Environment/Climate Change
Residential mobility and exposure misclassification of PM₂.₅ among pregnant Medicaid enrollees from 2001-2014 Stefania Papatheodorou* Stefania Papatheodorou Papatheodorou Papatheodorou Papatheodorou Papatheodorou Papatheodorou Papatheodorou Papatheodorou Papatheodorou Papatheodorou Papatheodorou Papatheodorou Papatheodorou Papatheodorou Papatheodorou Rutgers School of Public Health
Background:
Residential mobility during pregnancy can introduce exposure misclassification in environmental epidemiology and is socially patterned. However, large-scale evidence on prenatal mobility and the direction and magnitude of associated changes in PM₂.₅ exposure remains limited.
Methods:
We analyzed 1,548,303 Medicaid-covered pregnancies in the United States (2001–2014). Residential mobility was defined as a change in ZIP code between the last menstrual period (LMP) and delivery. Movers were classified by directional change in ambient PM₂.₅ exposure between LMP and delivery (High→High, High→Low, Low→High, Low→Low), based on the median annual ZIP code PM₂.₅ (10.8 µg/m³). Maternal sociodemographic, neighborhood, and clinical characteristics were compared using standardized mean differences (SMDs).
Results:
Overall, 259,918 pregnancies (16.8%) involved residential mobility. Movers were younger (mean age 23.4 vs. 24.8 years; SMD −0.23), more likely to be Black (37.0% vs. 31.6%), and had a median (IQR) move distance of 12.4 (6.5–23.8) km. Most moves were from high-to-high (32.7%) or low-to-low (31.9%) PM₂.₅ areas. The overall mean PM₂.₅ difference was −0.55 µg/m³ (SD 4.12), masking substantial heterogeneity, as indicated by the movement from low to high PM₂.₅.₅ areas (13%) were associated with a mean increase of 3.9 µg/m³, while moves from high- to low-PM₂.₅ areas (18%) showed a mean decrease of −4.7 µg/m³. Moves into higher-exposure areas were associated with greater neighborhood disadvantage, whereas moves into lower-exposure areas were associated with higher household income, homeownership, and lower neighborhood poverty. Movers also had modestly higher prevalence of selected comorbidities (all SMDs <0.20).
Conclusions:
Approximately one in six Medicaid-covered pregnancies involved residential mobility, often over short distances but frequently accompanied by meaningful changes in PM₂.₅ exposure and neighborhood context. These findings underscore residential mobility as an important source of differential exposure misclassification in perinatal environmental health studies.