Environmental Exposure Data Research Articles

Risk Assessment of Phthalate Esters in Baiyangdian Lake and Typical Rivers in China

Phthalate esters (PAEs) are frequently tracked in water environments worldwide. As a typical class of endocrine disruptor chemicals (EDCs), PAEs posed adverse effects on aquatic organisms at low concentration. Thus, they have attracted wide attention in recent years. In the present study, the concentrations of seven typical PAEs from 30 sampling sites in Baiyangdian Lake were measured, and the environmental exposure data of PAEs were gathered in typical rivers in China. Then, based on the aquatic life criteria (ALCs) derived from the reproductive toxicity data of aquatic organisms, two risk assessment methods, including hazard quotient (HQ) and probabilistic ecological risk assessment (PERA), were adopted to evaluate the ecological risks of PAEs in water. The sediment quality criteria (SQCs) of DEHP, DBP, BBP, DIBP and DEP were deduced based on the equilibrium partitioning method. Combined with the gathered environmental exposure data of seven PAEs in sediments from typical rivers in China, the ecological risk assessments of five PAEs in sediment were conducted only by the HQ method. The results of ecological risk assessment showed that in terms of water, DBP and DIBP posed low risk, while the risk of DEHP in Baiyangdian Lake cannot be ignored and should receive attention. In typical rivers in China, BBP and DEP posed no risk, while DIBP and DBP posed potential risk. Meanwhile, DEHP posed a high ecological risk. As far as sediment is concerned, DBP posed a high risk in some typical rivers in China, and the other rivers had medium risk. DEHP posed a high risk only in a few rivers and low to medium risk in others. This study provides an important reference for the protection of aquatic organisms and the risk management of PAEs in China.

Multi-omic approach associates blood methylome with bronchodilator drug response in pediatric asthma

Albuterol is the drug most widely used as asthma treatment among African Americans despite having a lower bronchodilator drug response (BDR) than other populations. Although BDR is affected by gene and environmental factors, the influence of DNA methylation is unknown. This study aimed to identify epigenetic markers in whole blood associated with BDR, study their functional consequences by multi-omic integration, and assess their clinical applicability in admixed populations with a high asthma burden. We studied 414 children and young adults (8-21 years old) with asthma in a discovery and replication design. We performed an epigenome-wide association study on 221 African Americans and replicated the results on 193 Latinos. Functional consequences were assessed by integrating epigenomics with genomics, transcriptomics, and environmental exposure data. Machine learning was used to develop a panel of epigenetic markers to classify treatment response. We identified 5 differentially methylated regions and 2 CpGs genome-wide significantly associated with BDR in African Americans located in FGL2 (cg08241295, P= 6.8×10-9) and DNASE2 (cg15341340, P= 7.8×10-8), which were regulated by genetic variation and/or associated with gene expression of nearby genes (false discovery rate< 0.05). The CpG cg15341340 was replicated in Latinos (P= 3.5×10-3). Moreover, a panel of 70 CpGs showed good classification for those with response and nonresponse to albuterol therapy in African American and Latino children (area under the receiver operating characteristic curve for training, 0.99; for validation, 0.70-0.71). The DNA methylation model showed similar discrimination as clinical predictors (P> .05). We report novel associations of epigenetic markers with BDR in pediatric asthma and demonstrate for the first time the applicability of pharmacoepigenetics in precision medicineof respiratory diseases.

Improved inference for doubly robust estimators of heterogeneous treatment effects.

We propose a doubly robust approach to characterizing treatment effect heterogeneity in observational studies. We develop a frequentist inferential procedure that utilizes posterior distributions for both the propensity score and outcome regression models to provide valid inference on the conditional average treatment effect even when high-dimensional or nonparametric models are used. We show that our approach leads to conservative inference in finite samples or under model misspecification and provides a consistent variance estimator when both models are correctly specified. In simulations, we illustrate the utility of these results in difficult settings such as high-dimensional covariate spaces or highly flexible models for the propensity score and outcome regression. Lastly, we analyze environmental exposure data from NHANES to identify how the effects of these exposures vary by subject-levelcharacteristics.

Evaluating robustness of a generalized linear model when applied to electronic health record data accessed using an Open API.

The Integrated Clinical and Environmental Exposures Service (ICEES) provides open regulatory-compliant access to clinical data, including electronic health record data, that have been integrated with environmental exposures data. While ICEES has been validated in the context of an asthma use case and several other use cases, the regulatory constraints on the ICEES open application programming interface (OpenAPI) result in data loss when using the service for multivariate analysis. In this study, we investigated the robustness of the ICEES OpenAPI through a comparative analysis, in which we applied a generalized linear model (GLM) to the OpenAPI data and the constraint-free source data to examine factors predictive of asthma exacerbations. Consistent with previous studies, we found that the main predictors identified by both analyses were sex, prednisone, race, obesity, and airborne particulate exposure. Comparison of GLM model fit revealed that data loss impacts model quality, but only with select interaction terms. We conclude that the ICEES OpenAPI supports multivariate analysis, albeit with potential data loss that users should be aware of.

Microplastics Scoping Review of Environmental and Human Exposure Data

Scientific studies of microplastics have expanded since 2015, propelling the topic to the forefront of scientific inquiry. Microplastics are ubiquitous in the environment and pose a potential risk to human health. The purpose of this review is to organize microplastics literature into areas of scientific research, summarize the state of the literature and identify the current data gaps in knowledge to promote a better understanding of human exposure to microplastics and their potential health effects. We searched for published literature from eight databases. Our search focused on three categories: (1) microplastics in the environment, (2) adsorption and absorption of chemicals to microplastics, and (3) human exposure to microplastics in the environment. We screened all abstracts to select articles that focused on microplastics. We then screened the remaining articles using criteria outlined in a questionnaire to identify and assign articles to the three scoping review categories. After screening abstracts, we selected 1186 articles (19%) to thoroughly assess their appropriateness for inclusion in the final review. Of the 1186 articles, 903 (76.1%) belonged to the environmental category, 268 (22.6%) to the adsorption and absorption category, and 16 (1.3%) to the human exposure category. Water was the most frequently studied environmental medium (440 articles). Our assessment resulted in 572 articles selected for the final review. Of the 572 publications, 268 (48.2%) included a geographic component and 110 (19.2%) were the product of literature reviews. We also show that relatively few publications have investigated human health effects associated with exposures to microplastics.

Abstract C005: Social determinants and environmental exposures and their localized contribution to risk of cancer mortality

Abstract When considering environmental exposures and/or social determinants of health (SDOH), data are typically available by counties, zip codes or census tracts. We consider two quantitative challenges in this setting when modeling cancer outcomes. The first is that multiple, potentially highly correlated exposures may contribute to risk. The second is that the spatial correlation of such features may not follow the original models designed for disease risk mapping. While several conditional autoregressive (CAR) models are typically utilized (Besag et al. (1991), Leroux et al. (1999), Stern and Cressie (1999)), a critical drawback is that they impose a global level of autocorrelation across the study area, presuming areas nearest each other are most similar. Spatial data may actually include step changes in the risk surface, indicative of more complex patterns either due to physical barriers and/or social isolation. Thus, identifying these step changes is of interest, as they may reflect the underlying complexity of the relationship between environmental exposures, SDOH and cancer mortality. We consider a two-step model to explore how clusters of environmental exposures and SDOH may be able to identify boundaries in the risk surface of cancer outcomes. In the first step, we estimate clusters of exposures using Bayesian profile regression, a nonparametric approach based on Dirichlet process mixture methods. In the second step, we model cancer mortality using locally adaptive spatial smoothing with the clusters as dissimilarity metrics (Lee and Mitchell, 2012). Several diagnostic tools are used to assess convergence in each step. With this work, we bridge the gap between incorporating mixtures of correlated exposures with the identification of localized areas with elevated risk of cancer mortality. We apply this approach to a nationwide database of county-level breast cancer mortality, utilize environmental exposure data from the National Air Toxics Assessment and assess SDOH from the American Community Survey. Citation Format: Kara McCormack. Social determinants and environmental exposures and their localized contribution to risk of cancer mortality [abstract]. In: Proceedings of the 15th AACR Conference on the Science of Cancer Health Disparities in Racial/Ethnic Minorities and the Medically Underserved; 2022 Sep 16-19; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2022;31(1 Suppl):Abstract nr C005.

Pan-cancer mutational signature surveys correlated mutational signature with geospatial environmental exposures and viral infections

BackgroundCancer has been disproportionally affecting minorities. Genomic-based cancer disparity analyses have been less common than conventional epidemiological studies. In the past decade, mutational signatures have been established as characteristic footprints of endogenous or exogenous carcinogens. MethodsIntegrating datasets of diverse cancer types from The Cancer Genome Atlas and geospatial environmental risks of the registry hospitals from the United States Environmental Protection Agency, we explored mutational signatures from the aspect of racial disparity concerning pollutant exposures. The raw geospatial environmental exposure data were refined to 449 air pollutants archived and modeled from 2007 to 2017 and aggregated to the census county level. Additionally, hepatitis B and C viruses and human papillomavirus infection statuses were incorporated into analyses for skin cancer, cervical cancer, and liver cancer. ResultsMutation frequencies of key oncogenic genes varied substantially between different races. These differences were further translated into differences in mutational signatures. Survival analysis revealed that the increased pollution level is associated with worse survival. The analysis of the oncogenic virus revealed that aflatoxin, an affirmed carcinogen for liver cancer, was higher in Asian liver cancer patients than in White patients. The aflatoxin mutational signature was exacerbated by hepatitis infection for Asian patients but not for White patients, suggesting a predisposed genetic or genomic disadvantage for Asians concerning aflatoxin. ConclusionsEnvironmental pollutant exposures increase a mutational signature level and worsen cancer prognosis, presenting a definite adverse risk factor for cancer patients.

Tyre wear nanoparticles as test for a nano risk governance framework

The Horizon 2020 consortia NanoRigo, Gov4Nano and RiskGONE jointly developed a nano risk governance framework (NRGF), fit for use for the practice of nanomaterials’ development and use, understandable and executable for stakeholders involved in manufacturing, using or regulating nanomaterials, or confronted with incidental exposure to generated nanomaterials. The NRGF includes risk management models, tools and approaches relevant to nanomaterials, as well as nano-oriented LCA and grouping, and takes into account socio-economic aspects and risk-benefit assessment. Special attention was paid to operationalizing concern assessment within the framework of the risk governance approach, emphasizing the different nature of quantitative risk assessment and qualitative concern assessment. The usefulness and added value of the NRGF was tested by means of a case study on rubber tyres. This case study performed a “mock-up” risk and concern assessment of the use of manufactured nanomaterials in the rubber tyre tread and the environmental release of tyre wear particles (TWP) from this tread during car driving. With a focus on the release and hazards of nano-TWP, an approach using the NRGF methodology was defined in collaboration with a group of real-life stakeholders. Simultaneously a group of scientific experts assessed the usefulness of the available tools and methods for assessing this nano-TWP release and investigated related broader societal concerns. The exercise concludes that so far, an exclusive risk assessment of nano-TWP is a step too far, due to the lack of nano-specific health- and environmental exposure and hazard data. Moreover, it seems that public concern about nano-sized TWP emissions is still limited, contrary to emerging scientific concerns. The case study underlines the complexity of deriving robust recommendations for “real-life” cases. Many questions remain on how to weigh the available technical and social evidence in nano risk governance.

A Description of the Imaging Innovations for Placental Assessment in Response to Environmental Pollution Study.

The aim of Placental Assessment in Response to Environmental Pollution Study (PARENTs) was to determine whether imaging of the placenta by novel multiparametric magnetic resonance imaging (MRI) techniques in early pregnancy could help predict adverse pregnancy outcomes (APOs) due to ischemic placental disease (IPD). Additionally, we sought to determine maternal characteristics and environmental risk factors that contribute to IPD and secondary APOs. Potential patients in their first trimester of pregnancy, who agreed to MRI of the placenta and measures of assessment of environmental pollution, were recruited into PARENTs, a prospective population-based cohort study. Participants were seen at three study visits during pregnancy and again at their delivery from 2015 to 2019. We collected data from interviews, chart abstractions, and imaging. Maternal biospecimens (serum, plasma, and urine) at antepartum study visits and delivery specimens (placenta, cord, and maternal blood) were collected, processed, and stored. The primary outcome was a composite of IPD, which included any of the following: placental abruption, hypertensive disease of pregnancy, fetal growth restriction, or a newborn of small for gestational age. In this pilot cohort, of the 190 patients who completed pregnancy to viable delivery, 50 (26%) developed IPD. Among demographic characteristics, having a history of prior IPD in multiparous women was associated with the development of IPD. In the multiple novel perfusion measurements taken of the in vivo placenta using MRI, decreased high placental blood flow (mL/100 g/min) in early pregnancy (between 14 and 16 weeks) was found to be significantly associated with the later development of IPD. Successful recruitment of the PARENTs prospective cohort demonstrated the feasibility and acceptability of the use of MRI in human pregnancy to study the placenta in vivo and at the same time collect environmental exposure data. Analysis is ongoing and we hope these methods will assist researchers in the design of prospective imaging studies of pregnancy. · MRI was acceptable and feasible for the study of the human placenta in vivo.. · Functional imaging of the placenta by MRI showed a significant decrease in high placental blood flow.. · Measures of environmental exposures are further being analyzed to predict IPD..

Representative oral microbiome data for the US population: the National Health and Nutrition Examination Survey

The human oral microbiome plays many important roles in human health.1 However, most oral microbiome studies have been conducted among convenience samples of volunteers and thus might not fully reflect the diversity of oral microbial communities in the human population.2 To understand the oral microbiome, large, diverse, population-representative studies are needed that collect demographic, behavioural, environmental, and other exposure data and that follow up individuals over time for health outcomes.

P099 Feasibility of assessing environmental factors in residential treatment settings: Preliminary associations between objective sleep and mental health outcomes

Abstract Introduction Circadian dysregulation is implicated in substance use and comorbid mental health problems, but is also exacerbated by these states. In addition to conventional treatment, environmental factors (light, temperature) are potential targets for improving sleep, circadian functioning, and recovery outcomes. This study explored the feasibility of objective sleep and environmental monitoring for recovering substance users in residential treatment, and also examined the relationship between objectively measured sleep-wake rhythms and mental health outcomes. Methods Fifty-two (Mage=38, SD=9.6; 56% female) individuals receiving treatment for substance use were included. Environmental light exposure, temperature, and activity data were collected via wrist-worn actigraphy, and in-room sensors for a minimum of 1-week of stay within a 6-week program. Actigraphy derived sleep-wake patterns were used to calculate Interdaily Stability (IS). Data were linked to routinely collected clinical measures of mental health. Results No individuals opted out of the environmental monitoring aspect of the study, indicating acceptability and feasibility. Environmental loggers mounted in individual rooms recorded average nightly temperatures &amp;gt;21 degrees for 5/8 months. Objectively measured IS (M=.79; 95%CI=.67-.73) was negatively associated with Anxiety (r=-.33, p=.024) and Depression (r=-.36, p=.012), while sleep duration was not. Relationships between participant-level light exposure and clinical outcomes were also examined. Discussion Preliminary results indicate that circadian regularity appears important for mental health outcomes in this population and setting. The structured residential environment provides opportunities for regulating sleep-wake rhythms via scheduling of activities at circadian-appropriate times. This research suggests that optimising circadian- appropriate lighting and thermal comfort may lead to improved individual outcomes.

Environmental Factors at the Periconceptional Period and the Occurrence of Cleft Lip and Palate in a Midwest Brazil Population: A Case-Control Study.

To investigate the association between periconceptional environmental exposures and the occurrence of cleft lips and palates. This case-control study analyzed 150 mothers of children with cleft lips and palates living in the same city as 250 mothers whose children did not present with this malformation (controls). Environmental exposure data were gathered through a questionnaire (Latin American Collaborative Study of Congenital Malformations methodology). Multivariate analysis revealed that monthly income below minimum wage, having another malformed child, other diseases in the first gestational trimester (urinary infection), use of pesticides in home gardens, and pesticide use in farms close to the home were risk factors associated with the malformation, whereas taking vitamins was a protective factor. Maternal and paternal exposure to pesticides is associated with cleft lip and palate in Mato Grosso State, Brazil.

Occupational and environmental exposures in the Genomic Research in Alpha-1 Antitrypsin Deficiency and Sarcoidosis (GRADS) study

IntroductionSarcoidosis is a granulomatous disorder thought to be caused by exposures in genetically susceptible individuals. This study investigated whether specific exposures were associated with different sarcoidosis phenotypes. MethodsExtensive demographic, occupational and environmental exposure data was analyzed from subjects enrolled in the NHLBI Genomic Research in Alpha-1 Antitrypsin Deficiency and Sarcoidosis (GRADS) study. ResultsIn patients with sarcoidosis, radiation exposure was significantly associated with an increased risk of cardiac sarcoidosis versus non-cardiac sarcoidosis. No exposures were significantly associated with pulmonary only disease versus extrapulmonary disease with or without pulmonary involvement, Scadding Stage II/III/IV versus Scadding Stage 0/I, acute or remitting disease versus non-acute or non-remitting disease, nor chronic versus non-chronic disease. Although not reaching statistically significance after adjustment for multiple comparisons, there were a number of exposures associated with specific disease phenotypes, including exposures where relationships to sarcoidosis have previously been described such as rural exposures and pesticide exposures. ConclusionsRadiation exposure may be a risk factor for cardiac sarcoidosis. Other exposures may also be associated with specific phenotypes and should be further explored. The study was limited by small groups of exposed subjects for individual exposures and multiple comparisons. The development of novel and innovative exposure assessment tools is needed.

Leveraging Open Electronic Health Record Data and Environmental Exposures Data to Derive Insights Into Rare Pulmonary Disease.

Research on rare diseases has received increasing attention, in part due to the realized profitability of orphan drugs. Biomedical informatics holds promise in accelerating translational research on rare disease, yet challenges remain, including the lack of diagnostic codes for rare diseases and privacy concerns that prevent research access to electronic health records when few patients exist. The Integrated Clinical and Environmental Exposures Service (ICEES) provides regulatory-compliant open access to electronic health record data that have been integrated with environmental exposures data, as well as analytic tools to explore the integrated data. We describe a proof-of-concept application of ICEES to examine demographics, clinical characteristics, environmental exposures, and health outcomes among a cohort of patients enriched for phenotypes associated with cystic fibrosis (CF), idiopathic bronchiectasis (IB), and primary ciliary dyskinesia (PCD). We then focus on a subset of patients with CF, leveraging the availability of a diagnostic code for CF and serving as a benchmark for our development work. We use ICEES to examine select demographics, co-diagnoses, and environmental exposures that may contribute to poor health outcomes among patients with CF, defined as emergency department or inpatient visits for respiratory issues. We replicate current understanding of the pathogenesis and clinical manifestations of CF by identifying co-diagnoses of asthma, chronic nasal congestion, cough, middle ear disease, and pneumonia as factors that differentiate patients with poor health outcomes from those with better health outcomes. We conclude by discussing our preliminary findings in relation to other published work, the strengths and limitations of our approach, and our future directions.

Association of ambient fine particulate air pollution with hospitalization costs and stays for depression in China: a multicity analysis

Objective: To explore the effect of fine particulate matter (PM2.5) pollution on depression hospitalization cost and length of stay in 57 cities of China. Methods: A total of 84 207 patients with depression in 57 cities of China from January 2013 to December 2017 were selected as the subjects. The demographic characteristics and hospitalization status of the patients were obtained from the database of basic medical insurance for urban workers and urban residents in China. The environmental exposure data of the same period were obtained from the national air quality real-time release platform of China Environmental Monitoring Station. A generalized additive model based on quasi-Poisson distribution was used to analyze PM2.5 exposure effect in each city, and the nonlinear mixing of moving average temperature, relative humidity and date was controlled by natural smooth spline function. Results: Among the included cities, southern cities accounted for 50.88% (29), and the number of female inpatients, hospitalization costs and hospitalization days accounted for 62.65%, 63.50% and 60.85% (42 735 cases, 567.78 million yuan and 1.14 million days, respectively). The proportion of hospitalized cases, hospitalization cost and length of stay in the age group of 40 to 64 years old were 59.15% (40 346 cases), 53.92% (482.15 million yuan) and 52.07% (0.98 million days), respectively. PM2.5 level was positively correlated with the number of hospitalized cases with depression, hospitalization cost and length of stay. When the 3-day moving average of PM2.5 exposure level increased by 10 μg/m3, the number of hospitalization cases increased by 0.64%. The attributed percentage (95%CI) of hospitalized cases, hospitalization costs and length of stay were 3.35% (0.57%-6.04%), 3.04% (0.52%-5.48%) and 3.07% (0.49%-5.56%), respectively. Subgroup analysis showed that the attributed percentage of hospitalization cases, hospitalization cost and length of stay to PM2.5 exposure ranged from 3.97% to 4.68%, 4.04% to 4.33% and 4.13% to 4.30% in northern China, male and cold season, respectively. Conclusion: PM2.5 exposure is associated with the increase of hospitalization cost and length of stay among Chinese urban population with depression.

Development and Application of an Open Tool for Sharing and Analyzing Integrated Clinical and Environmental Exposures Data: Asthma Use Case.

BackgroundThe Integrated Clinical and Environmental Exposures Service (ICEES) serves as an open-source, disease-agnostic, regulatory-compliant framework and approach for openly exposing and exploring clinical data that have been integrated at the patient level with a variety of environmental exposures data. ICEES is equipped with tools to support basic statistical exploration of the integrated data in a completely open manner.ObjectiveThis study aims to further develop and apply ICEES as a novel tool for openly exposing and exploring integrated clinical and environmental data. We focus on an asthma use case.MethodsWe queried the ICEES open application programming interface (OpenAPI) using a functionality that supports chi-square tests between feature variables and a primary outcome measure, with a Bonferroni correction for multiple comparisons (α=.001). We focused on 2 primary outcomes that are indicative of asthma exacerbations: annual emergency department (ED) or inpatient visits for respiratory issues; and annual prescriptions for prednisone.ResultsOf the 157,410 patients within the asthma cohort, 26,332 (16.73%) had 1 or more annual ED or inpatient visits for respiratory issues, and 17,056 (10.84%) had 1 or more annual prescriptions for prednisone. We found that close proximity to a major roadway or highway, exposure to high levels of particulate matter ≤2.5 μm (PM2.5) or ozone, female sex, Caucasian race, low residential density, lack of health insurance, and low household income were significantly associated with asthma exacerbations (P<.001). Asthma exacerbations did not vary by rural versus urban residence. Moreover, the results were largely consistent across outcome measures.ConclusionsOur results demonstrate that the open-source ICEES can be used to replicate and extend published findings on factors that influence asthma exacerbations. As a disease-agnostic, open-source approach for integrating, exposing, and exploring patient-level clinical and environmental exposures data, we believe that ICEES will have broad adoption by other institutions and application in environmental health and other biomedical fields.

Clouding Up Cognition? Secondhand Cannabis and Tobacco Exposure Related to Cognitive Functioning in Youth

BackgroundIncreasing legalization of cannabis, in addition to longstanding rates of tobacco use, raises concerns for possible cognitive decrements from secondhand smoke or environmental exposure, although little research exists. We investigate the relation between cognition and secondhand and environmental cannabis and tobacco exposure in youth. MethodsThe Adolescent Brain Cognitive Development (ABCD) Study year 2 follow-up (N = 5580; 48% female) cognitive performance and secondhand or environmental cannabis or tobacco exposure data were used. Principal components analysis identified a global cognition factor. Linear mixed-effects models assessed global cognition and individual cognitive task performance by cannabis and/or tobacco environmental exposure. Sociodemographics and other potential confounds were examined. p values were adjusted using the false discovery rate method. ResultsGlobal cognition was not related to any exposure group after testing corrections and considering confounds. Beyond covariates and family- and site-level factors, secondhand tobacco was related to poorer visual memory (p = .02), and environmental tobacco was associated with poorer visuospatial (p = .02) and language (p = .008) skills. Secondhand cannabis was related to cognition, but not after controlling for potential confounders (p > .05). Environmental cannabis was related to better oral reading (p = .01). Including covariates attenuated effect sizes. ConclusionsSecondhand tobacco exposure was associated with poorer visual memory, while environmental tobacco exposure was related to poorer language and visuospatial skills. Secondhand cannabis was not related to cognition after controlling for sociodemographic factors, but environmental cannabis exposure was related to better reading. Because, to our knowledge, this is the first known study of its kind and thus preliminary, secondhand cannabis should continue to be investigated to confirm results.

Leveraging Open Electronic Health Record Data and Environmental Exposures Data to Derive Insights Into Rare Pulmonary Disease

Leveraging Open Electronic Health Record Data and Environmental Exposures Data to Derive Insights Into Rare Pulmonary Disease

Using dynamic time warping self-organizing maps to characterize diurnal patterns in environmental exposures

Advances in measurement technology are producing increasingly time-resolved environmental exposure data. We aim to gain new insights into exposures and their potential health impacts by moving beyond simple summary statistics (e.g., means, maxima) to characterize more detailed features of high-frequency time series data. This study proposes a novel variant of the Self-Organizing Map (SOM) algorithm called Dynamic Time Warping Self-Organizing Map (DTW-SOM) for unsupervised pattern discovery in time series. This algorithm uses DTW, a similarity measure that optimally aligns interior patterns of sequential data, both as the similarity measure and training guide of the neural network. We applied DTW-SOM to a panel study monitoring indoor and outdoor residential temperature and particulate matter air pollution (PM2.5) for 10 patients with asthma from 7 households near Salt Lake City, UT; the patients were followed for up to 373 days each. Compared to previous SOM algorithms using timestamp alignment on time series data, the DTW-SOM algorithm produced fewer quantization errors and more detailed diurnal patterns. DTW-SOM identified the expected typical diurnal patterns in outdoor temperature which varied by season, as well diurnal patterns in PM2.5 which may be related to daily asthma outcomes. In summary, DTW-SOM is an innovative feature engineering method that can be applied to highly time-resolved environmental exposures assessed by sensors to identify typical diurnal (or hourly or monthly) patterns and provide new insights into the health effects of environmental exposures.

Development of childhood asthma prediction models using machine learning approaches.

Respiratory symptoms are common in early life and often transient. It is difficult to identify in which children these will persist and result in asthma. Machine learning (ML) approaches have the potential for better predictive performance and generalisability over existing childhood asthma prediction models. This study applied ML approaches to predict school-age asthma (age 10) in early life (Childhood Asthma Prediction in Early life, CAPE model) and at preschool age (Childhood Asthma Prediction at Preschool age, CAPP model). Clinical and environmental exposure data was collected from children enrolled in the Isle of Wight Birth Cohort (N=1368, ∼15% asthma prevalence). Recursive Feature Elimination (RFE) identified an optimal subset of features predictive of school-age asthma for each model. Seven state-of-the-art ML classification algorithms were used to develop prognostic models. Training was performed by applying fivefold cross-validation, imputation, and resampling. Predictive performance was evaluated on the test set. Models were further externally validated in the Manchester Asthma and Allergy Study (MAAS) cohort. RFE identified eight and twelve predictors for the CAPE and CAPP models, respectively. Support Vector Machine (SVM) algorithms provided the best performance for both the CAPE (area under the receiver operating characteristic curve, AUC=0.71) and CAPP (AUC=0.82) models. Both models demonstrated good generalisability in MAAS (CAPE 8-year=0.71, 11-year=0.71, CAPP 8-year=0.83, 11-year=0.79) and excellent sensitivity to predict a subgroup of persistent wheezers. Using ML approaches improved upon the predictive performance of existing regression-based models, with good generalisability and ability to rule in asthma and predict persistent wheeze.