ZIP Code and ZIP Code Tabulation Area Linkage: Implications for Bias in Epidemiologic Research.

Area of Science:

• Geospatial analysis
• Public health research
• Data linkage methodologies

Background:

• No standardized methods exist for linking U.S. Census ZIP Code Tabulation Areas (ZCTAs) with U.S. Postal Service ZIP Codes (ZIPs).
• One-to-one linkage can lead to data loss by excluding ZIPs without direct ZCTA matches.
• Crosswalk linkage offers a solution by allowing one ZCTA to map to multiple ZIPs, preventing data exclusion.

Purpose of the Study:

• To compare the outcomes of noncrosswalk versus crosswalk linkage methods for ZCTAs and ZIPs.
• To assess the sociodemographic implications of data exclusion in one-to-one linkage.
• To evaluate the impact of linkage methods on mortality and health insurance data.

Main Methods:

• National and California-specific comparisons of noncrosswalk and crosswalk linkage strategies.
• Utilized generalized additive models to analyze the relationship between sociodemographics and nonmatching ZIPs within ZCTAs.
• Examined data from mortality and health insurance datasets.

Main Results:

• Nationwide, 15% of ZCTAs contained nonmatching ZIPs, which were more prevalent in metropolitan cores and associated with lower socioeconomic status and non-White populations.
• In California, 34% of mortality ZIPs and 25% of health insurance ZIPs had nonmatching ZCTAs.
• These nonmatching ZIPs represented a small proportion of the overall data (0.03% for mortality, 0.44% for insurance).

Conclusions:

• Crosswalk linkage and using ZCTAs as the analytical unit are supported by study findings.
• One-to-one linkage may introduce bias by disproportionately excluding ZIP codes serving disadvantaged populations.
• While potentially biased, the population sizes affected by one-to-one linkage appear minimal.