1a. We suggest that the AMA definition of CME and the terms articulated in this guideline (or their modifications) be consistently employed by CME practitioners and researchers as a basis for the development and study of CME interventions.
1b. We suggest widespread dissemination, elaboration and clarification of these terms by journal editors, professional societies, and by the research community.
2. We suggest that increased funding be made available for CME research, enabling use of the most rigorous methods in original studies and systematic reviews. We recommend that such funding be carefully determined by the scope and precision of the research question in each case.
3. We suggest that searches employ an information specialist and extend beyond the traditional medical educational literature to incorporate databases established to encompass the role of CME in quality improvement, guideline utilization, managed care, business and organizational development, informatics, and other domains.
4. We suggest that systematic review processes of CME interventions undertake rigorous efforts to ensure high levels of definitional agreement, independent data abstraction by more than one researcher, and assessment of interrater reliability.
5. We suggest that systematic reviews of studies of CME interventions define and employ well-described and commonly agreed-on constructs of what constitutes positive, negative, and mixed outcomes. In this process, careful attention should be paid, where methodologically feasible, to questions of statistical, educational, and clinical significance, and of the magnitude of the effect (eg, effect size, coefficient of determination).
6. We suggest standardized definitions, methods, and reporting structures be developed and used for future research, systematic reviews, and guidelines.
7. We suggest that researchers explicitly consider the inclusion and documentation of teaching and learning principles in the design and implementation of further trials of CME. In addition, we suggest that, whenever possible, trials be designed to study the educational outcomes of such variables.
8a. We suggest that comprehensive models of change such as those developed in knowledge translation be employed when studies of the effect of CME are undertaken, in order to consider and assess the role of unaccounted and dependent variables.
8b. We suggest that future studies of CME interventions incorporate full descriptions of elements expressed in the continuing health-care education study template (or CHEST).
8c. We suggest that randomized controlled studies (1) be performed with a clear definition of intervention and comparison or control groups, (2) have their effects measured at multiple points after intervention, and (3) pay close attention to issues of participation and dropout.
8d. We suggest that researchers consider the value of rigorous observational, ethnographic, and other qualitative study methods, and use them either separately or in conjunction with quantitative methods and designs.
9. We suggest that leaders in medical education and related fields (1) foster the identification of high priority research topics in CME research that would span the broad scope of CME and (2) conduct scientifically rigorous studies of the process and effectiveness of CME.