Quality-assurance research: studying processes of care.

To conduct a successful research study, several criteria must be satisfied: the research question must be important (not only to the investigators but to clinicians and managers in other institutions); the research question must be answerable with the available resources; the investigators must be motivated and capable; and the research setting must be appropriate for the study (the institution must have a supportive culture and analytical resources, and the local institutional review board must approve the proposed study). Quality-assurance (QA) research poses some special challenges and requirements. First, although QA studies should be hypothesis-driven, they are usually before-and-after studies, rather than randomized controlled trials. Second, in before-and-after studies the investigators must address and minimize several possible sources of bias that could confound the results. For example, the compared groups must be similar in important features that could affect development of the outcome(s) of interest, and clinical practices other than the practice change that is being tested during the study must be shown not to independently affect the outcome(s) of interest. We discuss several examples of QA studies, and we offer a checklist for the process of considering, designing, executing, presenting, and publishing a QA study.