Disparity in the rural‐urban distribution of clinical trial sites used in studies cited by diabetes treatment guidelines

1 INTRODUCTION

The prevalence of diabetes and the associated complications are higher in rural areas in the United States.1 Representation of rural residents in diabetes clinical trials is thought to be low because trials are often conducted at healthcare institutions in urban areas. However, this has not been quantified. We sought to determine the proportion of rural US clinical trial sites that were used in guideline-changing studies evaluating pharmacological treatment for glycaemic control and to quantify the distance between each US county and the nearest clinical trial site.

2 METHODS

All citations in the 2021 American Diabetes Association “Pharmacologic Approaches to Glycemic Treatment–Standards of Medical Care in Diabetes” guideline were screened for inclusion.2 To be included, a study had to: (i) be a randomized controlled trial; (ii) include ≥1 US study site; (iii) evaluate pharmacological interventions for the glycaemic treatment of type 1 or type 2 diabetes; and (iv) report location information on clinicaltrials.gov. Original articles included in systematic reviews or meta-analyses cited by the guideline were also screened and individual studies were included if they met the screening criteria. Data on the location of US clinical trial sites (ie, zip code) for each study were obtained from clinicaltrials.gov. Rural and urban status for each site was classified based on the Rural Urban Commuting Area codes established by the US Department of Agriculture.3 The distance between each US county and the nearest clinical trial site was also ascertained. Analyses were performed and a map was generated in R version 4.04.

3 RESULTS

A total of 152 clinical trials evaluating pharmacological treatment for glycaemic control were identified for inclusion. A list of these studies can be found in Appendix S1. Most trials (86%) included individuals with type 2 diabetes (Table 1). The most common interventions evaluated included first- or second-line therapies for type 2 diabetes (66%).

TABLE 1. Characteristics of included studies (N = 152) n (%) Median (IQR) number of participants per trial Diabetes type Type 2 130 (85.5) Type 1 20 (13.2) Both Type 1 and 2 2 (1.3) Intervention Insulin 48 (31.5) Metformin or second-line therapya 100 (65.8) Other 4 (2.6) Multinational trial 124 (81.6) Median (IQR) number of US clinical trial sites per study 43 (21-74) Abbreviations: IQR, interquartile range.

The included clinical trials used 8898 study sites located in 1961 unique US zip codes. Of all sites, only 4.9% (n = 437) were in rural areas. Figure 1 shows the distance between each US county and the nearest clinical trial site. Of 3143 US counties, 51% (n = 1612) were > 30 miles from a clinical trial site (Appendix S2).

image

Distance between US counties and clinical trial sites for guideline-changing studies of pharmacological approaches to glycaemic treatment. Clinical trial sites indicated by dots

4 DISCUSSION

The vast majority (>95%) of US clinical trial sites for guideline-changing studies of pharmacological approaches to glycaemic treatment were in urban areas. Only 1961 unique zip codes were used for study sites, which represents approximately 5% of all US zip codes.

Substantial geographic disparities in diabetes exist in the United States.4, 5 Diabetes prevalence is 17% higher in rural areas when compared to urban areas.4 This difference is thought to be a result of several factors.4, 5 Obesity is a major risk factor for diabetes and is more common in rural areas. Rural residents are more likely to report inadequate physical activity and barriers to consuming healthy food. Decreased access to healthcare services and health insurance, higher rates of poverty, and differences in education may also result in an increased burden of diabetes among rural residents. Given these differences between urban and rural residents and our findings that most diabetes clinical trials are conducted in urban areas, the applicability of clinical trial results to individuals living in rural areas is unknown.

Rural residents face several barriers to clinical trial participation. For instance, poorer healthcare infrastructure is a persistent barrier.6 If rural patients do not see healthcare providers as often as their urban counterparts due to decreased access, they may not be screened for clinical trial inclusion as frequently. Further, rural populations are smaller in size and wider in geographical spread, hence conducting large-scale clinical trials in these populations is challenging. While 97% of US land is considered rural, approximately 20% of the US population lives in rural areas.7 Nonetheless, our data suggest efforts are needed to increase clinical trial participation among the 20% of the US population living in rural areas, especially given that diabetes prevalence is approximately 17% higher in these areas.4, 5 Strategies used to conduct clinical trials during the COVID-19 pandemic could be mechanisms for enabling clinical trial participation for individuals living in rural areas.6 For example, remote clinical trials that use telehealth or other digital technologies for virtual visits and monitoring may help rural residents attend study visits. Moreover, use of teleconsent could make it easier for rural residents to enroll in clinical trials in the first place.8

Studies evaluating the impact of rural residence on clinical trial outcomes are sparse. We could not evaluate this in our study because we did not have access to patient-level data. A previous study evaluated survival outcomes for rural patients with cancer enrolled in clinical trials.9 The study included 36 995 US patients with 17 different cancer types enrolled in 44 clinical trials between 1986 and 2012. Each patient's location of residence was classified as rural or urban. No differences in outcomes were observed for rural versus urban residence for most cancer types after adjusting for covariates. However, individuals with estrogen and progesterone receptor-negative breast cancer living in rural areas had worse overall (hazard ratio 1.27, 95% confidence interval [CI] 1.06-1.51) and cancer-related survival (hazard ratio 1.26, 95% CI 1.04-1.52) when compared to individuals living in urban areas. Similar analyses should be conducted to compare outcomes among individuals living in rural versus urban areas enrolled in diabetes clinical trials.

The clinical interpretation of our findings relies on the assumption that the participants of the studies live in the vicinity of the trial sites. This is a limitation. Although uncommon, there may be involvement of rural patients in trial sites in urban areas.

In conclusion, very few (<5%) US clinical trial sites for guideline-changing studies of pharmacological approaches to glycaemic treatment were in rural areas. Given this finding and the known burden of diabetes among rural residents, efforts are likely needed to increase the generalizability of diabetes clinical trials.

CONFLICT OF INTEREST

The authors report no conflicts of interest germane to this manuscript.

Data were derived from the following resources available in the public domain: clinicaltrials.gov. The R script generated for this analysis is available from the corresponding author upon reasonable request.

留言 (0)

沒有登入
gif