Long-Term Association of Air Pollution and Hospital Admissions Among Medicare Participants Using a Doubly Robust Additive Model

There are limited studies examining the non-fatal health outcomes of air pollution exposure due to the number of pollutants studied and focus on short-term exposures. The author Mahdieh Danesh Yazdi and colleagues (2021) published a study in the journal of American Heart Association under the title “Long-Term Association of Air Pollution and Hospital Admissions Among Medicare Participants Using a Doubly Robust Additive Model”. The summary of this paper is given below:

Objective:

To investigate the relationship between average annual fine particulate matter with an aerodynamic diameter of <2.5 micrometers [PM2.5, O3 and NO2, and NO4] and cardiovascular and respiratory hospitalization outcomes (ischemic stroke; myocardial infarction [MI]; atrial fibrillation and flutter; and pneumonia)

Method:

The doubly robust method was used for the statistical analysis. This analysis relied on both adjustments in the outcome model to account for confounding and inverse probability weighting. The outcome from this regression was on an additive scale. Further studies have been conducted to investigate this relationship at lower pollutant concentrations, which are not inconsistent with typical exposure levels in the United States, and among potentially susceptible subgroups.

Findings:

Investigators found that PM2.5 was associated with a high risk of hospital admissions, especially in elderly individuals. Whereas, there was an increased risk of stroke and atrial fibrillation and flutter associate with NO2. While O3 exposure was found to be associated with an increased probability of pneumonia, it is a risk factor for cardiovascular outcomes at modest concentrations. Additionally, all pollutants increased the probability of hospital admissions at lower concentrations with larger effect estimates than the primary results.

Subjects who are identified as White were at higher risk of atrial fibrillation attributable to NO2 than those who identified as Black. On the other hand, those who identified as Black were at higher risk of stroke attributable to NO2 than those who identified as White.

Limitations:

As the causal methodology used relied on the strong assumption of measured confounding which is not testable. Authors acknowledge that causality is not proven, and can only be an interpretation, including support from toxicology. There is a chance for over controlling for smoking and underestimating the true effect size as the study approach was more conservative and controlled for lung cancer rate as a proxy for smoking. Lastly, investigators assumed that the loss of follow-up among the subjects was unrelated to air pollution.

Image Credit: People photo created by jcomp – www.freepik.com