Examining differences in the association between apparent temperature difference and preterm birth by census-tract level utility-bill burden in North Carolina, 2011-2015
On this page:
Objectives: We examined the association between daily diurnal difference in apparent temperatures (DAT) and the risk of preterm birth (PTB) during a period of three days before birth and explored whether this association was modified by energy utility burden.
Material and Methods: Daily, census tract-level measures of DAT (degrees Fahrenheit [°F]) were created using the North American Land Data Assimilation System model. Study population included singleton live births at ≥20 weeks gestation with a geocoded North Carolina residence at date of birth linked to census tract. Contextual data on utility burden (percentage of gross household income spent on energy cost) from the U.S Department of Energy was linked by census tract. Time series regression analyses with distributed lag non-linear models (DLNM) were used to model associations between DAT and PTB (<37 weeks) for a 3-day lag period. We adjusted for long-term time trends and results were stratified by the median utility burden.
Results: Minimum PTB risk was observed on days with 15 °F DAT for the NC birth cohort and the risk peaked at both ends of the DAT distribution. For 1 °F and 35 °F DAT the risk ratios (RR) for PTB were 1.309 (95% CI: 1.107, 1.545) and 1.148 (1.065, 1.237). Comparing tracts with high energy burden to low energy burden tracts, the risk was higher for low DAT than high DAT. For a 1 and 35 °F DAT, the RR for PTB were 1.215 (0.922, 1.601) and 1.088 (0.938, 1.261) for low energy burden tracts, and 1.235 (0.948, 1.610) and 1.037 (0.924, 1.164) for high energy burden tracts.
Conclusion: There may be a non-linear association between DAT and PTB counts. The risk associated with low DAT may be higher in census tracts with a high utility burden.