Discussion

In this analysis, we report the rates of physician and patient refusal for trial enrollment within a sample of patients who were screened for ISCHEMIA based on moderate or severe ischemia on stress testing. Of those who apparently met clinical inclusion criteria on preliminary screening, over half did not participate due to physician or patient refusal. This rate of physician and patient refusal was disappointing and illustrates an important challenge to recruiting patients to participate in clinical trials. When there is uncertainty about the efficacy of an intervention, the best method to establish evidence is to conduct a randomized controlled trial. Physicians must have equipoise when optimal management is uncertain and explain to patients that there is a gap in evidence to define optimal management of their condition. Indeed, one could question the ethics of practicing without evidence or preventing a patient from being informed about or participating in a clinical trial when there is uncertainty regarding the best management strategy. In the case of ISCHEMIA, despite community equipoise whether revascularization reduces clinical event rates in patients with stable ischemic heart disease [Reference Stone, Hochman and Williams4], we encountered challenges to enrolling apparently eligible patients with moderate or severe ischemia into a trial with a 50% chance of being randomized to a conservative (i.e., non-revascularization) management strategy. These decisions are undoubtedly influenced by preexisting patient and physician beliefs regarding the risks and benefits of the treatment strategies tested. Notably, sites in the North American region had the lowest rates of enrollment and were more likely to have physicans refuse study participation than other study regions.

Data show that women are less likely to undergo stress testing, a phenomenon in part attributable to implicit sex bias of cardiologists [Reference Daugherty, Blair and Havranek5]. ISCHEMIA also found that women who were screened for study participation were less likely to consent as compared with men, with sex-based disparities more pronounced among clinically eligible participants [Reference Reynolds, Shaw and Min6]. This suggests that the observed sex-specific disparities are not fully explained by differences in the prevalence of obstructive coronary disease between men and women with similar degrees of ischemia. Female sex has been associated with lower clinical trial participation rates in other cardiovascular and stroke randomized trials, threatening the generalizability of study results [Reference Martin, Ou and Newby7, Reference O’Neill, Deptuck and Quong8]. Similar to our study that documents that women were more likely to decline study participation, one study found that women were 15% less willing to participate in cardiovascular disease prevention trials, perceiving a greater chance of harm than benefit from trial participation [Reference Ding, Powe, Manson, Sherber and Braunstein9]. Given the high burden of cardiovascular morbidity and mortality in women, increasing representation of women in practice-informing clinical trials is of paramount importance. Analyses from ISCHEMIA have already demonstrated sex differences between participants in terms of angina severity, anatomic severity of disease, degree of ischemia, and risk factor control [Reference Reynolds, Shaw and Min6, Reference Newman, Alexander and Gu10].

A limitation of this study is that the screening log sample does not reflect all patients screened, enrolled, and randomized into the trial. Screening logs were completed intermittently at only 81% of enrolling sites and were skewed toward lower enrolling sites in line with the purpose of screening logs as a tool to assess and act upon enrollment challenges. Two-thirds of those screened in this cohort had clinical exclusion criteria. The screening log did not assess race, ischemia severity (moderate vs. severe), angina severity, the reasons for physician and patient refusal, or clinical outcomes for excluded patients. The proportion of patient and physician refusals is subject to uncertainty as the assessment of clinical eligibility was preliminary, for example, details were not pursued further if there was a refusal. It is possible that the reasons for refusal could have biased the sample in way that affected trial results.

In conclusion, analyses from screening logs from ISCHEMIA, submitted by sites over varying durations of the enrollment period, demonstrate that two-thirds of patients screened had a clinical exclusion criterion. Among apparently eligible patients excluded from trial enrollment, patient and physician refusal accounted for over half of exclusions, and women were less likely than men to agree to participate. These findings underscore the importance of physician equipoise when advising patients about participation in randomized controlled trials. More work is also needed to dismantle reasons for sex-based differences in cardiovascular clinical trial participation.