Our recent experience with the COVID-19 pandemic has underscored several issues, importantly that health care disparities continue to exist and that they can drive outcomes. Real-world data provide insights into health care disparities that we do not see in clinical trials—disparities that may impact access to care and adherence to treatment.

Randomized clinical trials are necessary for drug approvals, and extended follow-up from these seminal CML trials has demonstrated the long-term efficacy of TKIs, while also highlighting significant differences in their respective toxicity profiles. However, for various reasons, individuals who participate in clinical trials are not your average patients and unfortunately seldom represent the demographic makeup of the general population. Clinical trials tend to represent a population of patients with a higher socioeconomic status and fewer minorities. This shift in clinical trial populations tends to be skewed as a result of the screening process and eligibility requirements, among other factors, including access to health care. Importantly, patients with a low socioeconomic standing may be excluded from clinical trials because they do not feel engaged or because of practical reasons, such as difficulty with transportation to and from academic centers, as the follow-up requirements are often more intensive.

Thus, real-world data can be very enlightening, and there may be patterns in the community that are not observed in academic practices. Real-world data can be used to estimate compliance or adherence to TKIs by examining pharmacy data and by looking at the proportion of days covered by the prescribed medications. At ASH 2021, Hou and colleagues used a simple method to estimate adherence (ie, the proportion of days covered [PDC]), and they defined suboptimal adherence as a PDC of less than 85% (abstract 4128). In their retrospective analysis of a random sample of 5000 patients who filled at least 1 prescription for a TKI therapy, they found that only 55% of patients had optimal adherence. These findings are eye opening. We see robust responses in clinical trials, and we expect to see those same responses in the community setting, but this is unlikely—especially if adherence is that poor. The reality is that adherence is lower in the general population than in clinical trials. So, I think that all clinicians need to acknowledge this observation and do what we can to try to make sure that our patients are more compliant.

Adequate follow-up and monitoring in CML are other areas that may be impacted by socioeconomic factors. Some patients may not be able to see their doctor and therefore have labs drawn at the recommended frequency because it requires them to take time off from work. As for monitoring, some patients may work in industries with working hours that conflict with the lab’s hours of operation. For example, maybe Saturday appointments are not offered, resulting in patients not being able to come in for their polymerase chain reaction (PCR).

At ASH 2021, Shallis et al described their use of the Surveillance, Epidemiology, and End Results–Medicare dataset (2007-2017) to assemble a cohort of newly diagnosed CML patients aged 66 to 99 years who had started TKI therapy (abstract 282).. The aim was to estimate monitoring frequency and treatment adherence. To investigate qPCR monitoring frequency, researchers examined qPCR tests at 3, 6, 9, and 12 months (all ±30 days) after TKI initiation. They defined optimal monitoring as having a qPCR test at 3 or more milestones during this year of treatment. During the study period of 2015 to 2017 (ie, the latest study period), only 32% had optimal monitoring by this definition. Those with optimal monitoring were more likely to reside in a high socioeconomic neighborhood and be non-Hispanic white. Treatment adherence was defined as a PDC greater than 80%, and, compared with less monitored patients, those with optimal qPCR monitoring were more likely to be adherent to TKI therapy.

Socioeconomic constraints being at odds with monitoring frequency is something that I see in my practice regularly. A patient will say something like, “Doc, I know that you want to see me every 3 months, but I can’t take the time off from work.” And, if we look at patients in the general population, we see the same thing. My interpretation is that real-world data are showing us the impact of health care disparities. For both treatment adherence and monitoring, compliance is not as it should be, and therefore one cannot expect to achieve the same outcomes that are observed in clinical trials when, in the real world, patients are not taking their TKI regularly and are not being monitored as frequently as needed. So, I think that the solution lies in trying to acknowledge these data and working to improve access to health care.