A18-01 Insights From the Use of Sub-daily PM2.5 in Respiratory Health Analyses

Abstract Rationale Public health assessments of PM2.5 typically rely on daily 24-hour averages; however, PM2.5 is dynamic on hourly timescales due to emissions, atmospheric chemistry, and meteorology. This temporal mismatch may hide short, high-intensity exposures, limiting our understanding of sub-daily PM2.5 events and their implications. This study has two goals: 1) to explore sub-daily peaks and associated respiratory outcomes and 2) to look at misclassification of hourly exposure with the Air Quality Index (AQI) and potential respiratory risk differences. Methods This study used an hourly 3-km gridded PM2.5 dataset for California (2018-2023), aggregated to ZIP-code levels via population-weighted centroid assignment. For the first goal, daily means and sub-daily metrics (e.g., daily 1-hour maximum) were calculated and compared using spearman correlation coefficients. For the second goal, hourly and daily values were assigned AQI categories using 2024 EPA-defined PM2.5 breakpoints to compare hourly category frequencies across days. To evaluate differences in health risks across daily and subdaily metrics, we queried UCSF Health system asthma- and COPD-related hospitalizations and emergency department visits (EDVs) identified by ICD codes for 2018 to 2023, and estimated their daily ZIP-code sums as a function of (1) daily means or sub-daily PM2.5 metrics and (2) daily AQI categorization or hourly counts of AQI categories in conditional quasi-Poisson models, adjusting for temperature and temporal trends. Results Across California, the average 1-hour maximum PM2.5 was approximately twice the 24-hour mean over the 6-year study period. Spatially, ZIPs with high 1-hour maxima had higher 24-hour averages (R2 = 0.76). All sub-daily and daily metrics were positively associated with respiratory outcomes. When scaled to metric-specific IQR, the average daily PM2.5 had the lowest risk (IRR: 1.014, 95% CI: 1.007, 1.022) whereas the 1-hour max had the highest risk (IRR: 1.018, 95% CI: 1.009, 1.027), though all metrics had overlapping intervals. Hours that occur on days with the same day-level AQI category span a wide range of PM2.5 concentrations. For example, 18% of hours occurring on days labeled “Unhealthy for Sensitive Groups” exceeded the category’s threshold, 55.4 μg m-3. However, for “Unhealthy for Sensitive Groups” and higher risk AQI categories, the positive risk associated with hourly AQI diverged with and was greater than daily AQI starting from 10-18 hours of exposure. Conclusion Sub-daily PM2.5 metrics identify acute exposures missed by daily averages and that materially change AQI-based characterizations. Incorporating hourly exposure metrics could improve acute-risk communication and interventions during pollution events. This abstract is funded by: EPA