Brief Report: Access to Care and Cardiovascular Health in US Counties with Low vs. Higher Broadband Internet Availability

INTRODUCTION

More than one in seven US adults with cardiovascular disease face barriers accessing clinical care.¹ The rapid adoption of telemedicine, which has been accelerated by the COVID-19 pandemic, has the potential to address these gaps in health care access.² Unfortunately, many areas of the US lack broadband internet access, raising concern that the selective proliferation of telemedicine in areas with greater broadband availability could widen the “digital divide.” As part of the 2021 Bipartisan Infrastructure Law, the Biden Administration recently invested $65 billion into broadband expansion. Despite this major federal investment, little is known about how health care access and cardiovascular health differ between communities with low vs. higher-broadband access.

Therefore, in this study, we aimed to answer 3 questions. First, how does clinician (cardiologist, primary care physician) supply compare in US counties with low broadband availability compared to those with higher broadband availability? Second, does the burden of cardiovascular risk factors and disease differ between these groups? And third, how do cardiovascular and all-cause mortality compare between low and higher-broadband counties?

METHODS

We identified all 3,143 US counties and county-equivalents in 2019. The 5-digit Federal Information Processing Standard (FIPS) codes for each county were used to create linkages between several datasets capturing information about broadband availability, demographics, health care resources, and cardiovascular health.

Federal Communications Commission (FCC) broadband availability data was obtained from the United States Broadband Usage Percentages Dataset, a publicly available data set created by Microsoft Corp. based on broadband availability data from the FCC Fourteenth Broadband Deployment Report. County-level demographic characteristics were obtained from the American Community Survey. Clinician and hospital bed supply data were obtained from the Area Health Resources Files. Cardiovascular disease risk factors and cardiovascular disease prevalence data was obtained from the PLACES: Local Data for Better Health, County Data 2021 release, which contains model-based county level estimates based on self-reported data from Behavioral Risk Factor Surveillance System. The PLACES multilevel regression and poststratification modeling approach has been validated against direct survey results.³ Finally, mortality data was obtained from the Center for Disease Control and Prevention Wide-ranging Online Data for Epidemiologic Resource (CDC WONDER).

Our primary exposure was low county-level broadband availability. Low-broadband counties were defined as those in which ≤50% of residents live in places where broadband providers offer internet subscriptions with speeds exceeding 25 megabits-per-second download and 3 megabits-per-second upload, similar to prior studies.^4,5 The main outcomes were county-level clinician (primary care physician, cardiologist) and hospital supply per 100,000 adults, prevalence of cardiovascular risk factors (high blood pressure, high cholesterol among adults screened within the past 5 years, diagnosed diabetes, obesity, and current smoking) and cardiovascular disease (coronary heart disease, stroke) per 100,000 adults, and cardiovascular and all-cause mortality rates.

Categorical and binary variables were compared using χ² tests and continuous variables were compared using 2-sided t-tests. Age-adjusted values for cardiovascular risk factors, cardiovascular disease, and mortality were obtained directly from the data sets and adjusted to the 2000 US Census population using the direct method of standardization.⁶ Differences in rates between low and higher-broadband counties were then calculated for descriptive purposes, as were 95% confidence intervals for these differences using 2-sample t-tests. Statistical tests were 2-sided at a significance level of 0.05. This study used only publicly available county-level data, and therefore Institutional Review Board approval was not required from Beth Israel Deaconess Medical Center, per institutional policy.

RESULTS

Among the 3,143 US counties, 461 (14.7%) had low broadband availability and 2,650 (84.3%) had higher availability in 2019. Counties with low broadband had higher mean age (44.3 vs. 41.5 years, p<0.001) and slightly lower percentage female sex (49.1% vs. 50.1%, p<0.001) (Table 1) compared with counties with higher broadband availability. Low-broadband counties were more likely to be rural than counties with higher-broadband availability (85.5% vs. 58.6%, p<0.001). In addition, college graduation rates were lower in counties with low broadband availability (47.1% vs. 53.9%, p<0.001) as was median annual income ($24,719 vs. $28,200, p<0.001).

Table.

Cardiovascular risk factors, disease, and mortality in counties with low vs. higher broadband availability

Outcome	Low Broadband Availability	Higher Broadband Availability	Absolute Difference (95% CI)	p-value
Age-adjusted prevalence, per 100,000 individuals1
Diabetes2	11,516 (11,293, 11,740)	10,649 (10,562, 10737)	+867 (639, 1095)	<0.001
Hypertension3	34,351 (33,869, 34,834)	32,402 (32,221, 32,583)	+1,950 (1,435, 2,465)	<0.001
Hyperlipidemia4	30,256 (30,038, 30,472)	29,405 (28,317, 29493)	+850 (621,1,079)	<0.001
Smoking5	22,038 (21683, 22392)	20,075 (19,918, 20232)	+1,963 (1,559, 2,366)	<0.001
Obesity6	37,127 (36,768, 37,485)	35,523 (35,359, 35,687)	+1,604 (1,210, 1,998)	<0.001
Coronary artery disease7	6,709 (6,618, 6,800)	6,127 (6,089, 6,164)	+582 (485, 680)	<0.001
Stroke8	3,683 (3,682, 3,614)	3,325 (3,299, 3,349)	+358 (286, 430)	<0.001
Age-adjusted rate, per 100,000 individuals9
Cardiovascular disease mortality	269 (262, 276)	241 (239, 244)	+27 (20, 35)	<0.001
All-cause mortality	854 (838, 870)	810 (804, 816)	+44 (28, 61)	<0.001

^1.Prevalence among adults age ≥18 years, obtained from PLACES based on data from the Behavioral Risk Factor Surveillance Survey

^2.Diabetes = respondents ever in their lifetime told by a healthcare professional that they have diabetes, annual prevalence

^3.Hypertension = respondents told by a healthcare professional that they have high blood pressure within the past year, biennial prevalence

^4.Hyperlipidemia = respondents told by a healthcare professional that they have had high cholesterol within the past five years, biennial prevalence

^5.Smoking = respondents who reported currently smoke every day or some days and having smoked ≥ 100 cigarettes in their lifetime, annual prevalence

^6.Obesity = respondents who have a body mass index ≥30.0 kg/m², calculated from selfreported weight and height, annual prevalence

^7.Coronary artery disease = respondents ever told by a health professional that they had angina or coronary heart disease, annual prevalence

^8.Stroke = respondents ever in their lifetime told by a health professional that they had a stroke, annual prevalence

^9.Mortality rates obtained from CDC WONDER based on based on death certificates with underlying cause of death data collected by the National Center for Health Statistics

Low-broadband counties had fewer cardiologists per capita (0.32 vs. 2.8 per 100,000 individuals, −2.5 [95% Confidence Interval (CI), −2.7, −2.2]) and fewer primary care physicians per capita (31.9 vs. 55.5 per 100,000 individuals, difference −23.6 [95% CI, −26.6, −20.7]) (Figure 1). However, there was no difference in hospital beds per capita between low and high broadband counties (276.4 vs. 303.6 per 100,000 individuals, −27.2 [95% CI, −97.1, 42.8]).

Figure.

The map shows the number of cardiologists (B) and primary care physicians (A) per 100,000 individuals in low (blue) vs. higher- (orange) broadband counties, with darker colors indicating more physicians per capita. Counties with low broadband access had fewer cardiologists and primary care physicians per capita than higher broadband counties. BB = broadband, PCP = primary care physicians, Card. = cardiologist

The age-adjusted prevalence of diabetes, hypertension, hyperlipidemia, smoking, and obesity were higher in low-broadband compared with higher-broadband counties (Table). In addition, counties with low broadband availability had higher age-adjusted prevalence of coronary artery disease and stroke. Unadjusted prevalence of cardiovascular risk factors and disease were similarly higher in low broadband counties. Counties with low broadband availability had higher age-adjusted cardiovascular deaths rates than those with higher broadband availability. In addition, age-adjusted all-cause deaths were higher in low broadband counties.

DISCUSSION

In the United States, counties with low broadband access had substantially fewer cardiologists and primary care physicians per capita than high-broadband counties. In addition, low broadband counties had a greater burden of cardiovascular risk factors and disease, as well as higher cardiovascular and all-cause death rates.

Our study highlights that significant gaps in cardiovascular health and access to care exist between counties with low and higher broadband access. These disparities are likely driven by the many cardioprotective health-related social resources associated with broadband access (e.g. education, employment).^7,8 Regardless, these high-cardiovascular risk counties would benefit from access to outpatient cardiovascular and primary care, which we found is worse in these low-broadband counties. Telehealth has been shown promise as a tool for improving monitoring and access to care, with studies demonstrating improvements in multiple cardiovascular outcomes.⁹ Unfortunately, telehealth access is impaired by low broadband availability.⁴ Our results highlight that, as part of a robust infrastructure policy to support health-related social needs, expanding broadband access and supporting telehealth extension could improve access to needed cardiovascular care in low-broadband counties.

This study has limitations. County-level cardiovascular risk factor and disease prevalence measures were model-based and derived from self-reported telephone survey data.^3,10 Additionally, the goal of this study was not to assess causal relationships, but rather, to describe access to care and cardiovascular health in low vs. higher broadband counties, insights with important implications for infrastructure policy efforts.

In conclusion, United States counties with low broadband access had substantially fewer cardiologists and primary care physicians per capita, a higher burden of cardiovascular risk factors and disease, and higher cardiovascular mortality rates than higher-broadband counties. Ongoing federal policy efforts to expand broadband and increase telemedicine use may have important implications for access to care in low broadband counties.