DISCUSSION

In this systematic review, reporting of older patients and ethnic/racial minorities was poor with <20% of the included RCTs mentioning these demographic subsets. Enrollment of women and American Indian/Alaskan natives in RCTs of AF matched their demographic share of disease burden. Hispanics and Asians were over-represented and Blacks, native Hawaiian or Pacific Islander and non-Whites were under-represented in RCTs compared with real world data. Representation of women was significantly higher in industry or academic center sponsored RCTs compared with government funded RCTs. Mean age of participants was less than the corresponding weighted mean age in the comparative epidemiological data, and varied significantly based on location, funding sources, baseline population and according to management strategies used for prevention of stroke or treatment of AF. The 1993 National Institute of Health Revitalization act legally required RCTs to include men and women consistent with the known sex related prevalence of the disease under study. More recently, the Government Accountability Office (GAO) issued a follow-up report calling for improved reporting of women’s enrollment in RCTs29. Since RCTs are regarded as the ”gold standard” for shaping up the management strategies and guidelines, inclusion of women and demographic subgroups sustain the generalizability of the findings to the population as a whole and allows subgroup analyses to determine influence of different ethnicities. The current findings are in line with recent review by Scott and colleagues who also reported adequate representation of women in the RCTs that supported FDA approval of cardiovascular drugs15. However, women with AF were reported to have lower quality of life, higher risk of ischemic stroke and higher mortality as compared to men 5-7. Therefore, such complexity of AF in women endorses the need for adequate enrollment in future RCTs as well. Our finding of lower proportion of women in government funded RCTs needs attention in this regard. The FDA guidelines in 1977 called for exclusion of women of childbearing potential from early phases of RCTs. Even though, such policies have been revised since that time, it is still plausible that this may have deterred female recruitment in RCTs especially the ones sponsored by the Government agencies30. Additionally, lack of awareness about RCTs and logistical barriers might also have contributed to the lower enrollment rates. Co-morbidities leading to AF vary based on age, sex and ethnicities31, 32. The prevalence of AF rises steadily with age, and Blacks have a 2-5 times higher risk of AF-associated stroke than Whites. However, the enrollment of older patients was reported in only 18.7% of the included RCTs. Similarly, only 12.7% RCTs reported ethnic/racial population, out of which Black patients made up disappointingly <2% of cohorts. In order to improve enrollment of these groups, following strategies can be considered. First, greater cultural sensitivity is needed to ensure adequate recruitment and consent procedures that are consistent with different ethnic/racial cultures 33. Second, targeting inner city population that are likely to have a high minority ethnic population and hiring special advocacy workers to provide a bridge between recruiters and minority population 34. Third, addressing the obstacles for women participation such as offering childcare or transportation, having special considerations when enrolling fertile women, and ensuring they have access to counseling and medical care 35. Fourth, an Office of Geriatric Health and Aging can be created to review protocols and enrollment of older population similar to the workings of Office of Women’s Health which was established by Congressional mandate in 1994 32. Fifth, simplifying consent forms, adding time to consult with family, the use of proxy data or remote follow up can be elicited to overcome barriers to participation of underrepresented demographics in RCTs32. Our study has certain limitations. Since we did not have access to individual participant’s data, this review was limited to trial-level information. We relied on published trials and thus risk of publication bias cannot be ignored. Overall, there was inadequate reporting of older patients and ethnic/racial minorities in AF RCTs. The representation of racial demographics did not start until 2007 in AF RCTs with reporting on certain races such as Native Hawaiian or Pacific Islanders beginning in 2011. This limit inferring strong conclusion around such racial demographics. Furthermore, some RCTs have only reported race as White or non-White which makes detailed analysis of non-White racial groups impossible from those studies. Additionally, one can argue that treatments in RCTs are examined to determine effects in population at risk, and naturally stroke prevention trials would reflect older cohorts given higher inherent risk of stroke compared with younger participants in catheter ablation trials. In conclusion, this systematic review suggests adequate enrollment of women, older patients and American Indian/Alaskan natives, over recruitment of Hispanics and Asians and under recruitment of Blacks, non-Whites and native Hawaiian or Pacific Islanders in Trials of AF compared with their demographic share of disease burden. These data demand a diligent review of the policies by organizations and investigators to ensure adequate representation of these demographic subgroups in future RCTs.