involved in the conduct and analysis of these types of intervention trials. Those interested in understanding how more complex study designs are used and what the basic statistical issues are in implementing such designs will benefit from workshop participation. Audience familiarity: The workshop is designed for the broad health services research audience with low to moderate levels of statistical understanding required.