About this dashboard
This dashboard maps three interlocking dimensions of US cardiovascular care at county and
census-tract resolution: the supply of cardiologists, the burden of cardiovascular disease in the
population, and the digital infrastructure that supports remote-care delivery.
Digital Health Prioritization (primary layer)
The default map answers a prioritization question: among US communities of greatest
cardiovascular need, which carry higher digital readiness and may be
positioned for digital health deployment, and which carry lower readiness
and may require infrastructure investment first?
We first define a single workforce-constrained county pool, the union of:
(a) counties with NO cardiologists in the latest panel year, and
(b) counties with a DECLINING PER-CAPITA cardiology workforce
(OLS slope of cardiologists per 100,000 < 0 over 2010-2023;
counties with mean < 3 use the last-vs-first rate)
Within that pool, at census-tract resolution, two anchored cuts form a
two-by-two matrix (the framing the manuscript uses):
high burden : burden_z > 0 (above the national-average tract)
higher readiness : DDI ≤ national median (18.77)
DEPLOYMENT PRIORITY = high burden AND higher readiness
→ may be positioned for digital health deployment
INVESTMENT PRIORITY = high burden AND lower readiness
→ may require infrastructure investment first
To rank tracts within each group, burden and DDI are standardized
within the pool and combined into a priority score: burden minus DDI
for deployment priority, burden plus DDI for investment priority.
These labels indicate relative priorities under the chosen thresholds.
They do not establish readiness for any specific intervention.
Counties are shaded by their tract mix: Deployment-lean, Investment-lean,
or Mixed (containing both priority groups, the within-county heterogeneity that
county-level analysis hides). Zoom in or open a county to see its individual tracts.
The two groups are disjoint; every priority tract is either deployment-priority or
investment-priority, never both.
Kentucky and Pennsylvania. CDC suppressed nine of the ten
PLACES measures at the tract level in these two states, so their tract burden
composite rests on short sleep alone. Their tracts keep box membership and
count toward the national totals, but they are ineligible for the ranked
national top lists, and the burden map views show them as no data.
The burden composite
Cardiometabolic burden is a tract-level composite z score across 10
conditions and risk factors: hypertension, high cholesterol, coronary
heart disease, stroke, diabetes, obesity, current smoking, physical
inactivity, short sleep duration, and binge drinking. For each measure,
prevalence is averaged across the two most recent CDC PLACES releases
(2022 and 2023), standardized across tracts, and the resulting z scores
are averaged. The Trend view under Disease burden compares the two
releases directly.
Workforce measures
Cardiology workforce counts are county-level, non-federal cardiovascular
disease physicians from the Area Health Resources Files (2010 to 2023).
The priority-pool decline rule uses the trend in cardiologists
per 100,000 residents, as in the manuscript. The Trend
map view classifies counties by the trend in cardiologist counts;
counties averaging fewer than 3 cardiologists across the panel show as
"no cardiologists or too few to assess." The + NPs/PAs view adds
cardiology-affiliated nurse practitioners and physician assistants from
the CMS Doctors and Clinicians file. A clinician counts as
cardiology-affiliated when cardiologists make up at least half of the
physicians at the same practice site.
Data sources
- Cardiology workforce, 2010-2023. HRSA Area Health Resources Files; non-federal cardiovascular disease physicians.
- Cardiology NPs/PAs. CMS Doctors and Clinicians (DAC) national file, practice-site affiliation method.
- Primary care physicians, 2023. AHRF; patient-care MD/DO excluding residents.
- Disease burden. CDC PLACES, 2022 and 2023 releases; 10 cardiometabolic measures.
- Digital readiness. Purdue Digital Divide Index (2022-2024), built from American Community Survey and FCC data; infrastructure/adoption (INFA) and socioeconomic (SE) sub-scores. Higher DDI = larger divide = lower readiness.
- Population and geography. ACS 2020-2024 five-year estimates; US Census Bureau cartographic boundaries.
Robustness
The manuscript tests the framework in eight sensitivity analyses:
readiness rebuilt from six ACS indicators (with and without FCC
broadband availability), the two DDI sub-scores used separately, the
matrix applied to all US tracts without the workforce screen, stricter
burden and readiness thresholds, a 25-mile geographic access credit,
the workforce broadened to cardiology-affiliated NPs/PAs, the pool
rebuilt on primary care physicians, and a KY/PA burden reconstruction
from the newest PLACES release. Classifications were broadly stable;
for example, a six-indicator readiness index placed tracts in the same
group about 93% of the time.
Display notes
The readiness tiers shown in county and tract panels are national
terciles of the tract-level DDI. They are a dashboard display aid, not
part of the manuscript classification.
Limitations
- Workforce counts are county-level only; every tract inherits its county's workforce measures.
- Tract-level PLACES reports crude prevalence, not age-adjusted.
- Kentucky and Pennsylvania tract burden rests on one measure (short sleep); their tracts hold group membership but are not ranked.
- The priority labels are relative to the chosen national thresholds. They do not certify any tract for a specific intervention.
Accessibility
Both dashboards were designed to meet WCAG 2.1 Level AA. All controls
work by keyboard; the search box and the tract table form the keyboard
path to any county or tract; and the palettes were checked under common
forms of colour vision deficiency.
CarDS Lab, Yale School of Medicine. A companion view, the
implementation decision space, plots
every workforce-constrained tract by burden and readiness.