It is well recognized that practice-based research represents an ideal laboratory for the conduct of primary care research.1-3 However, the realities of practice-based research can be problematic: How representative are the physicians who are willing to participate in practice-based research network (PBRN) studies? What factors affect the likelihood that a physician will participate in any given study? And what can be done to increase physician participation? Significant strides have been made in answering the question of representativeness of PBRN physicians.3,4 In reporting the factors influencing physician recruitment and participation in 2 practice-based studies in this issue of JFP, de Wit and colleagues and Temte and colleagues provide valuable information.
In the study by de Wit and colleagues,5 the authors are to be commended for a fairly inclusive examination of factors that influence study participation. Given the complexity of the dual cohort and clinical trial studies and the apparent initiation of the study by academic investigators, the low level of participation (8% of network members) was not unexpected. Unfortunately, information on the potential differences between participating and nonparticipating network physicians was not provided, thus precluding an examination of factors that may have played a role in the physician’s decision to initially participate in the research project. A separate analysis of the characteristics of physicians who participated only in the less complicated cohort study (versus both the cohort study and clinical trial) could have shed light on the influence of study complexity and time commitments on study participation. As is often seen in practice-based research, a number of clinicians were particularly dedicated to the project and recruited the majority of study subjects, while some of the physicians did not recruit any patients. In addition to the 3 possible reasons cited by the authors for nonrecruitment among physicians who had agreed to participate, changes in the practice (eg, staff turnover, change in location, partners on leave, vacation schedules), as well as the impact of the study on physician-patient relationships are potential reasons for lack of participation and recruitment.
In the article by Temte and Beasley,6 the authors examine physician study compliance in a smaller group of physicians who participated in a less intensive dyspepsia study. A single self-reported measure of physician study compliance was compared with the physician’s patient volume and the rate of dyspepsia visits. The authors concluded that practice volume had a significant effect on physician reporting, with high volume practices finding it more difficult to recruit study subjects. However, there is an alternative explanation for this finding. Previous research has indicated that physicians who had a personal interest in the research topic and those with a personal relationship with the researchers were significantly better at recruiting study subjects.7 It is highly likely that the 3 academic physician participants (who had the lowest patient volumes, the highest rate of dyspepsia, and the highest compliance with the study protocol) had a significant personal relationship to both the study and the investigators and were more likely to recruit study subjects irrespective of their practice volume.
It is important to consider factors that influence the decision to participate and those that affect the quality of physician participation; both qualitative and quantitative measures are needed. Factors already known to influence the physicians’ decision to participate include: personal interest in the research topic,8 clear clinical relevance of the research question,9,10 a personal connection to the researchers,10,11 collective ownership of the project,9 the influence of study champions,9 clinical revenue costs associated with research participation,12 and the effect of simplified protocols that decrease interference with patient care. Also, one should examine whether the number of recruited patients was influenced by the degree to which ancillary personnel were used to assist in the study, thus increasing the physician’s efficiency at patient enrollment.10 Finally, the quality of study participation may be measured in terms of the timeliness of patient recruitment, satisfaction with study participation, and disruption to the practice.
As each of us designs and implements studies, an evaluation of physician and patient participation at all levels of the study should be incorporated at implementation. Each of us must distinguish the characteristics of those physicians and patients who agree to participate in general practice-based research, as well as those who agree to participate in an individual PBRN study. We must provide an explicit and detailed description of our recruitment methods in an effort to address the questions posed by funding agencies and editors. And for those of us who participate in practice-based research as study subjects or who recruit patients, we must remember that the field of primary care will only be advanced if we thoroughly and thoughtfully participate in research.