Dataset Information

Concentrations of criteria pollutants in the contiguous U.S., 1979 - 2015: Role of prediction model parsimony in integrated empirical geographic regression.

ABSTRACT: National-scale empirical models for air pollution can include hundreds of geographic variables. The impact of model parsimony (i.e., how model performance differs for a large versus small number of covariates) has not been systematically explored. We aim to (1) build annual-average integrated empirical geographic (IEG) regression models for the contiguous U.S. for six criteria pollutants during 1979-2015; (2) explore systematically the impact on model performance of the number of variables selected for inclusion in a model; and (3) provide publicly available model predictions. We compute annual-average concentrations from regulatory monitoring data for PM10, PM2.5, NO2, SO2, CO, and ozone at all monitoring sites for 1979-2015. We also use ~350 geographic characteristics at each location including measures of traffic, land use, land cover, and satellite-based estimates of air pollution. We then develop IEG models, employing universal kriging and summary factors estimated by partial least squares (PLS) of geographic variables. For all pollutants and years, we compare three approaches for choosing variables to include in the PLS model: (1) no variables, (2) a limited number of variables selected from the full set by forward selection, and (3) all variables. We evaluate model performance using 10-fold cross-validation (CV) using conventional and spatially-clustered test data. Models using 3 to 30 variables selected from the full set generally have the best performance across all pollutants and years (median R2 conventional [clustered] CV: 0.66 [0.47]) compared to models with no (0.37 [0]) or all variables (0.64 [0.27]). Concentration estimates for all Census Blocks reveal generally decreasing concentrations over several decades with local heterogeneity. Our findings suggest that national prediction models can be built by empirically selecting only a small number of important variables to provide robust concentration estimates. Model estimates are freely available online.

SUBMITTER: Kim SY

PROVIDER: S-EPMC7028280 | biostudies-literature | 2020

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Concentrations of criteria pollutants in the contiguous U.S., 1979 - 2015: Role of prediction model parsimony in integrated empirical geographic regression.

Kim Sun-Young SY Bechle Matthew M Hankey Steve S Sheppard Lianne L Szpiro Adam A AA Marshall Julian D JD

PloS one 20200218 2

National-scale empirical models for air pollution can include hundreds of geographic variables. The impact of model parsimony (i.e., how model performance differs for a large versus small number of covariates) has not been systematically explored. We aim to (1) build annual-average integrated empirical geographic (IEG) regression models for the contiguous U.S. for six criteria pollutants during 1979-2015; (2) explore systematically the impact on model performance of the number of variables selec ...[more]

PMID: 32069301

Similar Datasets

Project description:Importance:Cardiac amyloidosis is an underdiagnosed disease and is highly fatal when untreated. Early diagnosis and treatment with the emerging novel therapies significantly improve survival. A comprehensive analysis of amyloidosis-related mortality is critical to appreciate the nature and distribution of underdiagnosis and improve disease detection. Objective:To evaluate the temporal and regional trends in age-adjusted amyloidosis-related mortality among men and women of various races/ethnicities in the United States. Design, Setting, and Participants:In this observational cohort study, death certificate information from the Centers for Disease Control and Prevention's Wide-ranging ONline Data for Epidemiologic Research database and the National Vital Statistics System from 1979 to 2015 was analyzed. A total of 30?764 individuals in the United States with amyloidosis listed as the underlying cause of death and 26?591 individuals with amyloidosis listed as a contributing cause of death were analyzed. Exposures:Region of residence. Main Outcomes and Measures:Age-adjusted mortality rate from amyloidosis per 1?000?000 population stratified by year, sex, race/ethnicity, and state and county of residence. Results:Of the 30?764 individuals with amyloidosis listed as the underlying cause of death, 17?421 (56.6%) were men and 27?312 (88.8%) were 55 years or older. From 1979 to 2015, the reported overall mean age-adjusted mortality rate from amyloidosis as the underlying cause of death doubled from 1.77 to 3.96 per 1?000?000 population (2.32 to 5.43 in men and 1.35 to 2.80 in women). Black men had the highest mortality rate (12.36 per 1?000?000), followed by black women (6.48 per 1?000?000). Amyloidosis contributed to age-adjusted mortality rates as high as 31.73 per 1?000?000 in certain counties. Most southern states reported the lowest US mortality rates despite having the highest proportions of black individuals. Conclusions and Relevance:The increased reported mortality over time and in proximity to amyloidosis centers more likely reflects an overall increase in disease diagnosis rather than increased lethality. The reported amyloidosis mortality is highly variable in different US regions. The lack of higher reported mortality rates in states with a greater proportion of black residents suggests underdiagnosis of amyloidosis, including cardiac forms of the disease, in many areas of the United States. Better understanding of the determinants of geographic and racial disparity in the reporting of amyloidosis deaths are warranted.

Project description:BackgroundA comprehensive clinical long-term survey over the complete spectrum of neuroblatoma disease is lacking in the literature.ObjectiveOur objective was to describe the incidence, risk profiles, therapies, and outcomes for the total cohort of German patients with neuroblastoma including all clinical stages and risk groups.MethodsEpidemiological, clinical, and outcome data of neuroblastoma patients who participated in one of the six consecutive national trials between 1979 and 2015 were analyzed retrospectively.ResultsOf all German neuroblastoma patients known to the national childhood cancer registry, ninety seven percent enrolled in one of the trials. The absolute neuroblastoma rate has increased slightly, whereas the median age at diagnosis has decreased. Except for the screening period (1995-2000), the risk factors lactate dehydrogenase (LDH), ferritin, chromosome 1p, and the MYCN oncogene have remained largely constant, with the exception of an increase in MYCN amplification at stage 4 for those aged ?18 months between trials NB97 (27%) and NB2004 (35%). The 10-year overall survival increased in patients with stage 1-3 neuroblastoma from 83 to 91%, for stage 4S from 80 to 85%, and for stage 4 aged ?18 months from 2 to 38%. The fraction of patients in stages 1-3 who never received chemotherapy (neither for frontline nor at recurrence) increased from 35 to 60%. The proportion of macroscopically complete surgical resections of the primary tumor decreased for the total population as well as for patients with stage 4 aged ?18 months. The impact of chemotherapy response on the outcome was trial dependent. The overall proportion of toxic death during the time of the protocol therapy was 6% for stage 4 patients aged ?18 months and 2% for low-/intermediate-risk patients. The most frequently reported late sequelae in stage 4 patients aged ?18 months were renal dysfunctions, hypothyroidism, major hearing impairment, and second malignancies.ConclusionThe body of data for incidences, risk profiles, and survival rates from this survey of more than 37 years provides a useful perspective for future studies on neuroblastoma sub-cohorts.

Project description:ObjectiveEndocrine-disrupting chemicals (EDCs) adversely affect human health. Our objective was to determine the association of EDC exposure with earlier age of menopause.MethodsCross-sectional survey using National Health and Nutrition Examination Survey (NHANES) data from 1999 to 2008 (n = 31,575 females). Eligible participants included: menopausal women >30 years of age; not currently pregnant, breastfeeding, using hormonal contraception; no history of bilateral oophorectomy or hysterectomy. Exposures, defined by serum lipid and urine creatinine-adjusted measures of EDCs, data were analyzed: > 90th percentile of the EDC distribution among all women, log-transformed EDC level, and decile of EDC level. Multi linear regression models considered complex survey design characteristics and adjusted for age, race/ethnicity, smoking, body mass index. EDCs were stratified into long (>1 year), short, and unknown half-lives; principle analyses were performed on those with long half-lives as well as phthalates, known reproductive toxicants. Secondary analysis determined whether the odds of being menopausal increased with EDC exposure among women aged 45-55 years.FindingsThis analysis examined 111 EDCs and focused on known reproductive toxicants or chemicals with half-lives >1 year. Women with high levels of ?-hexachlorocyclohexane, mirex, p,p'-DDE, 1,2,3,4,6,7,8-heptachlorodibenzofuran, mono-(2-ethyl-5-hydroxyhexyl) and mono-(2-ethyl-5-oxohexyl) phthalate, polychlorinated biphenyl congeners -70, -99, -105, -118, -138, -153, -156, -170, and -183 had mean ages of menopause 1.9 to 3.8 years earlier than women with lower levels of these chemicals. EDC-exposed women were up to 6 times more likely to be menopausal than non-exposed women.ConclusionsThis study of a representative sample of US women documents an association between EDCs and earlier age at menopause. We identified 15 EDCs that warrant closer evaluation because of their persistence and potential detrimental effects on ovarian function. Earlier menopause can alter the quantity and quality of a woman's life and has profound implications for fertility, human reproduction, and our global society.

Dataset Information

Concentrations of criteria pollutants in the contiguous U.S., 1979 - 2015: Role of prediction model parsimony in integrated empirical geographic regression.

Publications

Concentrations of criteria pollutants in the contiguous U.S., 1979 - 2015: Role of prediction model parsimony in integrated empirical geographic regression.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure