Use of Experimental and Quasi-Experimental Methods for Data-Based Decisions in QI

Abstract

Decisions made by quality improvement (QI) teams, as reported in the literature, are usually based on nonexperimental methods for data collection. Pretest-posttest designs, in particular; are common in reports of QI teams' evaluations of changes or interventions. Yet in such designs the results are inherently confounded; it is impossible to rule out alternative explanations for any differences found. As suggested by one study, QI teams can design and implement experimental interventions in relevant organizational processes. In an attempt to reduce the no-show rate for first appointments at a Residents Clinic at the Medical University of South Carolina (Charleston), the Youth Outpatient Improvement Team designed an experiment to test two possible modifications to the admission process. Those seeking services were randomly assigned to one of three groups (the control group or one of the experimental groups). Not only did results not support the team's hypothesis that one of the experimental procedures would produce a lower no-show rate, subjects in the experimental groups were less likely to enter treatment. Given these data, the decision was made to maintain current admission procedures. Since the quality of a decision is dependent on the quality of the data on which it is based, QI teams should consider experimental methods when planning data collection to evaluate their recommended interventions. When such methods are not feasible, quasi-experimental strategies can be used to strengthen the quality of nonexperimental data.