Purpose
To determine spatial variability in the rate of strabismus diagnosis among children enrolled in Medicaid in two US states. Understanding geographic variation illustrates issues in disparities and can improve the delivery of care.
Methods
Children age ≤10 enrolled in Medicaid in Michigan and North Carolina during 2009/2010 were identified from the Medicaid Analytic Extract (MAX) health care claims database. Residential location for each child was determined by last known 5 digit zip code, connected to a Zip Code Tabulation Area (ZCTA) for georeferencing and spatial analysis. ICD-9-CM billing codes were used to identify children diagnosed with strabismus (code 378.xx). Bayesian hierarchical intrinsic conditional autoregressive (ICAR) spatial probit models were used to model the prevalence across the Lower Peninsula of Michigan and the Raleigh-Durham-Cary Combined Statistical Area (CSA). Based on spatial random effects estimates, maps of the average predicted risk of strabismus diagnosis were created. ZCTAs with increased and decreased risk according to 95% credible intervals were identified.<br /> <br /> .
Results
Of approximately 500,000 eligible children in Michigan, roughly 7500 (~ 1.5%) received ≥ 1 strabismus diagnosis in the analysis time period, with an interquartile range (IQR) of approximately 0.9% to 2.1%. Communities with lower strabismus diagnosis rates included Flint, Saginaw, and portions of Lansing and Detroit. Areas of increased strabismus rates included Traverse City and portions of Grand Rapids and Detroit. Of over 90,000 eligible children in the Raleigh-Durham-Cary CSA, approximately 800 (~ 0.9%) received a strabismus diagnosis in the study period, with an IQR of roughly 0.4% to 1.2%). Areas of decreased strabismus diagnosis rate included Siler City and parts of eastern Raleigh. Areas of increased strabismus rates included Durham and Chapel Hill.<br /> <br />
Conclusions
Analysis of geographic patterns of care associated diagnoses can reveal differences across geographic regions at both state and local units of analysis. The information can inform decisions about resource allocation for expanding access to eye care services by targeting interventions at areas with lower than expected rates of diagnosis.