Statistical inference for complete and incomplete mobility trajectories under the flight-pause model

We formulate a statistical flight-pause model (FPM) for human mobility, represented by a collection of random objects, called motions, appropriate for mobile phone tracking (MPT) data. We develop the statistical machinery for parameter inference and trajectory imputation under various forms of missing data. We show that common assumptions about the missing data mechanism for MPT are not valid for the mechanism governing the random motions underlying the FPM, representing an understudied missing data phenomenon. We demonstrate the consequences of missing data and our proposed adjustments in both simulations and real data, outlining implications for MPT data collection and design.

Bayesian nonparametric adjustment of confounding

Analysis of observational studies increasingly confronts the challenge of determining which of a possibly high-dimensional set of available covariates are required to satisfy the assumption of ignorable treatment assignment for estimation of causal effects. We propose a Bayesian nonparametric approach that simultaneously (1) prioritizes inclusion of adjustment variables in accordance with existing principles of confounder selection; (2) estimates causal effects in a manner that permits complex relationships among confounders, exposures, and outcomes; and (3) provides causal estimates that account for uncertainty in the nature of confounding. The proposal relies on specification of multiple Bayesian additive regression trees models, linked together with a common prior distribution that accrues posterior selection probability to covariates on the basis of association with both the exposure and the outcome of interest. A set of extensive simulation studies demonstrates that the proposed method performs well relative to similarly-motivated methodologies in a variety of scenarios. We deploy the method to investigate the causal effect of emissions from coal-fired power plants on ambient air pollution concentrations, where the prospect of confounding due to local and regional meteorological factors introduces uncertainty around the confounding role of a high-dimensional set of measured variables. Ultimately, we show that the proposed method produces more efficient and more consistent results across adjacent years than alternative methods, lending strength to the evidence of the causal relationship between SO2 emissions and ambient particulate pollution.

Mortality risk from United States coal electricity generation

Policy-makers seeking to limit the impact of coal electricity-generating units (EGUs, also known as power plants) on air quality and climate justify regulations by quantifying the health burden attributable to exposure from these sources. We defined "coal PM2.5" as fine particulate matter associated with coal EGU sulfur dioxide emissions and estimated annual exposure to coal PM2.5 from 480 EGUs in the US. We estimated the number of deaths attributable to coal PM2.5 from 1999 to 2020 using individual-level Medicare death records representing 650 million person-years. Exposure to coal PM2.5 was associated with 2.1 times greater mortality risk than exposure to PM2.5 from all sources. A total of 460,000 deaths were attributable to coal PM2.5, representing 25% of all PM2.5-related Medicare deaths before 2009 and 7% after 2012. Here, we quantify and visualize the contribution of individual EGUs to mortality.

Inequitable Exposures to U.S. Coal Power Plant-Related PM2.5: 22 Years and Counting

BACKGROUND

Emissions from coal power plants have decreased over recent decades due to regulations and economics affecting costs of providing electricity generated by coal vis-à-vis its alternatives. These changes have improved regional air quality, but questions remain about whether benefits have accrued equitably across population groups.

OBJECTIVES

We aimed to quantify nationwide long-term changes in exposure to particulate matter (PM) with an aerodynamic diameter ≤2.5μm (PM2.5) associated with coal power plant SO2 emissions. We linked exposure reductions with three specific actions taken at individual power plants: scrubber installations, reduced operations, and retirements. We assessed how emissions changes in different locations have influenced exposure inequities, extending previous source-specific environmental justice analyses by accounting for location-specific differences in racial/ethnic population distributions.

METHODS

We developed a data set of annual PM2.5 source impacts ("coal PM2.5") associated with SO2 emissions at each of 1,237 U.S. coal-fired power plants across 1999-2020. We linked population-weighted exposure with information about each coal unit's operational and emissions-control status. We calculate changes in both relative and absolute exposure differences across demographic groups.

RESULTS

Nationwide population-weighted coal PM2.5 declined from 1.96μg/m3 in 1999 to 0.06 μg/m3 in 2020. Between 2007 and 2010, most of the exposure reduction is attributable to SO2 scrubber installations, and after 2010 most of the decrease is attributable to retirements. Black populations in the South and North Central United States and Native American populations in the western United States were inequitably exposed early in the study period. Although inequities decreased with falling emissions, facilities in states across the North Central United States continue to inequitably expose Black populations, and Native populations are inequitably exposed to emissions from facilities in the West.

DISCUSSION

We show that air quality controls, operational adjustments, and retirements since 1999 led to reduced exposure to coal power plant related PM2.5. Reduced exposure improved equity overall, but some populations continue to be inequitably exposed to PM2.5 associated with facilities in the North Central and western United States. https://doi.org/10.1289/EHP11605.

Optimal mode of delivery in pregnancy: Individualized predictions using national vital statistics data

Child birth via Cesarean section accounts for approximately 32% of all births each year in the United States. A variety of risk factors and complications can lead caregivers and patients to plan for a Cesarean delivery in advance before onset of labor. However, a non-trivial subset of Cesarean sections (∼25%) are unplanned and occur after an initial trial of labor is attempted. Unfortunately, patients who deliver via unplanned Cesarean sections have increased maternal morbidity and mortality rates and higher rates of neonatal intensive care admissions. In an effort to develop models aimed at improving health outcomes in labor and delivery, this work seeks to explore the use of national vital statistics data to quantify the likelihood of an unplanned Cesarean section based on 22 maternal characteristics. Machine learning techniques are used to ascertain influential features, train and evaluate models, and assess accuracy against available test data. Based on cross-validation results from a large training cohort (n = 6,530,467 births), the gradient-boosted tree algorithm was identified as the best performer and was evaluated on a large test cohort (n = 10,613,877 births) for two prediction scenarios. Area under the receiver operating characteristic curves of 0.77 or higher and recall scores of 0.78 or higher were obtained and the resulting models are well calibrated. Combined with feature importance analysis to explain why certain maternal characteristics lead to a specific prediction in individual patients, the developed analysis pipeline provides additional quantitative information to aid in the decision process on whether to plan for a Cesarean section in advance, a substantially safer option among women at a high risk of unplanned Cesarean delivery during labor.

Statistical and machine learning methods for evaluating trends in air quality under changing meteorological conditions

Evaluating the influence of anthropogenic-emission changes on air quality requires accounting for the influence of meteorological variability. Statistical methods such as multiple linear regression (MLR) models with basic meteorological variables are often used to remove meteorological variability and estimate trends in measured pollutant concentrations attributable to emission changes. However, the ability of these widely used statistical approaches to correct for meteorological variability remains unknown, limiting their usefulness in the real-world policy evaluations. Here, we quantify the performance of MLR and other quantitative methods using simulations from a chemical transport model, GEOS-Chem, as a synthetic dataset. Focusing on the impacts of anthropogenic-emission changes in the US (2011 to 2017) and China (2013 to 2017) on PM_2.5 and O₃, we show that widely used regression methods do not perform well in correcting for meteorological variability and identifying long-term trends in ambient pollution related to changes in emissions. The estimation errors, characterized as the differences between meteorology-corrected trends and emission-driven trends under constant meteorology scenarios, can be reduced by 30%-42% using a random forest model that incorporates both local- and regional-scale meteorological features. We further design a correction method based on GEOS-Chem simulations with constant-emission input and quantify the degree to which anthropogenic emissions and meteorological influences are inseparable, due to their process-based interactions. We conclude by providing recommendations for evaluating the impacts of anthropogenic-emission changes on air quality using statistical approaches.

Statistical and machine learning methods for evaluating trends in air quality under changing meteorological conditions

Evaluating the influence of anthropogenic-emission changes on air quality requires accounting for the influence of meteorological variability. Statistical methods such as multiple linear regression (MLR) models with basic meteorological variables are often used to remove meteorological variability and estimate trends in measured pollutant concentrations attributable to emission changes. However, the ability of these widely used statistical approaches to correct for meteorological variability remains unknown, limiting their usefulness in the real-world policy evaluations. Here, we quantify the performance of MLR and other quantitative methods using simulations from a chemical transport model, GEOS-Chem, as a synthetic dataset. Focusing on the impacts of anthropogenic-emission changes in the US (2011 to 2017) and China (2013 to 2017) on PM_2.5 and O₃, we show that widely used regression methods do not perform well in correcting for meteorological variability and identifying long-term trends in ambient pollution related to changes in emissions. The estimation errors, characterized as the differences between meteorology-corrected trends and emission-driven trends under constant meteorology scenarios, can be reduced by 30%-42% using a random forest model that incorporates both local- and regional-scale meteorological features. We further design a correction method based on GEOS-Chem simulations with constant-emission input and quantify the degree to which anthropogenic emissions and meteorological influences are inseparable, due to their process-based interactions. We conclude by providing recommendations for evaluating the impacts of anthropogenic-emission changes on air quality using statistical approaches.

RETRACTED: Neighborhood-level variability in asthma-related emergency department visits in Central Texas

Background:

The extent to which asthma-related ED visit incidence rates vary from neighborhood to neighborhood and predictors of neighorbood-level asthma ED visit burden are not well understood.

Objective:

To describe the census tract-level spatial distribution of asthma-related emergency department visits in Central Texas and identify neighborhood-level characteristics that explain variability in neighborhood-level asthma ED visit rates.

Methods:

Conditional autoregressive models were used to examine the spatial distribution of asthma-related ED visit incidence rates across Travis County, TX census tracts and to assess the contribution of census tract characteristics to their distribution.

Results:

There were distinct patterns in ED visit incidence rates at the census tract scale, which were largely unexplained by socioeconomic or selected built environment neighborhood characteristics. Racial and ethnic composition explained 33% of the variability of ED visit incidence rates across census tracts. Spatial patterns and the census tract predictors of ED visit incidence rates differed by racial and ethnic groups.

Conclusions:

Variability in asthma ED visit incidence rates are apparent at a smaller spatial scales than previously examined. The majority of the variability in census tract-level asthma ED visit rates in Central Texas is not explained by racial and ethnic composition or other neighborhood features. Race/ethnicity-specific estimates of neighborhood ED visit rates may be useful for identifying high burden neighborhoods for specific ethnic/racial groups, which otherwise would go unrecognized.

Asthma ED visit rates may vary among neighborhoods; neighborhood-level interventions or moving to a low incidence neighborhood may be effective in reducing asthma disparities and deserve further study.

Association between county-level coal-fired power plant pollution and racial disparities in preterm births from 2000 to 2018

Coal has historically been a primary energy source in the United States. The byproducts of coal combustion, such as fine particulate matter (PM_2.5), have increasingly been associated with adverse birth outcomes. The goal of this study was to leverage the current progressive transition away from coal in the United States (U.S.) to assess whether coal PM_2.5 is associated with preterm birth rates and whether this association differs by maternal Black/White race/ethnicity. Using a novel dispersion modeling approach, we estimated PM_2.5 pollution from coal-fired power plants nationwide at the county-level during the study period (2000-2018). We also obtained county-level preterm birth rates for non-Hispanic White and non-Hispanic Black mothers. We used a generalized additive mixed model to estimate the relationship between coal PM_2.5 and preterm birth rates, overall and stratified by maternal race. We included a natural spline to allow for non-linearity in the concentration-response curve. We observed a positive non-linear relationship between coal PM_2.5 and preterm birth rate, which plateaued at higher levels of pollution. We also observed differential associations by maternal race; the association was stronger for White women, especially at higher levels of coal PM_2.5 (> 2.0 μg/m³). Our findings suggest that the transition away from coal may reduce preterm birth rates in the U.S.

Bayesian data fusion: Probabilistic sensitivity analysis for unmeasured confounding using informative priors based on secondary data

Bayesian causal inference offers a principled approach to policy evaluation of proposed interventions on mediators or time-varying exposures. Building on the Bayesian g-formula method introduced by Keil et al., we outline a general approach for the estimation of population-level causal quantities involving dynamic and stochastic treatment regimes, including regimes related to mediation estimands such as natural direct and indirect effects. We further extend this approach to propose a Bayesian data fusion (BDF), an algorithm for performing probabilistic sensitivity analysis when a confounder unmeasured in a primary data set is available in an external data source. When the relevant relationships are causally transportable between the two source populations, BDF corrects confounding bias and supports causal inference and decision-making within the main study population without sharing of the individual-level external data set. We present results from a simulation study comparing BDF to two common frequentist correction methods for unmeasured mediator-outcome confounding bias in the mediation setting. We use these methods to analyze data on the role of stage at cancer diagnosis in contributing to Black-White colorectal cancer survival disparities.

Improved asthma outcomes observed in the vicinity of coal power plant retirement, retrofit, and conversion to natural gas

Coal-fired power plants release substantial air pollution, including over 60% of U.S. sulfur dioxide (SO2) emissions in 2014. Such air pollution may exacerbate asthma however direct studies of health impacts linked to power plant air pollution are rare. Here, we take advantage of a natural experiment in Louisville, Kentucky, where one coal-fired power plant retired and converted to natural gas, and three others installed SO2 emission control systems between 2013 and 2016. Dispersion modeling indicated exposure to SO2 emissions from these power plants decreased after the energy transitions. We used several analysis strategies, including difference-in-differences, first-difference, and interrupted time-series modeling to show that the emissions control installations and plant retirements were associated with reduced asthma disease burden related to ZIP code-level hospitalizations and emergency room visits, and individual-level medication use as measured by digital medication sensors.

Are per capita carbon emissions predictable across countries?

BACKGROUND

China and other developing countries in Asia follow similar economic growth patterns described by the flying geese (FG) model, which explains the "catching-up" process of industrialization in latecomer economies. Japan, newly industrialized economies, and China have followed this path, with similar economic development trajectories. Based on the FG model, we postulated a "flying S" hypothesis stating that if a country is located within an FG region and its energy matrix is relatively constant, its per capita CO₂ emission curve will mirror that of "leading geese" countries in the same FG group.

METHOD

Historical CO₂ emissions data were obtained from literature review and national reports and were calculated using bottom-up methods. A sigmoid-shaped, non-linear mixed effect model was applied to examine ex post data with 1000 simulated predictions to construct 95% empirical bands from these fits. By multiplying by estimated population, we predicted total emissions of selected FG countries.

RESULTS

Per capita CO₂ emissions from the same FG group mirror each other, especially among second and third industrial sectors. We estimated an annual 18,252.24 million tons of CO₂ emissions (MtCO₂) (95% CI = 9458.88-23,972.88) in China and 8281.76 MtCO2 (95% CI = 2765.68-14,959.12) in India in 2030.

CONCLUSION

This study bridges the macroeconomic FG paradigm to study climate change and proposes a "flying S" hypothesis to predict greenhouse gas emissions in East Asia. By applying our theory to empirical data, we provide an alternative framework to predict CO₂ emissions in 2030 and beyond.

A global perspective on coal-fired power plants and burden of lung cancer

BACKGROUND

Exposure to ambient particulate matter generated from coal-fired power plants induces long-term health consequences. However, epidemiologic studies have not yet focused on attributing these health burdens specifically to energy consumption, impeding targeted intervention policies. We hypothesize that the generating capacity of coal-fired power plants may be associated with lung cancer incidence at the national level.

METHODS

Age- and sex-adjusted lung cancer incidence from every country with electrical plants using coal as primary energy supply were followed from 2000 to 2016. We applied a Poisson regression longitudinal model, fitted using generalized estimating equations, to estimate the association between lung cancer incidence and per capita coal capacity, adjusting for various behavioral and demographic determinants and lag periods.

RESULTS

The average coal capacity increased by 1.43 times from 16.01 gigawatts (GW) (2000~2004) to 22.82 GW (2010~2016). With 1 kW (KW) increase of coal capacity per person in a country, the relative risk of lung cancer increases by a factor of 59% (95% CI = 7.0%~ 135%) among males and 85% (95% CI = 22%~ 182%) among females. Based on the model, we estimate a total of 1.37 (range = 1.34 ~ 1.40) million standardized incident cases from lung cancer will be associated with coal-fired power plants in 2025.

CONCLUSIONS

These analyses suggest an association between lung cancer incidence and increased reliance on coal for energy generation. Such data may be helpful in addressing a key policy question about the externality costs and estimates of the global disease burden from preventable lung cancer attributable to coal-fired power plants at the national level.

Propensity scores with misclassified treatment assignment: a likelihood-based adjustment

Propensity score methods are widely used in comparative effectiveness research using claims data. In this context, the inaccuracy of procedural or billing codes in claims data frequently misclassifies patients into treatment groups, that is, the treatment assignment ($T$) is often measured with error. In the context of a validation data where treatment assignment is accurate, we show that misclassification of treatment assignment can impact three distinct stages of a propensity score analysis: (i) propensity score estimation; (ii) propensity score implementation; and (iii) outcome analysis conducted conditional on the estimated propensity score and its implementation. We examine how the error in $T$ impacts each stage in the context of three common propensity score implementations: subclassification, matching, and inverse probability of treatment weighting (IPTW). Using validation data, we propose a two-step likelihood-based approach which fully adjusts for treatment misclassification bias under subclassification. This approach relies on two common measurement error-assumptions; non-differential measurement error and transportability of the measurement error model. We use simulation studies to assess the performance of the adjustment under subclassification, and also investigate the method's performance under matching or IPTW. We apply the methods to Medicare Part A hospital claims data to estimate the effect of resection versus biopsy on 1-year mortality among $10\,284$ Medicare beneficiaries diagnosed with brain tumors. The ICD9 billing codes from Medicare Part A inaccurately reflect surgical treatment, but SEER-Medicare validation data are available with more accurate information.

Guided Bayesian imputation to adjust for confounding when combining heterogeneous data sources in comparative effectiveness research

In comparative effectiveness research, we are often interested in the estimation of an average causal effect from large observational data (the main study). Often this data does not measure all the necessary confounders. In many occasions, an extensive set of additional covariates is measured for a smaller and non-representative population (the validation study). In this setting, standard approaches for missing data imputation might not be adequate due to the large number of missing covariates in the main data relative to the smaller sample size of the validation data. We propose a Bayesian approach to estimate the average causal effect in the main study that borrows information from the validation study to improve confounding adjustment. Our approach combines ideas of Bayesian model averaging, confounder selection, and missing data imputation into a single framework. It allows for different treatment effects in the main study and in the validation study, and propagates the uncertainty due to the missing data imputation and confounder selection when estimating the average causal effect (ACE) in the main study. We compare our method to several existing approaches via simulation. We apply our method to a study examining the effect of surgical resection on survival among 10 396 Medicare beneficiaries with a brain tumor when additional covariate information is available on 2220 patients in SEER-Medicare. We find that the estimated ACE decreases by 30% when incorporating additional information from SEER-Medicare.

A Global Perspective on Sulfur Oxide Controls in Coal-Fired Power Plants and Cardiovascular Disease

Sulfur oxides (SO_x), particularly SO₂ emitted by coal-fired power plants, produce long-term risks for cardiovascular disease (CVD). We estimated the relative risks of CVD and ischemic heart disease (IHD) attributable to SO_x emission globally. National SO_x reduction achieved by emissions control systems was defined as the average SO_x reduction percentage weighted by generating capacities of individual plants in a country. We analyzed the relative risk of CVD incidence associated with national SO_x reduction for 13,581 coal-fired power-generating units in 79 countries. A 10% decrease in SO_x emission was associated with 0.28% (males; 95%CI = −0.39%~0.95%) and 1.69% (females; 95%CI = 0.99%~2.38%) lower CVD risk. The effects on IHD were > 2 times stronger among males than females (2.78%, 95%CI = 1.99%~3.57% vs. 1.18%, 95%CI = 0.19%~2.17%). Further, 1.43% (males) and 8.00% (females) of CVD cases were attributable to suboptimal SOx reduction. Thus, enhancing regulations on SO_x emission control represents a target for national and international intervention to prevent CVD.

Best Practices for Gauging Evidence of Causality in Air Pollution Epidemiology

The contentious political climate surrounding air pollution regulations has brought some researchers and policy-makers to argue that evidence of causality is necessary before implementing more stringent regulations. Recently, investigators in an increasing number of air pollution studies have purported to have used "causal analysis," generating the impression that studies not explicitly labeled as such are merely "associational" and therefore less rigorous. Using 3 prominent air pollution studies as examples, we review good practices for how to critically evaluate the extent to which an air pollution study provides evidence of causality. We argue that evidence of causality should be gauged by a critical evaluation of design decisions such as 1) what actions or exposure levels are being compared, 2) whether an adequate comparison group was constructed, and 3) how closely these design decisions approximate an idealized randomized study. We argue that air pollution studies that are more scientifically rigorous in terms of the decisions made to approximate a randomized experiment are more likely to provide evidence of causality and should be prioritized among the body of evidence for regulatory review accordingly. Our considerations, although presented in the context of air pollution epidemiology, can be broadly applied to other fields of epidemiology.

Comparative effectiveness of radiotherapy with vs. without temozolomide in older patients with glioblastoma

It is unknown whether the addition of temozolomide (TMZ) to radiotherapy (RT) is associated with improved overall survival (OS) among older glioblastoma patients. We performed a retrospective cohort SEER-Medicare analysis of 1652 patients aged ≥65 years with glioblastoma who received ≥10 fractions of RT from 2005 to 2009, or from 1995 to 1999 before TMZ was available. Three cohorts were assembled based on diagnosis year and treatment initiated within 60 days of diagnosis: (1) 2005-2009 and TMZ/RT, (2) 2005-2009 and RT only, or (3) 1995-1999 and RT only. Associations with OS were estimated using Cox proportional hazards models and propensity score analyses; OS was calculated starting 60 days after diagnosis. Pre-specified sensitivity analyses were performed among patients who received long-course RT (≥27 fractions). Median survival estimates were 7.4 (IQR, 3.3-14.7) months for TMZ/RT, 5.9 (IQR, 2.6-12.1) months for RT alone in 2005-2009, and 5.6 (IQR, 2.7-9.6) months for RT alone in 1995-1999. OS at 2 years was 10.1 % for TMZ/RT, 7.1 % for RT in 2005-2009, and 4.7 % for RT in 1995-1999. Adjusted models suggested decreased mortality risk for TMZ/RT compared to RT in 2005-2009 (AHR, 0.86; 95 % CI, 0.76-0.98) and RT in 1995-1999 (AHR, 0.71; 95 % CI, 0.57-0.90). Among patients from 2005 to 2009 who received long-course RT, however, the addition of TMZ did not significantly improve survival (AHR, 0.91; 95 % CI, 0.80-1.04). In summary, among a large cohort of older glioblastoma patients treated in a real-world setting, the addition of TMZ to RT was associated with a small survival gain.

Developmental validation of a point-of-care, salivary α-amylase biosensor

The translation of salivary alpha-amylase (sAA) to the ambulatory assessment of stress hinges on the development of technologies capable of speedy and accurate reporting of sAA levels. Here, we describe the developmental validation and usability testing of a point-of-care, colorimetric, sAA biosensor. A disposable test strip allows for streamlined sample collection and a corresponding hand-held reader with integrated analytic capabilities permits rapid analysis and reporting of sAA levels. Bioanalytical validation utilizing saliva samples from 20 normal subjects indicates that, within the biosensor's linear range (10-230 U/ml), its accuracy (R(2)=0.989), precision (CV<9%), and measurement repeatability (range -3.1% to +3.1%) approach more elaborate laboratory-based, clinical analyzers. The truncated sampling-reporting cycle (<1 min) and the excellent performance characteristics of the biosensor has the potential to take sAA analysis out of the realm of dedicated, centralized laboratories and facilitate future sAA biomarker qualification studies.

Willingness of facial injury patients to change causal substance using behaviors

Many injuries due to interpersonal violence among patients presenting to urban trauma centers for treatment are preventable, with alcohol and illicit drug use presenting as common antecedent risk factors. However, many patients with such problems do not seek treatment. Substance use patients were surveyed to determine how many recognized they had a problem and whether they had previously received treatment for substance use problems. Almost 60% of the patients treated for a facial injury screened for problem alcohol use, and slightly more than 25% screened for problem drug use. Only approximately one third of patients indicated any movement towards dealing with these problems and of these, only 20% had actually sought treatment. Employment had an effect on treatment seeking, with fewer employed patients seeking help. Utilizing the critical window of opportunity for emergency department (ED) personnel to make referrals may have an impact on treatment seeking for problem level substance use.