The Hypothesis
Time-to-treatment is the single most critical determinant of stroke outcome. A patient receiving clot-busting therapy within 60 minutes of stroke onset ("golden hour") is dramatically more likely to survive without disability than one who arrives later. The location of certified stroke centers is therefore not an abstract planning concern — it is a life-and-death variable.
Urban proximity bias in U.S. hospital siting means that rural, low-income, and minority communities are systematically farther from certified stroke centers — and that this geographic inequity compounds the clinical stroke risk factors already concentrated in those same communities.
This project tests that hypothesis empirically for every ZIP code in California using drive-time data, Census demographics, and CMS hospital quality ratings as a proxy for stroke certification capability.
How the Score Works
The SAS uses a CSC-weighted scoring model with 15-minute drive time increments. Comprehensive Stroke Centers are weighted heavily as the only facilities capable of mechanical thrombectomy. Primary Stroke Centers provide modest mitigation only. TSC has been removed — it was assigned by CMS proxy, not authoritative certification data. LA County EMS recognizes only CSC and PSC.
| Drive Time |
CSC Score |
PSC Score |
| 0–15 min | 100 | 40 |
| 15–30 min | 85 | 32 |
| 30–45 min | 65 | 24 |
| 45–60 min | 50 | 16 |
| >60 min | 25 | 8 |
| No certified facility | 0 | 0 |
For each ZIP code, the SAS score is the highest score across all hospitals in the drive-time cache — reflecting the best certified facility realistically reachable. A ZIP with CSC access within 15 minutes scores 100; one with only a PSC more than 60 minutes away scores 8.
0–39
Low
40–59
Med-low
60–69
70–100
High
Scoring Models
The site includes two scoring models, selectable via the toggle in the navigation bar. Both use only CSC and PSC tiers — TSC is not a recognised EMS destination category in California and is excluded from both models.
The Primary Model is the default and is recommended for all general-purpose use. The Sensitivity Model is a research comparison tool — it isolates the methodological effect of switching from two-threshold flat penalties to 15-minute increments. It is not an alternative clinical recommendation.
Primary Model
Default
CSC-weighted, 15-minute drive time increments. Developed with James Beavis RN BSN SCRN (Cedars-Sinai).
| Drive Time | CSC | PSC |
|---|
| 0–15 min | 100 | 40 |
| 15–30 min | 85 | 32 |
| 30–45 min | 65 | 24 |
| 45–60 min | 50 | 16 |
| >60 min | 25 | 8 |
Sensitivity Model
Research
Same tiers, flat penalties at two thresholds. Use to compare against the Primary Model's 15-minute increment structure.
| Drive Time | CSC | PSC |
|---|
| 0–30 min | 100 | 50 |
| 30–60 min | 70 | 35 |
| >60 min | 55 | 20 |
Vulnerability Flags
Three flags identify ZIP codes where poor access intersects with community-level risk factors.
SEV — Socioeconomic Vulnerability
SAS score < 40, majority-minority community, AND poverty rate > 20% or uninsured rate > 15%.
CRF — Clinical Risk Factor
SAS score < 40, AND median age > 40 — indicating elevated population-level stroke risk.
Double Jeopardy
Both SEV and CRF flags are active — communities facing simultaneously poor access and high stroke risk burden.
Data Sources
-
LA County EMS Approved Stroke Center Directory
Authoritative source for LA County hospital stroke certifications. July 28, 2025 edition. Used for all LA County CSC and PSC designations.
-
CMS Hospital Compare — Rest of California
Hospital characteristics and facility type for non-LA-County hospitals. TSC proxy removed — TSC was assigned by CMS rating, not authoritative certification data. EMNet national certification data pending for future versions.
-
Manual Hospital Additions
8 hospitals present in the LA County EMS directory were absent from the CMS dataset and added manually, including drive-time collection for each.
-
U.S. Census — ACS 5-Year Estimates, 2022
ZCTA-level demographics including total population, race/ethnicity, median household income, poverty rate, health insurance coverage, and median age.
-
Census TIGER/Line Shapefiles, 2022
ZCTA boundary geometries for choropleth mapping and area calculations. Simplified to 0.005° tolerance for web performance.
-
Census 2020 Decennial — Block-level Population
PL 94-171 redistricting data used to compute population-weighted centroids for each ZCTA, replacing geometric centroids for more accurate geographic representation.
-
Drive-Time Cache
13,511 precomputed origin-destination drive times between California ZCTAs and geocoded hospitals. Captures up to 5 nearest hospitals per ZIP by straight-line distance. Single 8am weekday departure.
Coverage
Statistics computed live from the scoring dataset.
—
California ZIP codes scored statewide
—
of 312 LA County ZIP codes scored
(9 excluded — non-residential)
341
LA County hospitals in dataset
—
LA County — limited access (0–39)
—
LA County — moderate access (40–69)
—
LA County — strong access (70–100)
Limitations & Caveats
- Certification data outside LA County: Hospital certification outside LA County uses CMS proxy data. TSC has been removed as it was a CMS proxy artifact, not a recognized EMS certification tier in California. EMNet national certification data is pending for future versions.
- Drive time methodology: Drive times are based on a single 8am weekday departure point. A multi-time-point methodology capturing morning, midday, and evening departures (0800 / 1200 / 1700) is planned for future versions.
- Drive time cache coverage: The drive time cache captures the 5 nearest hospitals per ZIP code by straight-line distance. Some nearby certified facilities may be absent from a given ZIP's cache if a closer uncertified hospital displaced them in the pre-selection step.
About the Authors
JB
James Beavis RN BSN SCRN
Registered Nurse — Hypothesis, clinical methodology, and stroke care domain expertise
MM
Mike Morales
Data pipeline, geographic analysis, and visualization