E2SFCA Explained: The Math Behind Healthcare Access Scores

If you have ever wondered how CartoChrome turns millions of data points into a single 0-100 Healthcare Access Score, the answer starts with a method called E2SFCA -- the Enhanced Two-Step Floating Catchment Area. Developed by Wei Luo and Fahui Wang in 2003 and refined by Luo and Yi Qi in 2009, E2SFCA has become the most widely cited method for measuring spatial access to healthcare in the academic literature. CartoChrome brings it out of the research lab and into a consumer product for the first time.

The Intuition

Imagine you live in a small town with one family doctor who serves 2,000 patients. Your neighbor lives in a suburb with 50 doctors serving 80,000 people. Who has better access to primary care?

Simple provider counts would say the suburb is better -- 50 doctors versus 1. But when you account for population demand, the ratios tell a different story: the small-town doctor serves 1:2,000, while each suburban doctor serves 1:1,600. The difference is smaller than the raw count suggests, and other factors (travel distance, wait times, insurance acceptance) might tip the balance either way.

E2SFCA formalizes this intuition into a rigorous mathematical framework that accounts for both the supply of healthcare providers and the demand from the population they serve, weighted by the distance between them.

Step 1: Compute Provider-to-Population Ratios

For every healthcare provider (or facility) location *j*, the algorithm defines a catchment area -- the maximum distance within which patients would reasonably travel to reach that provider. Then it sums the population of all ZIP codes that fall within that catchment, applying a **distance-decay function** that gives more weight to nearby populations and less weight to distant ones.

The formula for the supply-to-demand ratio at provider *j* is:

**R_j = S_j / SUM(P_k * f(d_kj))** for all ZIP codes *k* within the catchment

Where: - **S_j** is the supply capacity of provider *j* (e.g., number of physicians at that location) - **P_k** is the population of ZIP code *k* - **f(d_kj)** is the distance-decay function applied to the distance between ZIP code *k* and provider *j* - The sum runs over all ZIP codes within the catchment radius

Step 2: Sum Access Scores for Each ZIP Code

For every ZIP code *i*, the algorithm identifies all providers whose catchment areas overlap that ZIP code. It then sums their supply-to-demand ratios, again weighted by distance decay:

**A_i = SUM(R_j * f(d_ij))** for all providers *j* whose catchment overlaps ZIP *i*

This produces a single access score for each ZIP code that reflects both the quantity of nearby providers and the competition for their services from other nearby populations.

Why Distance Decay Matters

The choice of distance-decay function is critical. In the original 2SFCA method (before the "Enhanced" version), all locations within the catchment were weighted equally -- a provider 1 mile away counted the same as one 29 miles away. This created artificial boundary effects and unrealistic access estimates.

E2SFCA introduced continuous decay functions that model how willingness to travel decreases with distance. CartoChrome uses different decay functions for different healthcare types, calibrated to real-world travel behavior:

• Gaussian decay for primary care, dental, mental health, and preventive services -- utilization drops gradually with distance in a bell-curve pattern
• Sigmoid decay for emergency and trauma care -- utilization is roughly constant up to a critical threshold (about 15-20 minutes), then drops sharply, reflecting the time-sensitive nature of emergencies
• Logistic decay for specialist care -- a moderate decline that reflects patients' willingness to travel further for specialized services

Catchment Radii: One Size Does Not Fit All

A critical refinement in CartoChrome's implementation is the use of RUCA-based catchment tiers. The Rural-Urban Commuting Area (RUCA) classification system divides the country into urbanicity tiers based on commuting patterns and population density. CartoChrome defines 5 tiers:

• Urban core -- Dense city centers (smallest catchment radii)
• Suburban -- Surrounding metro areas
• Large rural -- Towns with 10,000+ population
• Small rural -- Towns with 2,500-10,000 population
• Isolated rural -- Areas with fewer than 2,500 people (largest catchment radii)

Each of the 8 healthcare components has a distinct catchment radius for each urbanicity tier, producing a 5x8 matrix of 40 calibrated radii. This prevents rural areas from being penalized for having providers at distances that are perfectly normal for rural life, while still recognizing that urban residents expect and need shorter travel times.

From Raw Scores to the 0-100 Scale

The raw E2SFCA scores are in arbitrary units that vary by healthcare type and are not directly interpretable. CartoChrome normalizes these into a 0-100 scale using a log-softcap at the 95th percentile. This means:

• The bottom of the distribution maps to scores near 0
• The median maps to roughly 50
• The 95th percentile maps to 95
• The top 5% is log-compressed into the 95-100 range

This approach preserves the full rank ordering (every ZIP code maintains its relative position) while producing human-readable scores where 50 genuinely means "middle of the pack."

Validation: Does It Actually Work?

A scoring methodology is only as good as its correlation with real-world outcomes. CartoChrome validates E2SFCA scores against three independent benchmarks:

• CDC PLACES utilization measures (target: r > 0.65) -- Do areas with high access scores actually show higher utilization of preventive services?
• County Health Rankings (target: r > 0.70) -- Do our scores align with the Robert Wood Johnson Foundation's county-level health assessments?
• HRSA HPSA concordance (target: > 70%) -- Do areas we flag as low-access overlap with HRSA's official Health Professional Shortage Area designations?

Meeting these thresholds gives us confidence that E2SFCA, as implemented by CartoChrome, captures meaningful variation in healthcare access rather than just statistical noise.