Structural inequality shapes spatiotemporal patterns of heat-related emergency calls.

Discrepancies between projected and observed impacts of heat hazard underscore the need to understand how heat exposure interacts with sociodemographic and environmental factors to shape health outcomes. However, current methodological approaches overlook unobserved spatial and temporal heterogeneity, limiting their policy relevance. This study examines how advances in heat vulnerability assessment can reveal key predictors of heat-related emergency calls for improved heat resilience planning. Using persistent change analysis, year-on-year difference maps, and a Poisson Pseudo-Maximum Likelihood spatial lag of X (PPML-SLX) model with two-way fixed effects, we identify which sociodemographic and environmental characteristics buffer or amplify heat risk over time in the City of Phoenix. Results reveal increasing emergency calls in south and central Phoenix, and decreasing in the north. Tracts with increasing call trends had higher rates of uninsurance, poverty, and disability, and lower educational attainment. They also had higher proportions of non-US citizens, individuals with limited English proficiency, households without a vehicle, and Hispanic residents (standardized mean difference [SMD] = 0.55-0.75). In contrast, tracts with decreasing trends had more non-Hispanic White residents and higher housing costs (SMD = 0.57-0.77). Our findings indicate that the spatial structure of heat-related emergencies is primarily driven by structural inequality when accounting for time and tract fixed effects. The negative association between emergency calls and social attributes highlights barriers to service use, rather than reduced heat-related emergencies. Future studies should integrate methodological improvements with social relevance, ensuring that advances in modeling translate into more equitable and targeted interventions.