Disparities in routine healthcare utilization disruptions during COVID-19 pandemic among veterans with type 2 diabetes

Emerging studies suggest that the COVID-19 pandemic caused significant disruption in routine healthcare, including reduced in-person primary care visits [1‐5]. Gaps in routine care can have a negative impact on health outcomes, especially among those with chronic diseases, such as type 2 diabetes (referred to as diabetes hereafter) [6]. While over 30 million Americans [7], including many veterans [8], are living with diabetes, changes in their care utilization as a result of the pandemic have not been well characterized, particularly with respect to variations by patient characteristics, disease severity, or geography [9]. Understanding the pandemic’s impact on healthcare utilization among patients with diabetes is needed to design strategies to reduce the negative impact that gaps in healthcare services may have on this large, vulnerable population.

While frequent routine care is crucial for improving diabetes control and reducing related complications risks [10, 11], there were challenges providing in-person diabetes management care during COVID-19 [12]. Benefits of routine in-person care include face-to-face patient-provider interactions, optimized setting for lab measurements, and lower technology burden on patients [13]. However when these barriers exist, telehealth may provide alternative solution to replicate in-person care [12, 13]. Many published trials and meta-analyses have shown improved glycemic control and increased frequency of care via telehealth visits among patients with diabetes [14‐16]. While telehealth can provide a stopgap solution to loss of in-person services during the pandemic, little is known about which patient subgroups used or benefited from this healthcare alternative.

In March 2020, in response to the initial wave of the pandemic, the Veterans Administration (VA) ordered an almost-complete, national shutdown of in-person outpatient care at VA hospitals and clinics. Shortly thereafter, the VA expanded telehealth care via telephone and video conferencing. While studies have begun to characterize the uptake of telehealth visits [2, 17], the extent to which these disruptions and changes impacted routine care among veterans with diabetes, including regular glycohemoglobin (HbA1C) testing and medication prescription fills, remains unknown.

To address this gap and to inform mitigation strategies, we examined longitudinal trends in routine healthcare utilization during the first year of the pandemic among veterans with diabetes from an existing large national cohort using VA electronic health records (EHR). Using an interrupted time series design with the pandemic onset as an external shock, we quantified the pandemic’s disruption on routine medical care among veterans with diabetes. To determine if baseline levels of glycemic control were associated with subsequent utilization patterns during the pandemic, we compared utilization trends among patients with different baseline glycemia levels. To assess the potential differential impacts of the pandemic on healthcare utilization by patient demographics or geographic context, we compared trends among racial/ethnic groups and across the urban-rural continuum.

Table 1 shows the demographic characteristics of 733,006 veterans with well-documented diabetes. The mean age as of March 2018 was 67 years, with more than 50% over the age of 60. Veterans were pre-dominantly male (94%) and non-hispanic White (69%) followed by non-hispanic Black (21%) and 6% Hispanic. Among these veterans, 40% were categorized as low-income (not disabled), and 38% were disabled. Among 648,213 veterans who had pre-pandemic glycemic control data, 282,666 (43%) had HbA1c > 9.0, 224,658 (35%) had HbA1c between 6.5 and 8.9, 78,762 (12%) had HbA1c between 5.7 and 6.49, and 50,075 (7.7%) had HbA1c < 5.7. In this sample, 17.8% veterans were from rural communities, 24.7% from suburban, and 38.1% from urban communities, with 19% missing community type. Only 55% of the veterans in the cohort filled prescriptions for diabetes and 76% filled prescriptions for hypertension from VA pharmacy during March 2018–2021. Veterans with prescription fills were similar in sex and low-income status compared to those without (Table S3). However, those with prescription fills were slightly younger (76% over 60 years vs 83%) and predominantly non-hispanic Black (22% vs 16%).

Since March 2020, there were sharp reductions in the monthly rates of HbA1C measurements and in-person outpatient visits, with the lowest rates in April 2020, Fig. 1. Telehealth visits increased starting March 2020, peaking in April 2020 and slowly declining throughout the first year of the pandemic as in-person visits returned. Most telehealth visits were via telephone rather than video. Among veterans who filled their prescriptions in VA pharmacies, we saw no major changes in fill rates after the pandemic onset for hypertension medication, although fill rates for diabetes medication showed a slight increase after July 2020. Trend plots (Figs. S1, S2 and S3) stratified by pre-pandemic glycemic control, race/ethnicity and community types showed similar patterns.

Rate ratios comparing monthly rates of utilization outcomes pre-pandemic (March 2018 through February 2020) and throughout the first year of the pandemic starting on March 2020 are reported in Table 2. We observed a 57% reduction in monthly rates of HbA1c measurements starting March 2020 compared to the pre-pandemic period. Reductions were greatest in the first 3 months, March–June 2020 with monthly predicted rate of 60 (95%CI,50–70) per 1000 veterans compared to 130 (95%CI,110–140) prior to March 2020. By June 2020, monthly rates of HbA1c measurements had largely returned to pre-pandemic rates (110;95%CI,90–130 per 1000 veterans). In the first year of the pandemic, in-person outpatient visits decreased by 56%, while telehealth visits increased by more than two-fold. Monthly rates of in-person visits decreased from 1830 visits (95%CI,1640–2040) to 810 (95%CI,710–930) per 1000 in March–June 2020 and to 1190 (95%CI,1040–1360) July–December 2020, before returning to pre-pandemic rates on January 2021. However, telehealth visits consistently increased throughout March 2020–2021 from 330 (95%CI,300–350) pre-pandemic to 770 (95%CI,720–820) visits per 1000. Overall combined in-person and telehealth outpatient visits decreased in the initial 3 months of the pandemic from 2130 (95%CI,2000–2280) to 1630 (95%CI,1520–1760) per 1000 and rebounded to pre-pandemic rates starting in July 2020. We observed similar trends when the visits were limited to primary care. Monthly rates of prescription fills for diabetes and for hypertension medications did not differ from pre-pandemic period. Plot of the autocorrelation functions (Fig. S6) suggested no evidence of residual autocorrelation, increasing our confidence in the model assumptions.

Stratified analyses (Fig. 2) showed differential impacts on telehealth visits by community type and race/ethnicity. Increase in monthly rates of telehealth visits starting March 2020 was lowest in rural community (RR = 2.06;95%CI,1.85–2.30) and highest in high density urban community (RR = 2.78;95%CI,2.50–3.09). Moreover, rates of telehealth visits increased most among non-hispanic Black veterans (RR = 2.66;95%CI,2.39–2.95) followed by Hispanics and non-Hispanic AAPI, and least among non-hispanic White (RR = 2.15;95%CI,1.94–2.38). However, we did not observe any difference in the associations by geography or race on HbA1c measurements, medication fills, and in-person visits. When stratified by pre-pandemic glycemic controls, relative rates of monthly HbA1c measurements were consistent across the strata. Similarly, in-person visits decreased and telehealth visits increased consistently for all levels of pre-pandemic glycemic control. There was also no evidence of variable changes in prescription fills.

In this large, national cohort of US veterans with diabetes diagnosed in the decade prior to the COVID-19 pandemic, we quantified the reduction in healthcare utilization after the onset of the pandemic compared to the pre-pandemic period. We observed fewer HbA1c measurements and fewer in-person outpatient visits in the first year of the pandemic period, with most pronounced reductions in March–June 2020, when nearly all elective procedures and routine care were reduced nationally due to public health protocols and pandemic demands on the healthcare system. Telehealth visits increased consistently starting March 2020, with important variation in uptake by different communities and racial groups. While there was substantial reduction in total number of outpatient visits March–June 2020, it returned to pre-pandemic levels starting July 2020. Trends in prescription fills for diabetes and hypertension medications, however, remained stable throughout the study period.

We observed disparate uptake of telehealth visits, with substantially lower uptake in rural communities compared to other community types, while urban communities had the largest increase. While we did not find disparate utilization for HbA1c measurements and prescription fills by race, we did see the highest increase in telehealth utilization among Black and lowest increase among White veterans. Our analyses further showed that utilization did not vary according to pre-pandemic glycemic levels, suggesting that the pandemic impacted veterans with diabetes across all levels of pre-pandemic control. Our findings are consistent with pior research that investigated utilizations in chronic diseases after external shock from a natural disaster [6, 30]. Since the onset of the pandemic, the overall trends of reduced in-person visits and increased telehealth visits among veterans who received medical care from VA health centers [2, 17] and among older veterans with diabetes [31] have been reported. However, our study is among the first to investigate variations in the impact of the pandemic on veterans with diabetes nationwide.

The trends in reduced utilization, including HbA1c testing and in-person visits, among veterans with diabetes are concerning as they highlight the missed opportunities to manage the chronic condition, which can result in worse outcomes including poor diabetes control [32‐34]. While increased telehealth visits partially compensated for the missed in-person visits by July 2020, it is unclear whether replacement of in-person visits with telehealth helps in disease management, such as routine HbA1c measurement, among patients with diabetes. Furthermore, substantial geographic and racial variation in telehealth uptake is likely a premonition for future disparities in long-term clinical complications among this patient population.

While Black and Hispanic racial groups in urban communities have disproportionately experienced the burden of the pandemic [35], there is a mixed evidence indicating disparities on healthcare utilization through the pandemic [36, 37]. Our finding on lower telehealth utilization among rural residents is consistent with observed trends among patients with chronic diseases [38]. Limited broadband availability and low preparedness prior to the pandemic are potential key structural barriers to telehealth utilization among rural communities [39]. However, the increased telehealth utilization by Black veterans, who primarily resided in urban areas, is contrary to previous findings [40]. Further research is needed to understand whether telehealth could be helpful in mitigating health disparities.

While the pandemic disrupted routine healthcare utilization among veterans with diabetes, their access to medications were unaffected. Further, while in-person visits decreased since the pandemic onset, telehealth visits have increased. The rise in telehealth visits along with mail-order pharmacy could have played a role in ensuring that the medication orders were filled in timely manner with none to minimal disruption on the prescription fills.

Healthcare utilization among veterans with diabetes was disrupted in the initial months of the COVID-19 pandemic before rebounding to pre-pandemic levels. There were large reductions in monthly rates of HbA1c measurements and in-person outpatient visits, whereas telehealth visits nearly doubled with substantial disparities by community type and race/ethnicity.