African Americans had differentially lower enrollment rates in the Look AHEAD trial compared to all other volunteers screened. The reasons for this lower enrollment were not attributable to losses to follow-up or attrition during the screening process; the percent of African American screenees who were lost to follow-up or refused to continue screening procedures did not differ from other screenees. The differential rates of enrollment were due primarily to failure of eligibility criteria, including criteria related to poor control of chronic disease, to conditions that were thought to limit the lifespan of participants and to interfere with ability to participate safely in the trial, and characteristics and behaviors that were judged to predict poor adherence procedures if enrolled.
The Look AHEAD trial did not provide direct medical care and required participants to identify their source of care. Individuals with poorly controlled hypertension or diabetes were excluded from the trial. In addition, participants had to have verification of T2DM (e.g., by medical records, current treatment, verification from personal health care provider, or test result). These health-related eligibility criteria differentially excluded African Americans, who as a group in the United States have higher rates of uncontrolled hypertension and diabetes and lower rates of access to health care [17
]. Look AHEAD allowed volunteers who initially did not meet these criteria to be rescreened at later dates, and assistance was provided, when needed, to find sources of medical care. However, the trial funding and design did not provide for direct intervention. Some of the reasoning behind this was to separate the source of medical care from the investigators administering the trial’s unmasked behavioral intervention, thereby reducing the potential for confounding. However, had there existed separate avenues and funding for administering the health care needed to bring hypertension and diabetes into control, it is possible that greater numbers of African Americans would have been eligible and would have enrolled.
Our findings raise the issue of whether the Look AHEAD criteria for exclusion were justified. Abnormal heart rates, chronic heart conditions, and evidence of renal disease were adopted as exclusion criteria because the behavioral interventions in Look AHEAD may not have been safe for individuals with these conditions and because such conditions may have interfered with their ability to complete the trial. These criteria differentially excluded African Americans who have greater burdens of renal and heart diseases than non-African Americans [22
Look AHEAD used a behavioral run-in task as part of screening for participants who were likely to adhere to trial procedures; in the run-in, candidates were required to record information about diet and physical activity daily during a 2-week period. The use of behavioral run-ins and reliance on staff judgment are important components of trial enrollment. Look AHEAD clinic staffs were centrally trained to promote cultural awareness both in providing instructions regarding run-in tasks and in evaluating participants. Prior to randomization, the local study team met to review each volunteer’s screening data to determine whether there was consensus that the individual was an appropriate candidate for the trial. Consideration was given to safety, whether there were inconsistencies between entries on forms and self-reports, and concerns about adherence. However, unlike most reasons for exclusion, this one could be challenged as subjective, despite the training in cultural sensitivity.
The requirements related to diabetes control and access to a regular health care provider may have acted to account for the higher rates of exclusions related to diabetes treatment. However, for most subgroups of screenees we examined, the differential rates of exclusion of African Americans were consistent.
The lower enrollment rates for African Americans had a relatively larger effect on the enrollment of male patients and those whose diabetes was either untreated or required insulin treatment (). Many trials have reported difficulties in enrolling African American male patients [26
]. Speculations concerning low participation rates of African American male patients are multipronged, ranging from prior history of biomedical research mistreatment, racial concordance, awareness about research studies, clinical trial literacy, and factors that are more socioeconomically driven: job flexibility, transportation, housing, and neighborhood factors, which could adversely affect participation continuity [31
The African American enrollees in the Look AHEAD trial tended to differ from others with respect to many demographic, health, and behavioral characteristics, thus stressing the importance of including these individuals in clinical trials to ensure generalizability of findings regarding the effects of interventions. Because losses to follow-up and refusal were reported as a single category, we cannot comment on the approximately 41% of volunteers who were possibly eligible but then did not continue with screening or enroll in the trial. Although Look AHEAD included clinical sites located throughout the United States, recruitment areas necessarily were limited geographically and may not reflect the general populations. Furthermore, this analysis was conceived post hoc; more information could have been collected, for example, in an ancillary study, if it had been planned ‘up front’.