Search

BackgroundTime-resolved quantification of physical activity can contribute to both personalized medicine and epidemiological research studies, for example, managing and identifying triggers of asthma exacerbations. A growing number of reportedly accurate machine learning algorithms for human activity recognition (HAR) have been developed using data from wearable devices (eg, smartwatch and smartphone). However, many HAR algorithms depend on fixed-size sampling windows that may poorly adapt to real-world conditions in which activity bouts are of unequal duration. A small sliding window can produce noisy predictions under stable conditions, whereas a large sliding window may miss brief bursts of intense activity. ObjectiveWe aimed to create an HAR framework adapted to variable duration activity bouts by (1) detecting the change points of activity bouts in a multivariate time series and (2) predicting activity for each homogeneous window defined by these change points. MethodsWe applied standard fixed-width sliding windows (4-6 different sizes) or greedy Gaussian segmentation (GGS) to identify break points in filtered triaxial accelerometer and gyroscope data. After standard feature engineering, we applied an Xgboost model to predict physical activity within each window and then converted windowed predictions to instantaneous predictions to facilitate comparison across segmentation methods. We applied these methods in 2 datasets: the human activity recognition using smartphones (HARuS) dataset where a total of 30 adults performed activities of approximately equal duration (approximately 20 seconds each) while wearing a waist-worn smartphone, and the Biomedical REAl-Time Health Evaluation for Pediatric Asthma (BREATHE) dataset where a total of 14 children performed 6 activities for approximately 10 min each while wearing a smartwatch. To mimic a real-world scenario, we generated artificial unequal activity bout durations in the BREATHE data by randomly subdividing each activity bout into 10 segments and randomly concatenating the 60 activity bouts. Each dataset was divided into ~90% training and ~10% holdout testing. ResultsIn the HARuS data, GGS produced the least noisy predictions of 6 physical activities and had the second highest accuracy rate of 91.06% (the highest accuracy rate was 91.79% for the sliding window of size 0.8 second). In the BREATHE data, GGS again produced the least noisy predictions and had the highest accuracy rate of 79.4% of predictions for 6 physical activities. ConclusionsIn a scenario with variable duration activity bouts, GGS multivariate segmentation produced smart-sized windows with more stable predictions and a higher accuracy rate than traditional fixed-size sliding window approaches. Overall, accuracy was good in both datasets but, as expected, it was slightly lower in the more real-world study using wrist-worn smartwatches in children (BREATHE) than in the more tightly controlled study using waist-worn smartphones in adults (HARuS). We implemented GGS in an offline setting, but it could be adapted for real-time prediction with streaming data.

The Environmental Influences on Child Health Outcomes (ECHO)-Wide Cohort Study (EWC), a collaborative research design comprising 69 cohorts in 31 consortia, was funded by the National Institutes of Health (NIH) in 2016 to improve children's health in the United States. The EWC harmonizes extant data and collects new data using a standardized protocol, the ECHO-Wide Cohort Data Collection Protocol (EWCP). EWCP visits occur at least once per life stage, but the frequency and timing of the visits vary across cohorts. As of March 4, 2022, the EWC cohorts contributed data from 60,553 children and consented 29,622 children for new EWCP data and biospecimen collection. The median (interquartile range) age of EWCP-enrolled children was 7.5 years (3.7-11.1). Surveys, interviews, standardized examinations, laboratory analyses, and medical record abstraction are used to obtain information in 5 main outcome areas: pre-, peri-, and postnatal outcomes; neurodevelopment; obesity; airways; and positive health. Exposures include factors at the level of place (e.g., air pollution, neighborhood socioeconomic status), family (e.g., parental mental health), and individuals (e.g., diet, genomics).

Tools for assessing multiple exposures across several domains (e.g., physical, chemical, and social) are of growing importance in social and environmental epidemiology because of their value in uncovering disparities and their impact on health outcomes. Here we describe work done within the Environmental influences on Child Health Outcomes (ECHO)-wide Cohort Study to build a combined exposure index. Our index considered both environmental hazards and social stressors simultaneously with national coverage for a 10-year period. Our goal was to build this index and demonstrate its utility for assessing differences in exposure for pregnancies enrolled in the ECHO-wide Cohort Study. Our unitless combined exposure index, which collapses census-tract level data into a single relative measure of exposure ranging from 0-1 (where higher values indicate higher exposure to hazards), includes indicators for major air pollutants and air toxics, features of the built environment, traffic exposures, and social determinants of health (e.g., lower educational attainment) drawn from existing data sources. We observed temporal and geographic variations in index values, with exposures being highest among participants living in the West and Northeast regions. Pregnant people who identified as Black or Hispanic (of any race) were at higher risk of living in a "high" exposure census tract (defined as an index value above 0.5) relative to those who identified as White or non-Hispanic. Index values were also higher for pregnant people with lower educational attainment. Several recommendations follow from our work, including that environmental and social stressor datasets with higher spatial and temporal resolutions are needed to ensure index-based tools fully capture the total environmental context.

Limited studies examine how prenatal environmental and social exposures jointly impact perinatal health. Here we investigated relationships between a neighborhood-level combined exposure (CE) index assessed during pregnancy and perinatal outcomes, including birthweight, gestational age, and preterm birth. Across all participants, higher CE index scores were associated with small decreases in birthweight and gestational age. We also observed effect modification by race; infants born to Black pregnant people had a greater risk of preterm birth for higher CE values compared to White infants. Overall, our results suggest that neighborhood social and environmental exposures have a small but measurable joint effect on neonatal indicators of health.

Background: Exposure to air pollution is associated with acute pediatric asthma exacerbations, including reduced lung function, rescue medication usage, and increased symptoms; however, most studies are limited in investigating longitudinal changes in these acute effects. This study aims to investigate the effects of daily air pollution exposure on acute pediatric asthma exacerbation risk using a repeated-measures design. Methods: We conducted a panel study of 40 children aged 8−16 years with moderate-to-severe asthma. We deployed the Biomedical REAI-Time Health Evaluation (BREATHE) Kit developed in the Los Angeles PRISMS Center to continuously monitor personal exposure to particulate matter of aerodynamic diameter < 2.5 µm (PM2.5), relative humidity and temperature, geolocation (GPS), and asthma outcomes including lung function, medication use, and symptoms for 14 days. Hourly ambient (PM2.5, nitrogen dioxide (NO2), ozone (O3)) and traffic-related (nitrogen oxides (NOx) and PM2.5) air pollution exposures were modeled based on location. We used mixed-effects models to examine the association of same day and lagged (up to 2 days) exposures with daily changes in % predicted forced expiratory volume in 1 s (FEV1) and % predicted peak expiratory flow (PEF), count of rescue inhaler puffs, and symptoms. Results: Participants were on average 12.0 years old (range: 8.4−16.8) with mean (SD) morning %predicted FEV1 of 67.9% (17.3%) and PEF of 69.1% (18.4%) and 1.4 (3.5) puffs per day of rescue inhaler use. Participants reported chest tightness, wheeze, trouble breathing, and cough symptoms on 36.4%, 17.5%, 32.3%, and 42.9%, respectively (n = 217 person-days). One SD increase in previous day O3 exposure was associated with reduced morning (beta [95% CI]: −4.11 [−6.86, −1.36]), evening (−2.65 [−5.19, −0.10]) and daily average %predicted FEV1 (−3.45 [−6.42, −0.47]). Daily (lag 0) exposure to traffic-related PM2.5 exposure was associated with reduced morning %predicted PEF (−3.97 [−7.69, −0.26]) and greater odds of “feeling scared of trouble breathing” symptom (odds ratio [95% CI]: 1.83 [1.03, 3.24]). Exposure to ambient O3, NOx, and NO was significantly associated with increased rescue inhaler use (rate ratio [95% CI]: O3 1.52 [1.02, 2.27], NOx 1.61 [1.23, 2.11], NO 1.80 [1.37, 2.35]). Conclusions: We found significant associations of air pollution exposure with lung function, rescue inhaler use, and “feeling scared of trouble breathing.” Our study demonstrates the potential of informatics and wearable sensor technologies at collecting highly resolved, contextual, and personal exposure data for understanding acute pediatric asthma triggers.

IntroductionEstimating PM2.5 concentrations and their prediction uncertainties at a high spatiotemporal resolution is important for air pollution health effect studies. This is particularly challenging for California, which has high variability in natural (e.g, wildfires, dust) and anthropogenic emissions, meteorology, topography (e.g. desert surfaces, mountains, snow cover) and land use.MethodsUsing ensemble-based deep learning with big data fused from multiple sources we developed a PM2.5 prediction model with uncertainty estimates at a high spatial (1 km × 1 km) and temporal (weekly) resolution for a 10-year time span (2008-2017). We leveraged autoencoder-based full residual deep networks to model complex nonlinear interrelationships among PM2.5 emission, transport and dispersion factors and other influential features. These included remote sensing data (MAIAC aerosol optical depth (AOD), normalized difference vegetation index, impervious surface), MERRA-2 GMI Replay Simulation (M2GMI) output, wildfire smoke plume dispersion, meteorology, land cover, traffic, elevation, and spatiotemporal trends (geo-coordinates, temporal basis functions, time index). As one of the primary predictors of interest with substantial missing data in California related to bright surfaces, cloud cover and other known interferences, missing MAIAC AOD observations were imputed and adjusted for relative humidity and vertical distribution. Wildfire smoke contribution to PM2.5 was also calculated through HYSPLIT dispersion modeling of smoke emissions derived from MODIS fire radiative power using the Fire Energetics and Emissions Research version 1.0 model.ResultsEnsemble deep learning to predict PM2.5 achieved an overall mean training RMSE of 1.54 μg/m3 (R2: 0.94) and test RMSE of 2.29 μg/m3 (R2: 0.87). The top predictors included M2GMI carbon monoxide mixing ratio in the bottom layer, temporal basis functions, spatial location, air temperature, MAIAC AOD, and PM2.5 sea salt mass concentration. In an independent test using three long-term AQS sites and one short-term non-AQS site, our model achieved a high correlation (>0.8) and a low RMSE (

ObjectiveTo describe a configurable mobile health (mHealth) framework for integration of physiologic and environmental sensors to be used in studies focusing on the domain of pediatric asthma.Materials and methodsThe Biomedical REAl-Time Health Evaluation (BREATHE) platform connects different sensors and data streams, contextualizing an individual's symptoms and daily activities over time to understand pediatric asthma's presentation and its management. A smartwatch/smartphone combination serves as a hub for personal/wearable sensing devices collecting data on health (eg, heart rate, spirometry, medications), motion, and personal exposures (eg, particulate matter, ozone); securely transmitting information to BREATHE's servers; and interacting with the user (eg, ecological momentary assessments). Server-side integration of electronic health record data and spatiotemporally correlated information (eg, weather, traffic) elaborates on these observations. An initial panel study involving pediatric asthma patients was conducted to assess BREATHE.ResultsTwenty subjects were enrolled, during which BREATHE accrued seven consecutive days of continuous data per individual. The data were used to confirm knowledge about asthma (use of controller inhalers, time-activity behaviors, personal air pollution exposure), and additional analyses provided insights into within-day associations of environmental triggers and asthma exacerbations. Exit surveys focusing on mHealth usability, while positive, noted several translational challenges.DiscussionBased on these promising results, a longitudinal panel study to evaluate individual microenvironments and exposures is ongoing. Lessons learned thus far reflect the need to address various usability aspects, including convenience and ongoing engagement.ConclusionBREATHE enables multi-sensor mHealth studies, capturing new types of information alongside an evolving understanding of personal exposomes.

Spatiotemporal air pollution models are increasingly being used to estimate health effects in epidemiological studies. Although such exposure prediction models typically result in improved spatial and temporal resolution of air pollution predictions, they remain subject to shared measurement error, a type of measurement error common in spatiotemporal exposure models which occurs when measurement error is not independent of exposures. A fundamental challenge of exposure measurement error in air pollution assessment is the strong correlation and sometimes identical (shared) error of exposure estimates across geographic space and time. When exposure estimates with shared measurement error are used to estimate health risk in epidemiological analyses, complex errors are potentially introduced, resulting in biased epidemiological conclusions. We demonstrate the influence of using a three-stage spatiotemporal exposure prediction model and introduce formal methods of shared, multiplicative measurement error (SMME) correction of epidemiological health risk estimates. Using our three-stage, ensemble learning based nitrogen oxides (NOx) exposure prediction model, we quantified SMME. We conducted an epidemiological analysis of wheeze risk in relation to NOx exposure among school-aged children. To demonstrate the incremental influence of exposure modeling stage, we iteratively estimated the health risk using assigned exposure predictions from each stage of the NOx model. We then determined the impact of SMME on the variance of the health risk estimates under various scenarios. Depending on the stage of the spatiotemporal exposure model used, we found that wheeze odds ratio ranged from 1.16 to 1.28 for an interquartile range increase in NOx. With each additional stage of exposure modeling, the health effect estimate moved further away from the null (OR=1). When corrected for observed SMME, the health effects confidence intervals slightly lengthened, but our epidemiological conclusions were not altered. When the variance estimate was corrected for the potential "worst case scenario" of SMME, the standard error further increased, having a meaningful influence on epidemiological conclusions. Our framework can be expanded and used to understand the implications of using exposure predictions subject to shared measurement error in future health investigations.

Aerosols have adverse health effects and play a significant role in the climate as well. The Multiangle Implementation of Atmospheric Correction (MAIAC) provides Aerosol Optical Depth (AOD) at high temporal (daily) and spatial (1 km) resolution, making it particularly useful to infer and characterize spatiotemporal variability of aerosols at a fine spatial scale for exposure assessment and health studies. However, clouds and conditions of high surface reflectance result in a significant proportion of missing MAIAC AOD. To fill these gaps, we present an imputation approach using deep learning with downscaling. Using a baseline autoencoder, we leverage residual connections in deep neural networks to boost learning and parameter sharing to reduce overfitting, and conduct bagging to reduce error variance in the imputations. Downscaled through a similar auto-encoder based deep residual network, Modern-Era Retrospective analysis for Research and Applications Version 2 (MERRA-2) GMI Replay Simulation (M2GMI) data were introduced to the network as an important gap-filling feature that varies in space to be used for missingness imputations. Imputing weekly MAIAC AOD from 2000 to 2016 over California, a state with considerable geographic heterogeneity, our full (non-full) residual network achieved mean R2 = 0.94 (0.86) [RMSE = 0.007 (0.01)] in an independent test, showing considerably better performance than a regular neural network or non-linear generalized additive model (mean R2 = 0.78-0.81; mean RMSE = 0.013-0.015). The adjusted imputed as well as combined imputed and observed MAIAC AOD showed strong correlation with Aerosol Robotic Network (AERONET) AOD (R = 0.83; R2 = 0.69, RMSE = 0.04). Our results show that we can generate reliable imputations of missing AOD through a deep learning approach, having important downstream air quality modeling applications.

BackgroundAccurate estimation of nitrogen dioxide (NO2) and nitrogen oxide (NOx) concentrations at high spatiotemporal resolutions is crucial for improving evaluation of their health effects, particularly with respect to short-term exposures and acute health outcomes. For estimation over large regions like California, high spatial density field campaign measurements can be combined with more sparse routine monitoring network measurements to capture spatiotemporal variability of NO2 and NOx concentrations. However, monitors in spatially dense field sampling are often highly clustered and their uneven distribution creates a challenge for such combined use. Furthermore, heterogeneities due to seasonal patterns of meteorology and source mixtures between sub-regions (e.g. southern vs. northern California) need to be addressed.ObjectivesIn this study, we aim to develop highly accurate and adaptive machine learning models to predict high-resolution NO2 and NOx concentrations over large geographic regions using measurements from different sources that contain samples with heterogeneous spatiotemporal distributions and clustering patterns.MethodsWe used a comprehensive Kruskal-K-means method to cluster the measurement samples from multiple heterogeneous sources. Spatiotemporal cluster-based bootstrap aggregating (bagging) of the base mixed-effects models was then applied, leveraging the clusters to obtain balanced and less correlated training samples for less bias and improvement in generalization. Further, we used the machine learning technique of grid search to find the optimal interaction of temporal basis functions and the scale of spatial effects, which, together with spatiotemporal covariates, adequately captured spatiotemporal variability in NO2 and NOx at the state and local levels.ResultsWe found an optimal combination of four temporal basis functions and 200 m scale spatial effects for the base mixed-effects models. With the cluster-based bagging of the base models, we obtained robust predictions with an ensemble cross validation R2 of 0.88 for both NO2 and NOx [RMSE (RMSEIQR): 3.62 ppb (0.28) and 9.63 ppb (0.37) respectively]. In independent tests of random sampling, our models achieved similarly strong performance (R2 of 0.87-0.90; RMSE of 3.97-9.69 ppb; RMSEIQR of 0.21-0.27), illustrating minimal over-fitting.ConclusionsOur approach has important implications for fusing data from highly clustered and heterogeneous measurement samples from multiple data sources to produce highly accurate concentration estimates of air pollutants such as NO2 and NOx at high resolution over a large region.

BackgroundIncreasingly ensemble learning-based spatiotemporal models are being used to estimate residential air pollution exposures in epidemiological studies. While these machine learning models typically have improved performance, they suffer from exposure measurement error that is inherent in all models. Our objective is to develop a framework to formally assess shared, multiplicative measurement error (SMME) in our previously published three-stage, ensemble learning-based nitrogen oxides (NOx) model to identify its spatial and temporal patterns and predictors.MethodsBy treating the ensembles as an external dosimetry system, we quantified shared and unshared, multiplicative and additive (SUMA) measurement error components in our exposure model. We used generalized additive models (GAMs) with a smooth term for location to identify geographic locations with significantly elevated SMME and explain their spatial and temporal determinants.ResultsWe found evidence of significant shared and unshared multiplicative error (p

Rationale: Few studies have examined the effects of long-term childhood air pollution exposure on adult respiratory health, including whether childhood respiratory effects underlie this relation. Objectives: To evaluate associations between childhood air pollution exposure and self-reported adult bronchitic symptoms while considering child respiratory health in the Southern California Children's Health Study. Methods: Exposures to nitrogen dioxide (NO2), ozone, and particulate matter <2.5 μm and <10 μm in diameter (PM10) assessed using inverse-distance-squared spatial interpolation based on childhood (birth to age 17 yr) residential histories. Bronchitic symptoms (bronchitis, cough, or phlegm in the past 12 mo) were ascertained via a questionnaire in adulthood. Associations between mean air pollution exposure across childhood and self-reported adult bronchitic symptoms were estimated using logistic regression. We further adjusted for childhood bronchitic symptoms and asthma to understand whether associations operated beyond childhood respiratory health impacts. Effect modification was assessed for family history of asthma, childhood asthma, and adult allergies. Measurements and Main Results: A total of 1,308 participants were included (mostly non-Hispanic White [56%] or Hispanic [32%]). At adult assessment (mean age, 32.0 yr; standard deviation [SD], 4.7), 25% reported bronchitic symptoms. Adult bronchitic symptoms were associated with NO2 and PM10 childhood exposures. Odds ratios per 1-SD increase were 1.69 (95% confidence interval, 1.14–2.49) for NO2 (SD, 11.1 ppb) and 1.51 (95% confidence interval, 1.00–2.27) for PM10 (SD, 14.2 μg/m3). Adjusting for childhood bronchitic symptoms or asthma produced similar results. NO2 and PM10 associations were modified by childhood asthma, with greater associations among asthmatic individuals. Conclusions: Childhood NO2 and PM10 exposures were associated with adult bronchitic symptoms. Associations were not explained by childhood respiratory health impacts; however, participants with childhood asthma had stronger associations. [ABSTRACT FROM AUTHOR]

Spatiotemporal models to estimate ambient exposures at high spatiotemporal resolutions are crucial in large-scale air pollution epidemiological studies that follow participants over extended periods. Previous models typically rely on central-site monitoring data and/or covered short periods, limiting their applications to long-term cohort studies. Here we developed a spatiotemporal model that can reliably predict nitrogen oxide concentrations with a high spatiotemporal resolution over a long time span (>20 years). Leveraging the spatially extensive highly clustered exposure data from short-term measurement campaigns across 1-2 years and long-term central site monitoring in 1992-2013, we developed an integrated mixed-effect model with uncertainty estimates. Our statistical model incorporated nonlinear and spatial effects to reduce bias. Identified important predictors included temporal basis predictors, traffic indicators, population density, and subcounty-level mean pollutant concentrations. Substantial spatial autocorrelation (11-13%) was observed between neighboring communities. Ensemble learning and constrained optimization were used to enhance reliability of estimation over a large metropolitan area and a long period. The ensemble predictions of biweekly concentrations resulted in an R2 of 0.85 (RMSE: 4.7 ppb) for NO2 and 0.86 (RMSE: 13.4 ppb) for NOx. Ensemble learning and constrained optimization generated stable time series, which notably improved the results compared with those from initial mixed-effects models.

BACKGROUND: Air pollution has been associated with gestational hypertension (GH) and preeclampsia, but susceptible windows of exposure and potential vulnerability by comorbidities, such as prenatal depression, remain unclear. METHODS: We ascertained GH and preeclampsia cases in a prospective pregnancy cohort in Los Angeles, CA. Daily levels of ambient particulate matters (with a diameter of ≤10 μm [PM10] or ≤2.5 μm [PM2.5]), nitrogen dioxide, and ozone were averaged for each week from 12 weeks preconception to 20 gestational weeks. We used distributed lag models to identify susceptible exposure windows, adjusting for potential confounders. Analyses were additionally stratified by probable prenatal depression to explore population vulnerability. RESULTS: Among 619 participants, 60 developed preeclampsia and 42 developed GH. We identified a susceptible window for exposure to PM2.5 from 1 week preconception to 11 weeks postconception: higher exposure (5 µg/m3) within this window was associated with an average of 8% (95% CI, 1%–15%) higher risk of GH. Among participants with probable prenatal depression (n=179; 32%), overlapping sensitive windows were observed for all pollutants from 8 weeks before to 10 weeks postconception with increased risk of GH (PM2.5, 16% [95% CI, 3%–31%]; PM10, 39% [95% CI, 13%–72%]; nitrogen dioxide, 65% [95% CI, 17%–134%]; and ozone, 45% [95% CI, 9%–93%]), while the associations were close to null among those without prenatal depression. Air pollutants were not associated with preeclampsia in any analyses. CONCLUSIONS: We identified periconception through early pregnancy as a susceptible window of air pollution exposure with an increased risk of GH. Prenatal depression increases vulnerability to air pollution exposure and GH. [ABSTRACT FROM AUTHOR]

Introduction Secondhand smoke (SHS) exposure during pregnancy is linked to adverse birth outcomes, such as low birth weight and preterm birth. While questionnaires are commonly used to assess SHS exposure, their ability to capture true exposure can vary, making it difficult for researchers to harmonize SHS measures. This study aimed to compare self-reported SHS exposure with measurements of airborne SHS in personal samples of pregnant women. Methods SHS was measured on 48-hour integrated personal PM2.5 Teflon filters collected from 204 pregnant women, and self-reported SHS exposure measures were obtained via questionnaires. Descriptive statistics were calculated for airborne SHS measures, and analysis of variance tests assessed group differences in airborne SHS concentrations by self-reported SHS exposure. Results Participants were 81% Hispanic, with a mean (standard deviation [SD]) age of 28.2 (6.0) years. Geometric mean (SD) personal airborne SHS concentrations were 0.14 (9.41) µg/m3. Participants reporting lower education have significantly higher airborne SHS exposure (p  = .015). Mean airborne SHS concentrations were greater in those reporting longer duration with windows open in the home. There was no association between airborne SHS and self-reported SHS exposure; however, asking about the number of smokers nearby in the 48-hour monitoring period was most correlated with measured airborne SHS (Two + smokers: 0.30 µg/m3 vs. One: 0.12 µg/m3 and Zero: 0.15 µg/m3; p  = .230). Conclusions Self-reported SHS exposure was not associated with measured airborne SHS in personal PM2.5 samples. This suggests exposure misclassification using SHS questionnaires and the need for harmonized and validated questions to characterize this exposure in health studies. Implications This study adds to the growing body of evidence that measurement error is a major concern in pregnancy research, particularly in studies that rely on self-report questionnaires to measure SHS exposure. The study introduces an alternative method of SHS exposure assessment using objective optical measurements, which can help improve the accuracy of exposure assessment. The findings emphasize the importance of using harmonized and validated SHS questionnaires in pregnancy health research to avoid biased effect estimates. This study can inform future research, practice, and policy development to reduce SHS exposure and its adverse health effects. [ABSTRACT FROM AUTHOR]