Diabetes and prediabetes affect 13 % and 34.5 % of U.S. adults, respectively [1]. Testing adults’ blood glucose levels to screen for prediabetes or diabetes is important for early detection and subsequent management. Regular screening of asymptomatic individuals is vital to achieve the Healthy People 2030 objectives: reduce the number of diabetes cases diagnosed yearly — D-01; and reduce the proportion of adults who don’t know they have prediabetes — D-02 [2]. The U.S. Preventive Services Task Force (USPSTF) guideline in 2015 recommended screening for prediabetes and type 2 diabetes in adults aged 40–70 years who are overweight or obese [3]. In 2021, the USPSTF updated the recommendation by lowering that age criteria for screening from 40 to 35 years for earlier detection [4].

The COVID-19 pandemic has negatively impacted preventive health services, and identification and management of chronic diseases [5]. Previous research showed that people delayed or avoided routine medical care during the pandemic [6]. Reports from Europe and the United Kingdom (U.K.) showed the COVID-19 pandemic adversely affected testing for glycated hemoglobin in people with or at risk for diabetes, delaying delivery of diabetes care [7,8]. One U.S. report showed that preventive health screenings for cardiometabolic diseases—including diabetes— and cancers declined in 2021 relative to 2019, with variation by educational attainment and race and ethnicity [9].

However, among adults without diagnosed diabetes, no study has assessed how changes in blood glucose testing differed by demographic factors and geographic locations. We hypothesized that disparities in demographic characteristics and geographic locations exist in blood glucose testing among adults without diagnosed diabetes. Capturing changes in preventive screening uptake is important, as its consequences may extend long after the pandemic, and it can inform future pandemic readiness. Therefore, we used the National Health Interview Survey (NHIS), 2019–2021, to investigate the prevalence of blood glucose testing among U.S. adults without diagnosed diabetes—overall and by selected sociodemographic characteristics.

The study sample included 82,594 respondents without diabetes in 2019–2021, with a mean age of 46.4–46.8 years (48.4 %–48.7 % were men). Population characteristics by survey year are shown in Table 1. Most demographic features showed similar distributions over the 3 years. However, compared with adults in 2019, those in 2021 were on average older, more likely to report diagnosed prediabetes, living in large central metro areas, having health insurance, reporting higher income, and having a college degree or greater.

Overall, the prevalence of adults without diagnosed diabetes who received blood glucose testing within 12 months was 64.2 % (95 % CI 63.3 %, 65.1 %) in 2019 and 60.0 % (95 % CI 59.1 %, 60.9 %) in 2021. These findings are equivalent to an absolute change in percentage points of −4.2 (95 % CI −5.3, −3.1) and a relative change of −6.6 % (95 % CI, −8.2 %,−4.9 %). Among the adults who fulfilled the USPSTF 2015 recommendation for screening of prediabetes and type 2 diabetes, the absolute and relative change in prevalence from 2019 to 2021 was −3.9 (95 % CI, −5.6, −2.3) and –5.4 % (95 % CI, −7.5 %, −3.2 %), respectively. Fig. 1 shows the age-adjusted prevalence of adults who received testing for blood glucose within 12 months by year, overall and for those meeting USPSTF 2015 recommendation.

Fig. 2 shows age-adjusted absolute and relative change in testing prevalence from 2019 to 2021, and Supplemental Table 1 also shows crude prevalence of adults who received blood glucose testing for the 3 years by selected characteristics. Significantly lower testing was observed for almost every demographic and geographic subgroup except non-Hispanic (NH) American Indian and Alaska Native adults, other single or multiple race adults, those with underweight, and those with family income < 100 % of FPL.

In 2019, prevalence of blood glucose testing within 12 months was lowest among adults without health insurance and in the youngest age group. The highest testing prevalence both in 2019 and 2021 was observed among adults aged 40–64 years and ≥ 65 years, NH Black adults, and those with self-reported prediabetes.

The largest age-adjusted decline in testing prevalence, in both absolute and relative terms, was observed in NH Asian adults (−10.0 and − 14.8 %, respectively), the West region (−7.2 and − 12.0 %, respectively), large central metro areas (−6.9 and − 10.5 %, respectively), those without health insurance (−6.3 and − 17.3 %, respectively), and the youngest age group (−5.0, − 10.3 %, respectively). Notably, among racial and ethnic groups, Hispanic adults experienced the second-largest decline in testing, from 60.7 % in 2019 to 54.7 % in 2021 (−6.1 % absolute and − 10.0 % relative change).

More frequent testing in 2021 than 2019 was observed among NH American Indian and Alaska Native adults, other single or multiple race adults, and those with BMI < 18.5 kg/m². However, these changes were not statistically significant, and the 51.9 %–61.4 % testing prevalence attained in 2021 remained at the lower end of the overall prevalence distribution.

Supplemental Table 2 showed multivariable-adjusted change in prevalence between 2019 and 2021 where similar trends to age-adjusted change were observed.

In a nationally representative sample of US adults without diabetes, the prevalence of those who received blood glucose testing within 12 months was significantly lower in 2021 than in 2019, suggesting that screening for prediabetes or type 2 diabetes was negatively impacted during the COVID-19 pandemic. Lower testing prevalence from 2019 to 2021 was observed particularly in young adults, non-Hispanic Asian and Hispanic populations, those without health insurance, adults living in central metropolitan areas and adults living in the West. Although testing was consistently higher among adults meeting the 2015 USPSTF criteria, it declined over time in a manner similar to that of the general population without diagnosed diabetes.

Reduced use of preventive health services such as screening during the pandemic was consistently observed in our study and in reports from various countries, including the US [9,18], the UK [7,19], Canada [20], and Israel [21]. Because 31.5 % of US adults refrained from routine care during the pandemic due to reduced accessibility, decreased availability of public transportation or fear of exposure to COVID-19 [6, 22], it is not surprising that people delayed or avoided blood glucose testing as well.

The low prevalence of blood glucose testing in the youngest population (aged 18–39 years) before the COVID-19 pandemic and that group’s further 10 % reduction during the pandemic reflects the generally low prevalence of preventive care in young adults [23] and a decreased usage of preventive care services [24]. This finding is concerning for at least two reasons. First, approximately 32 million people aged 18–44 years have prediabetes [25], and they are less likely to be aware of their prediabetes status than older adults [26]. The observed delay in blood glucose testing might lead to more undetected prediabetes cases and a rise in type 2 diabetes incidence after 2021, as suggested by other population-based studies [27]. Secondly, national data indicate a resurgence in the incidence of diabetes complications among young and middle-aged adults after 2015 [28], suggesting that early detection and treatment of diabetes is critical in this population [29]. It remains to be seen if lowering the age for prediabetes and diabetes screening from 40 to 35 years, as recommended by the USPSTF [4], will encourage higher rates of blood glucose testing to screen for prediabetes or type 2 diabetes in this segment of the population.

The racial and ethnic heterogeneity in testing for blood glucose persisted during the COVID-19 pandemic, with the highest testing prevalence among NH Black adults and the lowest in Hispanic adults and those of other single or multiple races. The largest attrition in testing frequency occurring among NH Asian adults is consistent with one previous report which included only adults aged 40–75 [9]. NH Asian adults had relatively high testing prevalence in 2019; also they were on average less likely to be in the lower income bracket, or to lack health care access or health insurance than other non-White racial and ethnic groups. A combination of potential factors may have prevailed in this population, including its larger concentration in the West—where testing was consistently lower than in other regions—experiences of discrimination [30], language barriers [31], and possibly stricter adoption of isolation practices for COVID-19 prevention [32].

The increasing trend in testing uptake from 2019 to 2021 was unexpected in some groups such as NH American Indian and Alaska Native adults although the differences were not statistically significant, possibly due to the low sample numbers. Use of remote technology in the Indian Health Service might partly explain this population’s increase in health care visits and subsequent testing uptake [33]. One study from a single integrated academic health system in the Upper Midwest showed that American Indian and Alaska Native adults were more likely to use a video or audio visit vs. in-person visits, compared with other race and ethnicity groups in 2020, during the early months of the COVID-19 pandemic [34]. It may be worth considering further research about the effects of telemedicine on testing during the pandemic.

Change in testing is linked to locality. People living in large central metro areas may have been more affected by the pandemic, possibly due to enforcement of quarantine measures. In addition, the higher proportion of younger people in urban areas may partly explain the lower testing rates [35]. Consistently lower testing rates in rural areas may be explained by disparities in health care access linked to financial constraints [36,37], in addition to demographics (e.g., rural residents were more likely to be NH White and older).

We acknowledge several limitations. NHIS collected self-reported data, which are subject to recall biases. There was also a change to phone interviews during 2020 due to the pandemic, with a lower response rate that may reflect selection bias. Due to a major questionnaire redesign in 2019, we could not assess trends in blood glucose testing to compare frequency before and after 2019.

Testing for blood glucose decreased in the US during the COVID-19 pandemic, with differences by socio-demographic, clinical, and geographic characteristics. Delayed blood glucose testing could result in a higher burden of diabetes management in the post-pandemic era. The study findings reveal the need to encourage diabetes screenings post-pandemic and subsequently, as well as awareness and management of diabetes—especially among people who are at increased risk for diabetes. In addition, our findings may be helpful in preparing for a potential pandemic in the future. Routine blood glucose testing among adults without diagnosed diabetes may be an important first step in the cascade of appropriate care for people with diabetes and developing resilient screening systems is warranted.