Question 
Is integrating rapid diabetes screening into a community-based COVID-19 testing site feasible and effective in reaching socioeconomically disadvantaged Latinx persons without health care?

Findings 
In this health care improvement study of 6631 participants presenting for COVID-19 testing, 923 (13.9%) underwent hemoglobin A_1c testing, of which 313 (34%) and 113 (12%) had prediabetes and diabetes, respectively. Persons reached were mostly Latinx (83%), low-income, had not previously been tested for diabetes, and were not engaged in formal health care services.

Meaning 
These results suggest that leveraging community COVID-19 infrastructure and partnerships has the potential to address disparities in diabetes.

Importance 
Community-based COVID-19 testing and vaccination programs play a crucial role in mitigating racial and ethnic disparities in COVID-19 service delivery. They also represent a platform that can be leveraged to expand access to testing for chronic diseases, including diabetes, that disproportionately affect the Latinx community and other marginalized communities.

Objective 
To evaluate outcomes associated with a diabetes testing strategy designed to reach low-income Latinx persons by leveraging COVID-19 testing infrastructure and community trust developed during the COVID-19 pandemic.

Design, Setting, and Participants 
This health care improvement study was conducted from August 1 to October 5, 2021, at an outdoor, community-based COVID-19 testing site at a transport hub in the Mission Neighborhood in San Francisco, California. Because the program was designed to expand access to diabetes screening to the local community, all individuals presenting for on-site testing were eligible. Data were analyzed in November 2021.

Interventions 
Integration of rapid, point-of-care hemoglobin A_1c screening as a testing option in an existing low-barrier COVID-19 testing program.

Main Outcomes and Measures 
Evaluation was guided by the Reach, Effectiveness, Adoption, Implementation and Maintenance (RE-AIM) framework and utilized programmatic data and structured surveys among clients and staff.

Results 
Of 6631 individuals tested (median [IQR] age 39.3 [29.7-51.3] years; 3417 [52.3%] female, 4348 [65.6%] Latinx), 923 (13.9%) underwent hemoglobin A_1c testing with or without COVID-19 testing and 5708 (86.1%) underwent COVID-19 testing only. Individuals tested for diabetes were more likely to be Latinx (763 of 923 individuals [82.7%] who underwent testing were Latinx vs 3585 of 5708 [62.8%] not undergoing testing), have an annual household income of less than $50 000 (450 individuals [81.2%] vs 2409 individuals [66.0%]), and not have health insurance (381 individuals [47.2%] vs 1858 individuals [39.9%]), and 206 (48.0%) had never tested for diabetes before. Overall, 313 (33.9%) and 113 (12.2%) individuals had prediabetes and diabetes, respectively; only 141 of 354 of these individuals (39.8%) had a primary care clinician whom they had seen in the prior 12 months, which was lower among Latinx individuals (113 of 307 individuals [36.8%] vs 28 of 47 [59.6%]). Acceptability of the rapid testing program was high—98% were satisfied with their visit and 96% said they would return for future services; key factors underpinning acceptability included friendly staff, efficiency, and a convenient location.

Conclusions and Relevance 
In this health care improvement study conducted within an existing community-based COVID-19 testing program, integrating rapid testing for diabetes was feasible, reached low-income Latinx individuals, and identified many persons with prediabetes and diabetes, most of whom lacked access to services in formal health care settings. Leveraging pandemic-related public health responses represents an important opportunity for engaging socioeconomically disadvantaged populations into care for diabetes.

Beginning August 1, 2021, we began offering rapid hemoglobin A_1c and rapid HIV antibody testing at the Unidos en Salud COVID-19 testing program as part of a health care improvement study; this directly coincided with the arrival of the COVID-19 Delta variant in San Francisco. At the time of registration, individuals were asked whether they wanted to undertake COVID-19 testing, HIV testing, and/or hemoglobin A_1c testing. For those wanting COVID-19 testing, laboratory assistants performed a bilateral anterior nasal swab, and their status was determined using a rapid COVID-19 antigen test (Abbott Diagnostics). Fingerstick whole blood was collected by laboratory assistants for individuals who agreed to hemoglobin A_1c or HIV antibody testing. Hemoglobin A_1c levels were determined using a A_1C whole blood test (PTS Diagnostics). All diagnostic tests were performed at the point-of-care according to manufacturer instructions. Clients were provided results disclosure for all tests within 1 hour of testing via secure messaging; for those found to have COVID-19, prediabetes or diabetes, or HIV, a Spanish-speaking bilingual community health worker (CHW) disclosed the results of the test, provided brief health education, and facilitated linkage to primary care. Health education included information about prediabetes or diabetes pathophysiology, brief counseling on lifestyle modifications, and counseling for all patients on the need for confirmatory testing in a primary care clinic. Clients with diabetes who were uninsured were directly referred to a community health clinic for a new patient appointment to establish primary care; clients with diabetes who were insured and had an established primary care clinic but were out of care were provided coaching from a CHW to reengage in care. Uninsured clients with prediabetes were provided coaching on how to call and request primary care through the San Francisco Health Network. All clients with prediabetes and diabetes who had an established primary care clinician and were engaged in care were encouraged to schedule a follow-up visit to discuss their results.

We used descriptive statistics to characterize participants. As our diabetes screening strategy and the broader community-based testing program was designed and tailored for Latinx community members, all outcomes, aside from representativeness of clients reached with hemoglobin A_1c testing, were undertaken overall and according to clients’ ethnicity (Latinx [self-identifying Latinx/Hispanic and/or American Indian from Central or South America] vs non-Latinx) to evaluate whether the strategy reached Latinx individuals. Wilcoxon rank-sum, Fisher exact, and χ² tests were applied as appropriate to compare characteristics. A cascade-of-care analysis was undertaken to visualize care engagement status among those with prediabetes and diabetes, respectively. All analyses were undertaken using Stata version 17.0 (Stata Corp). All statistical tests were 2-sided at α = .05.

Among all persons who underwent rapid hemoglobin A_1c testing, 497 (53.9%) had normal hemoglobin A_1c levels, while 313 (33.9%) and 113 (12.2%) individuals met criteria for prediabetes and diabetes, respectively. There was a higher prevalence of prediabetes (34.6% vs 30.6%) and diabetes (12.7% vs 10.0%) among Latinx clients compared with non-Latinx clients; Latinx clients accounted for 84.4% of prediabetes and 85.8% of diabetes diagnoses, respectively. Of clients completing the survey, 41 of 149 (27.5%) and 22 of 55 (40.0%) of those with prediabetes and diabetes, respectively, stated that they had been previously tested for diabetes, while only 15 of 54 (27.9%) of those with diabetes had a prior diabetes diagnosis.

Of 6591 clients who underwent rapid COVID-19 antigen testing, 385 (5.8%) were positive; persons who underwent hemoglobin A_1c testing were slightly more likely to test COVID-19 positive compared with those who declined hemoglobin A_1c testing (6.0% vs 5.8%; P < .001). Among clients with both rapid A_1c and COVID-19 results (883 individuals), the prevalence of COVID-19 did not differ according to A_1c classification: 32 (6.7%) in those with a normal A_1c, 15 (5.0%) in those with prediabetes, and 6 (5.5%) in those with diabetes (P = .82). Of 598 clients who underwent rapid HIV testing, 2 (0.3%) were positive; however, upon disclosure it was confirmed that both had a prior HIV diagnosis.

In this study, we found that a strategy to make point-of-care diabetes testing available to low-income Latinx community members seeking COVID-19 testing at a community site in the Mission District in San Francisco was feasible to implement, highly acceptable to individuals tested, and reached our priority population of low-income Latinx persons not engaged in formal health care services. Among the more than 900 persons who underwent rapid hemoglobin A_1c testing during the 2-month evaluation period, 83% were Latinx, of whom more than 80% reported a household income of less than $50 000 and 54% that they were not engaged in care, despite the availability of a health coverage program in which all San Francisco residents are eligible regardless of documentation status. The program effectively identified a high proportion of people with prediabetes and diabetes—over a third of testers met criteria for prediabetes and 12% for diabetes. Notably, most of these community members had never been tested for diabetes before. This testing program was paired with education and linkage to care by CHWs, and 83% of participants received diabetes-related counseling from a CHW; in a subsample, 71% scheduled primary care follow-up. These data highlight the many unmet needs of low-income Latinx individuals and the critical need to identify innovative, effective, and sustainable strategies to surmount any preexisting access-related barriers to primary care for socioeconomically disadvantaged populations.

Our community-based COVID-19 and diabetes testing program represents a novel care engagement model beyond traditional brick-and-mortar health care services that could increase access to care for diabetes and other chronic diseases among socioeconomically disadvantaged populations. There were several design features crucial to our low-barrier testing approach that helped us successfully reach low-income Latinx participants (many of whom were first generation immigrants) and drove high levels of acceptability and satisfaction among the clients we served. This included a highly visible outdoor, neighborhood location that was conveniently located at a busy transportation and Latinx commercial and cultural hub and that was open on weekend days. Also, community members could drop in for free testing without a prior appointment or need to show identification at a site staffed by monolingual and bilingual Spanish speaking staff that were friendly and trained in good customer service principles. Furthermore, the site provided an efficient process using rapid point-of-care diagnostics with results available for COVID-19 and diabetes tests within an hour. While clients cited several reasons for getting tested for diabetes, the majority said that they came for COVID-19 testing, but that they might as well get tested for diabetes since they were already there. This suggests that an important additional feature driving the successful uptake of diabetes testing in this context was the act of colocating diabetes testing among other services for which there was significant community demand (eg, COVID-19 testing and vaccination).

This study had several strengths, including that the multidisease testing strategy was developed in collaboration with community and public health partners and was tailored to specifically meet the needs of the local community. There were also some limitations to this study. Our multidisease testing strategy was conducted at a single site and leveraged community trust that we developed with the local Latinx community over a more than 15-month period during the pandemic; therefore, some of our findings may not be generalizable to other settings. Nonetheless, our site staff strongly felt that it was feasible to integrate point-of-care diabetes testing into an existing community COVID-19 testing site, and many communities have developed tailored COVID-19 testing and vaccination approaches to reach disadvantaged populations that could potentially be leveraged to implement a similar approach in other settings. Furthermore, despite reaching many low-income Latinx clients without linkage to primary care with hemoglobin A_1c testing, a majority of clients declined testing; reasons for declining hemoglobin A_1c testing were not directly elicited, but may reflect several factors, including knowledge of diabetes status through prior testing, perceived lack of time, and feeling too unwell to pursue additional testing, among others. Finally, linkage ascertainment was undertaken among a random subsample of clients. While these data are not fully generalizable because of their small sample size, this random sample provided insights on linkage. Although 71% had a follow-up appointment, the median follow-up time was 6 weeks; furthermore, the majority of uninsured clients did not have a follow-up primary care appointment at the time of the follow-up call. Additional research on the optimal strategies to overcome the multilevel barriers to care linkage following new diabetes diagnoses in community-based testing programs is urgently needed.

In this health care improvement study of implementing rapid testing hemoglobin A_1c testing into an existing community-based COVID-19 testing program, we found that integrated, point-of-care testing reached and was acceptable to predominantly socioeconomically disadvantaged Latinx community members, was feasible to deliver, and resulted in the detection of a large number of community members with either prediabetes or diabetes—most of whom had never been tested for diabetes and lacked access to formal health care services. Leveraging public health responses to COVID-19 and future pandemics represents a highly promising but largely underexplored opportunity to improve health equity by engaging socioeconomically disadvantaged populations into care for diabetes, and this model has potential to expand testing and linkage for other highly prevalent chronic diseases.

Accepted for Publication: March 20, 2022.