Long-term PM2.5 exposure as a risk factor for 30-day hospital readmissions among heart failure patients
On this page:
Background:
Hospital readmissions represent a health and economic burden to patients and the healthcare system. We understand little about the relationship between long-term environmental exposures and hospital readmission risk. Here, we use electronic health records linked to exposure data (EPA CARES) to quantify the associations between long-term PM2.5 exposure at the time of heart failure (HF) diagnosis and the number of hospital readmissions – a key outcome metric for quality of care for hospitals and healthcare systems.
Methods:
Annual average PM2.5 was estimated at 1x1 km resolution for each primary residence. We used stabilized inverse probability weights to account for imbalanced confounder distributions and the competing risk of death. We adjusted for age at HF diagnosis, race, sex, year of HF diagnosis, and neighborhood socioeconomic status in quasi-Poisson regression models with an offset of log(follow-up time). Associations are presented as the percent increase from the mean and associated 95% confidence interval (CI).
Results:
Out of 34,459 HF patients, 18,078 had at an inpatient or emergency admission, thus qualifying them to have a readmission and enter this analysis. A 1 µg/m3 increase in annual average PM2.5 was associated with an 11% (95% CI = 6 – 16) increase in 30-day readmissions; independent of short-term (5-day average) exposures prior to the readmissions. PM2.5 exposure was also associated with a 13% (95% CI = 4 – 22) and 14% (95% CI = 10 – 18) increase in 7 and 90 day readmissions, respectively.
Conclusion:
Long-term PM2.5 exposure is associated with increased hospital readmissions, including 30-day readmissions, a closely tracked metric by which hospitals and healthcare systems are evaluated. Associations were independent of short-term variations in PM2.5 prior to readmission, indicating that long-term exposures may be an important contributor to hospital readmission risk that should be further explored. This abstract does not necessarily represent EPA policy.