In 2010, 26% of US adults lived with 2 or more chronic conditions (1), a number expected to increase as the prevalence of chronic conditions increases (2). The most common pairing of multiple chronic conditions across ages is diabetes and hypertension (1). In Minnesota, the adult diabetes prevalence nearly doubled in the past 15 years to more than 7%, and 26% of adults report receiving a hypertension diagnosis (3,4).

Recent evidence-based guidelines for prevention of diabetes and cardiovascular risk emphasize expanded delivery of preventive clinical services, which may improve referral to diabetes prevention programs and early clinical management for adults with diabetes (5,6). Blood pressure control among adults with diabetes significantly reduces cardiovascular events (7,8). On the basis of blood pressure data, the US Preventive Services Task Force (USPSTF) rated blood glucose screening for people with diagnosed hypertension a Grade B recommendation, making it a covered preventive service under the Affordable Care Act (9).

Despite the value of diabetes screening in patients with hypertension, estimates of the implementation of USPSTF screening guidelines are unavailable. National Health and Nutrition Examination Survey (NHANES) data from 2005–2006 indicate that 53.2% of adults with hypertension have been screened (10); recent analyses were limited to attributes of different screening criteria, including sensitivity, specificity, and absolute numbers of cases detected (11,12). Assessment of screening rates and patient-level characteristics associated with screening status are needed.

This analysis describes blood glucose screening rates and patient-level characteristics among Minnesota adults with hypertension. The results provide baseline information for developing public health and clinical strategies to identify patients with hypertension who are at risk for diabetes.

To assess blood glucose screening rates, we obtained data for Minnesota from the 2011 Behavioral Risk Factor Surveillance System (BRFSS) (13), a state-based survey of health-related behaviors among noninstitutionalized US adults aged 18 years or older and administered by states for the Centers for Disease Control and Prevention (CDC). Standardized clinical measures assessing diabetes management do exist, but standardized clinical measures for assessing diabetes screening rates do not exist. BRFSS provides a unique opportunity to assess statewide trends.

We limited analysis to 2011 data because differences in sampling and weighting initiated for the 2011 survey make results noncomparable to those of previous years. In Minnesota, the percentage of people with complete or partially complete interviews (ie, cooperation rate), was 82.2%, and the completion rate was 51.9%, higher than the national medians of 73.8% (cooperation rate) and 49.7% (completion rate) (14). We identified a cross-sectional sample of adults (N = 3,847) with self-reported hypertension, available covariate data, self-reported diabetes or prediabetes status, and blood glucose screening in the previous 3 years.

We identified all people with hypertension and then determined subsets of people who would closely meet the definition of asymptomatic for diabetes used in the USPSTF recommendation (Figure). All responses of “don’t know/not sure” or “refused” were treated as missing.

Weighted frequencies were used to describe Minnesota adults with diagnosed hypertension (Table 1). Most were 45 years or older (81%) and non-Hispanic white (90%); there were slightly more men than women in this group. Nearly 80% were classified as either overweight or obese. Approximately 22% were college graduates. More than 90% had insurance coverage and had a check-up in the past 2 years. Less than 1 in 7 reported coronary heart disease, and 20% reported a diagnosis of diabetes. Characteristics were not appreciably different for the subset of individuals with hypertension who had no diabetes diagnosis (Table 1), and 67.6% of these adults reported blood glucose testing within the past 3 years.

Other descriptive analysis focused on 4 groups of people with hypertension: those diagnosed with diabetes (19.6%), those diagnosed with prediabetes (10.7%), those diagnosed with neither diabetes nor prediabetes and screened in the past 3 years (no diagnosis, screened, 44.6%), and those diagnosed with neither diabetes nor prediabetes but not screened in the past 3 years (no diagnosis, unscreened, 25.1%) (Table 2). We examined the proportion of adults in each composite diagnosis-screening group stratified by diabetes risk factors and access to the health care system (Table 2). The percentage of adults with diabetes tended to be higher among adults who were obese, had lower educational attainment, were of a race/ethnicity other than non-Hispanic white, and not current smokers. Compared with their counterparts, the percentage of adults who had no diagnosis and were unscreened was greater among adults who were younger, had lower BMI, were current smokers, were not enrolled in a health plan, and did not have a recent check-up. We also conducted analyses excluding the 14.6% of adults who had coronary heart disease; associations were unchanged.

To describe associations between demographic, lifestyle, and health care access and use variables and screening among respondents at risk of developing diabetes, we calculated the odds of not being tested within the past 3 years (Table 3). Adults aged 18 to 44 were 2.64 (95% CI, 1.89–3.67) times as likely to not be screened compared with adults aged 65 years or older; there was no meaningful difference for adults aged 45 to 64 versus those aged 65 years or older. Odds of not being screened were elevated for adults with low/normal BMI compared with those who were obese, for those without a college degree compared with those with a college degree, and for current smokers compared with never smokers. Adults not taking medication to manage their hypertension were less likely to be screened (OR, 2.94; 95% CI, 2.25–3.85) than those who did take medication. Nonenrollment in a health plan (OR, 1.97; 95% CI, 1.22–3.17) and lack of a check-up in the past 2 years (OR, 3.70; 95% CI, 2.28–6.02) were associated with not being screened. In multivariable models, most associations remained, with similar magnitude of effect. Smoking was no longer associated with screening, nor was health plan enrollment. Removing covariates that were significant at P < .05 did not appreciably change the overall results or point estimates of associations between covariates and no screening. None of the 3 sensitivity analyses changed results appreciably. For example, in data limited to non-Hispanic whites only, associations between respondents aged 18 to 44 years (OR, 1.90; 95% CI, 1.28–2.81) and not having a check-up (OR, 3.08; 95% CI, 1.93–4.91) were strengthened slightly. If limited to respondents without a prediabetes diagnosis, the association with not having a check-up was strengthened (OR, 3.10; 95% CI, 1.83–5.24).

Nearly one-third of Minnesota adults who have hypertension without a diabetes diagnosis reported not receiving recommended blood glucose screening in the past 3 years. Adults with hypertension are a group at high risk for developing diabetes; 19.6% and 10.7% of Minnesota adults with hypertension have diagnoses of diabetes and prediabetes, respectively. Because 26% (or approximately 1.1 million) of Minnesota adults report having a hypertension diagnosis (3), approximately 345,000 Minnesota adults with hypertension would report not having recommended blood glucose screening. This is the first state or national assessment of USPSTF diabetes screening rates that also identifies subpopulations less frequently screened. Other state-based analyses of blood glucose screening using BRFSS data described rates in the general adult population or a subset of obese adults that do not align with national guidelines (18–20). Standardized clinical measures for diabetes screening for large population subgroups do not exist and use of All Payer Claims Databases is in its infancy (21). BRFSS provides a unique opportunity to assess statewide trends and monitor the progress of programmatic strategies to improve diabetes screening and hypertension management.

Use of self-reported data adds complexity to data interpretation but does not make the effort less meaningful. Failure to report screening in the past 3 years could reflect not having been screened, lack of awareness that screening was needed or occurred, or both. Lack of awareness could be a function of low health literacy (22,23) and could be used to guide programmatic activity and intervention development.

Consistent with the concept of variable health literacy (22) among adults with hypertension is the association between education and blood glucose screening. Compared with adults with a college degree, adults with less than a college degree were approximately 40% more likely not to have blood glucose testing, as has been observed for people with CHD (16). Higher educational attainment has been associated with greater agreement between self-report of health conditions and the medical record (24).

Health systems factors may influence self-reported screening rates. Not receiving a check-up was associated with lower self-reported screening. Lack of health insurance showed the same relationship in crude models but was not significant after adjustment, likely because of not having had a check-up. In Minnesota, more than 80% of adults with diagnosed hypertension reported visiting a provider in the last year, whereas the US average is 70% (12). Adults who did not use medication to manage their hypertension were less likely to be screened, consistent with results showing that adults who take medication for hypertension or hypercholesterolemia were more aware of their prediabetes status (25). Additional visits for medication management may be required, providing more opportunities for screening; conversely, enhanced screening may reflect clinical assessment of this subgroup being at higher risk.

Variation in screening rates by BMI and age suggest that provider or health systems strategies may also be warranted. After adjustment for other factors, adults younger than 45 years and those with low/normal and overweight BMI were less likely to report screening, as reported previously for adults with CHD (16). The results suggest that providers do not screen all people with hypertension with equal probability, and providers may screen according to guidelines that take age and BMI into account (ie, guidelines from the American Diabetes Association [ADA] or National Institute of Diabetes and Digestive and Kidney Diseases [NIDDK]) (17,26). The results also suggest that awareness of diabetes risk and the need for screening is lower among these groups (ie, adults younger than 45 years and adults with lower BMI) and they are less likely to recall provider-ordered testing. Screening rates among Minnesota adults aged 45 years or older show BMI-related patterns consistent with NIDDK guidelines; in 2011, 63% of overweight or obese adults and 52.1% with low/normal BMI reported screening. NIDDK recommends screening for overweight or obese adults aged 45 years or older and that testing be considered for normal-weight adults of the same age (26). ADA recommends screening all adults aged 45 years or older (17). Given these multiple guidelines, there may be lack of clarity regarding populations eligible for screening.

Strategies addressing barriers related to providers, health systems, and health literacy may improve blood glucose screening rates among adults with hypertension in Minnesota. To address provider awareness and screening, several approaches can be considered: 1) analyzing screening rates using electronic health record data or assessing All-Payer Claims Databases’ provider-ordered screening rates (21), 2) emphasizing that screening adults with hypertension is largely consistent with other screening guidelines, and 3) using clinical decision tools or including blood glucose screening in hypertension management plans. Recent policy changes also may improve access to care. The Affordable Care Act expands the number of people insured and provides an annual preventive visit that could address blood glucose screening at no cost to the patient. Efforts to address health literacy could include awareness campaigns that emphasize the need for blood glucose screening for all adults with hypertension and stress that preventive visits, now a covered benefit under the Affordable Care Act, are opportunities for recommended screening.

Although our analysis is meant to inform programmatic efforts in Minnesota, our findings may be useful more broadly. National-level analysis of diabetes screening rates among adults with CHD found similar rates and screening patterns. Blood glucose screening rates from the 16 states with available 2011 BRFSS blood glucose screening data for all adults with hypertension ranged from 61.9% to 74.1% . Minnesota rates fall midrange, failing to provide evidence that our rates would be highly divergent from results in other states. Univariate analysis of blood glucose screening rates for all adults in Montana (BRFSS 2009) and New York (BRFSS 2008–2009) demonstrated lower screening rates with lower BMI, lower age, and lower educational attainment, consistent with our findings (18,19).

Our study has 5 key limitations. First, we were unable to identify asymptomatic adults with hypertension, the population to which the USPSTF recommendation applies (9), because BRFSS lacks questions about signs and symptoms of diabetes among those without a diagnosis. Second, self-reported measures like diagnosed hypertension are a combination of testing and diagnosis rates and awareness of the diagnosis. Third, the reliability of hypertension and diabetes diagnoses and blood glucose screening is not well-defined. Earlier work suggests high sensitivity for hypertension, moderate sensitivity for diabetes, and high specificity for both (27) (κ, range, 0.7–0.8, for both) (24). Reliability is important from a clinical perspective but becomes less so when addressing issues of patient awareness and health literacy. Fourth, small numbers limited our ability to examine patterns of screening among adults of nonwhite, non-Hispanic race/ethnicity. Finally, we analyzed only 1 year of data. It is unlikely that using a single year of data misrepresents larger trends, given the similarity between 2011 BRFSS estimates in Minnesota and other states. Also, the 19.6% prevalence rate for diabetes among Minnesota adults with hypertension is similar to the rate of 18% to 18.5% reported in 2005–2010 NHANES (28).

Nearly one-third of Minnesota adults with hypertension without diabetes did not report having had blood glucose testing in the past 3 years, as recommended by the USPSTF. Failure to screen and failure to understand the importance of screening may mean missed opportunities for 1) early detection and good clinical management of diabetes and 2) identification of adults with prediabetes and referral to evidence-based interventions that delay or prevent the onset of diabetes (5,6,29) and that may confer additional cardiovascular benefit (30–32). Efforts to improve screening rates should consider a multifaceted approach, addressing provider-, health systems–, and health literacy-related barriers suggested by surveillance data.