Estimating Population Average Causal Effects in the Presence of Non-Overlap: A Bayesian Approach

May 24, 2018 By Rachel Nethery, Fabrizia Mealli, and Francesca Dominici Most causal inference studies rely on the assumption of positivity, or overlap, to identify population or sample average causal effects. When this assumption is violated, these estimands are unidentifiable without some degree of reliance on model specifications, due to poor data support. Existing methods to address non-overlap, such as trimming or down-weighting data in regions of poor support, all…

High-dimensional confounding adjustment using continuous spike and slab priors

In observational studies, estimation of a causal effect of a treatment on an outcome relies on proper adjustment for confounding. If the number of the potential confounders ( p ) is larger than the number of observations ( n ), then direct control for all potential confounders is infeasible. Existing approaches for dimension reduction and penalization are generally aimed at predicting the outcome, and are less suited for estimation of…

Best Practices for Gauging Evidence of Causality in Air Pollution Epidemiology

American Journal of Epidemiology | September 6, 2017 The contentious political climate surrounding air pollution regulations has brought some researchers and policy makers to argue that evidence of causality is necessary for more stringent regulations. Recently, an increasing number of air pollution studies purport the use of “causal analysis,” generating the impression that studies not explicitly labeled as such are merely “associational” and therefore less rigorous. Using three prominent air…

Air Pollution and Mortality in the Medicare Population

New England Journal of Medicine | June 29, 2017 Studies have shown that long-term exposure to air pollution increases mortality. However, evidence is limited for air-pollution levels below the most recent National Ambient Air Quality Standards. Previous studies involved predominantly urban populations and did not have the statistical power to estimate the health effects in underrepresented groups. Read more here

Air Pollution Still Kills

Editorial, New England Journal of Medicine | June 29, 2017 In late October 1948, a dense smog descended over the town of Donora, Pennsylvania. The town was home to a zinc plant and a steel mill, both run by the United States Steel Corporation. Susan Gnora, a 62-year-old resident of Donora, started to gasp and cough as the smog descended. She died the next day. Dr. William Rongaus, a physician…

Drought linked with human health risks in US analysis, Berman et al.

In a retrospective study of health claims for 618 U.S. counties over 14 years, published in The Lancet, researchers found that severe drought conditions increased the risk of mortality among adults 65 or over. They also found that individuals in places where droughts were rare, such as counties in Minnesota, showed a larger risk of mortality and cardiovascular disease compared with counties where drought is more common. Read more here

Cardiovascular disease-related hospital admissions jump on second day after major snowfall, Bobb et al.

According to a new study by researchers from Harvard T.H. Chan School of Public Health, hospital admissions for cardiovascular diseases decline on days with major snowfalls compared to days with no snowfall, but they jump by 23% two days later. Lead author Jennifer Bobb of the Group Health Research Institute in Seattle, who worked on the study as a postdoctoral researcher in the Department of Biostatistics at Harvard Chan School,…

Readmission Rates After Passage of the Hospital Readmissions Reduction Program: A Pre–Post Analysis, Wasfy et al.

Passage of the Patient Protection and Affordable Care Act (ACA) in March 2010 created the Medicare Hospital Readmissions Reduction Program (HRRP), which introduced the prospect of financially penalizing hospitals based on their previous performance. This program represents one of several efforts to encourage provider organizations to enhance safety and value. Medicare has reported lower readmission rates since passage of the law, but whether the lowest-performing hospitals might experience less rapid improvement…

Hierarchical Models for Semicompeting Risks Data With Application to Quality of End-of-Life Care for Pancreatic Cancer, Lee et al.

Readmission following discharge from an initial hospitalization is a key marker of quality of healthcare in the United States. For the most part, readmission has been studied among patients with “acute” health conditions, such as pneumonia and heart failure, with analyses based on a logistic-Normal generalized linear mixed model. Naïve application of this model to the study of readmission among patients with “advanced” health conditions such as pancreatic cancer, however,…

“Direct” Approach Evaluates Air Quality Interventions, Zigler et al.

Causal Inference Methods for Estimating Long-Term Health Effects of Air Quality Regulations, was funded as part of HEI’s Accountability research program, aimed at understanding whether actions to improve air quality have resulted in improved health outcomes. Corwin Zigler and his colleagues used existing and newly developed statistical methods to assess whether an intervention was causally related to changes in pollutant levels or health outcomes, and applied their methods in two…