Diabetes clinical trials located in few socially disadvantaged counties in US
Click Here to Manage Email Alerts
Measures of social determinants of health are substantially different in U.S. counties with diabetes clinical trial sites compared with counties with no trial sites, according to a research letter published in Diabetes Care.
In an analysis of diabetes clinical trials conducted in the U.S., counties used for clinical trial sites had a lower percentage of participants whose household income fell below the federal poverty line, had no high school diploma and were uninsured compared with counties not used in clinical trials.
“County-level social determinants of health were significantly different in U.S. counties that were used for clinical trial sites in studies evaluating glycemic treatment when compared to U.S. counties not used for these sites,” Erin R. Weeda, PharmD, assistant professor of clinical pharmacy and outcome sciences in the College of Pharmacy at the Medical University of South Carolina in Charleston, told Healio. “County-level social determinants of health may be an important factor to consider as clinical trial recruitment is planned.”
Researchers collected data from 152 clinical trials cited in the pharmacological approaches to glycemic treatment section of the American Diabetes Association’s 2021 Standards of Medical Care in Diabetes. Data on the location of the clinical trial sites were obtained from ClinicalTrials.gov. County-level social determinants of health data from 2015 to 2019 were obtained from CDC data sets.
There were 612 unique U.S. counties included in the 152 clinical trials, with each trial including a median of 43 counties. The majority of trials have been published since 2015.
Counties included in clinical trials had a lower proportion of homes falling below the poverty line compared with counties not included in trials (9% vs. 10%; P < .001). After stratifying counties by quartiles according to the percentage of homes falling below the federal poverty line, counties used for clinical trial sites had a lower percentage of counties in the highest quartile compared with counties not used in clinical trials (17% vs. 27%; P < .001).
Counties used in clinical trials also had a lower median percentage of the population without a high school diploma (10.4% vs. 12.2%; P < .001), a lower median percentage of residents without health insurance (8.1% vs. 8.9%; P < .001) and a lower percentage of vacant homes (10.4% vs. 18%; P < .001) than counties not used for clinical trial sites.
“These findings suggest that county-level social determinants of health measures should be considered in the planning and execution of clinical trial recruitment,” the researchers wrote.
Weeda said the findings revealed differences in social determinants of health in clinical trials on a larger scale, but more research must be performed on how social determinants of health at the individual level may impact clinical trial results.
For more information:
Erin R. Weeda, PharmD, can be reached at weeda@musc.edu.
Perspective
Back to Top
In a recent letter, Weeda and colleagues describe an analysis comparing diabetes clinical trials in counties where social determinants of health — including higher poverty levels, absence of high school diplomas, lack of insurance and increased vacant housing units — were more common to trials done in counties where the frequency of such findings was significantly less. The studies analyzed in their report were meaningful inasmuch as the results of these trials were influential in changing treatment guidelines for pharmacologic treatment for glycemic control in diabetes.
Importantly, it has been demonstrated that the level of burden of social determinants of health in persons can have significant impact on outcomes in clinical studies, with those affected by the worst profiles of social determinants of health being the least likely to achieve study goals and more likely to experience worse outcomes. It has also been demonstrated that persons with such profiles of social determinants of health are the most likely to be underrepresented in clinical trials. Thus, the findings from such trials do not adequately capture, and often do not measure, the influence of social determinants of health on outcomes of guideline-directed treatments. When patients fail to demonstrate outcomes predicted by the guidelines, based on clinical trials evidence, it may often be assumed that such failures stem from biological or even racial differences.
To more accurately reflect the effects of a specific clinical intervention in diabetes, clinical trials must be designed in ways that allow assessment of social determinants of health, and there must be adequate numbers of persons with different burdens of social determinants of health to facilitate analysis of the effects of these factors on clinical outcomes. It might be advisable to consider including social determinants of health in the inclusion criteria and the baseline characteristics, captured as county-level data from available datasets. Such an approach would prompt redoubled efforts to increase representation from socially disadvantaged counties in clinical trials, but just as important, it may suggest that we need to broaden and recodify the clinical and social characteristics that we consider most influential on trial outcomes in diabetes. This is clearly an area ripe for further, more intensive study.
Collapse
Click Here to Manage Email Alerts