Mapping the Availability of Rehabilitation Providers Using Public Licensure and Population Data for a Geographic Information System-Based Approach to Workforce Planning: Cross-Sectional Feasibility Study.

BACKGROUND: Access to rehabilitation services is a critical yet under-studied dimension of health equity. Among the 6 domains of access, health care provider availability, defined as the presence of sufficient health care providers to meet population needs, is particularly underexplored in rehabilitation professions such as physical and occupational therapy. Current data reporting often lacks the geographic granularity required for effective workforce planning. OBJECTIVE: The purpose of this study was to demonstrate the feasibility of mapping rehabilitation provider availability at the census tract level using geographic information systems and integrating public licensure and population data to inform equitable workforce planning. METHODS: A descriptive, cross-sectional study was conducted using publicly available state licensure data for physical and occupational therapists and demographic data from the American Community Survey. Residential addresses of rehabilitation providers were geocoded and matched to 2020 census tracts. Population-to-provider ratios were calculated and mapped using choropleth and bivariate mapping techniques. Population-to-provider ratios were calculated per tract and summarized overall and by rurality using 2020 Rural-Urban Commuting Area (RUCA) codes (urban: RUCA of 1-3; rural: RUCA of ≥4). The spatial dependence of ratios was tested using a spatial autocorrelation statistic, the global Moran I, in ArcGIS Pro using edge contiguity neighbors and row standardization. RESULTS: Across 6896 tracts, ratios ranged from 4.5 to 11,147 persons per provider (median 1131, IQR 537-2501). By rurality, urban tracts (n=5734, 83.1%) had a median ratio of 1141 (IQR 2054), and rural tracts (n=1162, 16.9%) had a median ratio of 1093 (IQR 1690), indicating a broadly similar central tendency with somewhat greater variability in urban areas. The population-to-provider ratio exhibited significant positive spatial autocorrelation (global Moran I=0.305; Z=40.28; P<.001), consistent with clustered pockets of high and low availability rather than random dispersion. CONCLUSIONS: A replicable geographic information system protocol can integrate licensure and demographic data to produce interpretable population-to-provider metrics and spatial diagnostics at the census-tract level. In Texas, rehabilitation workforce availability is spatially clustered and not explained solely by an urban-rural divide, underscoring the value of small-area mapping for equitable workforce planning and policy decisions.