Paul Sampson
University of Washington, Statistics, Faculty Member
We propose novel methods for predictive (sparse) PCA with spatially misaligned data. These methods identify principal component loading vectors that explain as much variability in the observed data as possible, while also ensuring the... more
We propose novel methods for predictive (sparse) PCA with spatially misaligned data. These methods identify principal component loading vectors that explain as much variability in the observed data as possible, while also ensuring the corresponding principal component scores can be predicted accurately by means of spatial statistics at locations where air pollution measurements are not available. This will make it possible to identify important mixtures of air pollutants and to quantify their health effects in cohort studies, where currently available methods cannot be used. We demonstrate the utility of predictive (sparse) PCA in simulated data and apply the approach to annual averages of particulate matter speciation data from national Environmental Protection Agency (EPA) regulatory monitors.
Research Interests:
There is growing evidence in the epidemiologic literature of the relationship between air pollution and adverse health outcomes. Prediction of individual air pollution exposure in the Environmental Protection Agency (EPA) funded... more
There is growing evidence in the epidemiologic literature of the relationship between air pollution and adverse health outcomes. Prediction of individual air pollution exposure in the Environmental Protection Agency (EPA) funded Multi-Ethnic Study of Atheroscelerosis and Air Pollution (MESA Air) study relies on a flexible spatio-temporal prediction model that integrates land-use regression with kriging to account for spatial dependence in pollutant concentrations. Temporal variability is captured using temporal trends estimated via modified singular value decomposition and temporally varying spatial residuals. This model utilizes monitoring data from existing regulatory networks and supplementary MESA Air monitoring data to predict concentrations for individual cohort members. In general, spatio-temporal models are limited in their efficacy for large data sets due to computational intractability. We develop reduced-rank versions of the MESA Air spatio-temporal model. To do so, we ap...
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Large-scale animal feeding operations compromise regional air quality in the rural US through emission of pollutants, such as ammonia gas. Exposure to airborne pollution from animal feeding operations may cause pediatric asthma... more
Large-scale animal feeding operations compromise regional air quality in the rural US through emission of pollutants, such as ammonia gas. Exposure to airborne pollution from animal feeding operations may cause pediatric asthma exacerbations in surrounding communities. To describe spatial and temporal patterns in ambient ammonia concentrations in an agricultural region, and to investigate associations between short-term fluctuations in ammonia and subsequent changes in respiratory health in children with asthma. For 13 months in the Yakima Valley of Washington State, 14 monitors sampled ammonia in outdoor air for 24-hour periods every 6 days. School-age children with asthma (n = 51) were followed for two health outcomes: biweekly reports of asthma symptoms and quick relief medication usage, and daily measurements of forced expiratory volume in 1 second. We assessed associations between each outcome and ammonia using generalized estimating equations. Twenty-four-hour ammonia concentr...
Research Interests: Statistics, Epidemiology, Environmental Health, Asthma, Agriculture, and 15 moreAir pollution, Medicine, Animal Husbandry, Community Based Participatory Research, Humans, Child, Female, Male, Weather, Ammonia, Washington, Indoor air pollution, Longitudinal Studies, Public health systems and services research, and Environmental Exposure
Studies estimating health effects of long-term air pollution exposure often use a two-stage approach, building exposure models to assign individual-level exposures which are then used in regression analyses. This requires accurate... more
Studies estimating health effects of long-term air pollution exposure often use a two-stage approach, building exposure models to assign individual-level exposures which are then used in regression analyses. This requires accurate exposure modeling and careful treatment of exposure measurement error. To illustrate the importance of carefully accounting for exposure model characteristics in two-stage air pollution studies, we consider a case study based on data from the Multi-Ethnic Study of Atherosclerosis (MESA). We present national spatial exposure models that use partial least squares and universal kriging to estimate annual average concentrations of four PM2.5 components: elemental carbon (EC), organic carbon (OC), sulfur (S), and silicon (Si). Our models perform well, with cross-validated Rs ranging from 0.62 to 0.95. We predict PM2.5 component exposures for the MESA cohort and estimate cross-sectional associations with carotid intima-media thickness (CIMT), adjusting for subje...
Research Interests:
Given the increasing interest in the association between exposure to air pollution and adverse health outcomes, the development of models that provide accurate spatio-temporal predictions of air pollution concentrations at small spatial... more
Given the increasing interest in the association between exposure to air pollution and adverse health outcomes, the development of models that provide accurate spatio-temporal predictions of air pollution concentrations at small spatial scales is of great importance when assessing potential health effects of air pollution. The methodology presented here has been developed as part of the Multi-Ethnic Study of Atherosclerosis and Air Pollution (MESA Air), a prospective cohort study funded by the US EPA to investigate the relationship between chronic exposure to air pollution and cardiovascular disease. We present a spatio-temporal framework that models and predicts ambient air pollution by combining data from several different monitoring networks with the output from deterministic air pollution model(s). The model can accommodate arbitrarily missing observations and allows for a complex spatio-temporal correlation structure. We apply the model to predict long-term average concentratio...
Seeking improved speech intelligibility in a university classroom. [The Journal of the Acoustical Society of America 107, 2833 (2000)]. Dean R. Heerwagen, Paul D. Sampson. Abstract. Speech intelligibility has been studied while... more
Seeking improved speech intelligibility in a university classroom. [The Journal of the Acoustical Society of America 107, 2833 (2000)]. Dean R. Heerwagen, Paul D. Sampson. Abstract. Speech intelligibility has been studied while systematically modifying an existing classroom. ...
Research Interests:
We sought to evaluate potential mediators of the relationship between heat exposure and traumatic injuries in outdoor agricultural workers. Linear mixed models were used to estimate associations between maximum work-shift Wet Bulb Globe... more
We sought to evaluate potential mediators of the relationship between heat exposure and traumatic injuries in outdoor agricultural workers. Linear mixed models were used to estimate associations between maximum work-shift Wet Bulb Globe Temperature (WBGT) and post-shift vigilance (reaction time) and postural sway (total path length) in a cross-sectional sample of 46 Washington State tree fruit harvesters in August-September 2015. The mean (SD) WBGT was 27.4 (3.2)°C in August and 21.2 (2.0)°C in September. The mean pre-work-shift participant urine specific gravity indicated minimal dehydration. Twenty-four percent of participants exhibited possible excessive sleepiness. There was no association between WBGT and post-shift reaction time or total path length. Heat exposure was not associated with impaired vigilance or balance in this study, in which the overall mean (SD) WBGT was 25.9 (4.2)°C. However, the study identified opportunities to ensure adequate pre-work-shift hydration and to optimize sleep and work-shift timing in order to reduce occupational injury and heat-related illness risk.
Research Interests:
Recent studies suggest that heat exposure degrades work productivity, but such studies have not considered individual- and workplace-level factors. Forty-six tree-fruit harvesters (98% Latino/a) from 6 orchards participated in a... more
Recent studies suggest that heat exposure degrades work productivity, but such studies have not considered individual- and workplace-level factors. Forty-six tree-fruit harvesters (98% Latino/a) from 6 orchards participated in a cross-sectional study in central/eastern Washington in 2015. The association between maximum measured work-shift wet-bulb globe temperature (WBGTmax) and productivity (total weight of fruit bins collected per time worked) was estimated using linear mixed-effects models, adjusting for relevant confounders. The mean (standard deviation) WBGTmax was 27.9°C (3.6°C) in August and 21.2°C (2.0°C) in September. There was a trend of decreasing productivity with increasing WBGTmax, but this association was not statistically significant. When individual- and workplace-level factors were included in the model, the association approached the null. Not considering individual, work, and economic factors that affect rest and recovery in projections of the effects of climate...
Research Interests:
Minimally invasive surgery (MIS) technique is becoming the standard tissue sparing approach for decompression of lumbar central and lateral recess stenosis, intervertebral disc herniation, or any situation that would have required... more
Minimally invasive surgery (MIS) technique is becoming the standard tissue sparing approach for decompression of lumbar central and lateral recess stenosis, intervertebral disc herniation, or any situation that would have required extensive open decompression laminectomy. Full-endoscopic or arthroscopic assisted surgery is arguably the "ultra-MIS" approach to lumbar spinal pathology. Age and body mass index (BMI) are significant risk factors to be considered in full-endoscopic assisted ultra-MIS. With limited medical literature published on complication rates for MIS, reports on the ultra-MIS approach are even scarcer for free-standing, outpatient ambulatory settings. The primary goal of this study is to compare outcomes for full-endoscopic assisted ultra-MIS lumbar decompression surgical techniques, performed in a free-standing, outpatient ambulatory facility, with other spine surgery techniques. This is a Western Institutional Review Board (WIRB)-approved retrospective r...
Research Interests:
Research Interests:
Research Interests: Environmental Engineering, Atmospheric Science, Time Series, Statistical Analysis, Spatial Statistics, and 15 moreAtmospheric sciences, Extreme Value Theory, Variable Selection, Regression, Regression Analysis, Trend Estimation, Space Time, Regional scale, Tropospheric Ozone, Ozone, EXTREME VALUES, Troposphere, Atmospheric Environment, Extreme value analysis, and Statistical Method
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Current epidemiologic studies rely on simple ozone metrics which may not appropriately capture population ozone exposure. For understanding health effects of long-term ozone exposure in population studies, it is advantageous for exposure... more
Current epidemiologic studies rely on simple ozone metrics which may not appropriately capture population ozone exposure. For understanding health effects of long-term ozone exposure in population studies, it is advantageous for exposure estimation to incorporate the complex spatiotemporal pattern of ozone concentrations at fine scales. To develop a geo-statistical exposure prediction model that predicts fine scale spatiotemporal variations of ambient ozone in six United States metropolitan regions. We developed a modeling framework that estimates temporal trends from regulatory agency and cohort-specific monitoring data from MESA Air measurement campaigns and incorporates land use regression with universal kriging using predictor variables from a large geographic database. The cohort-specific data were measured at home and community locations. The framework was applied in estimating two-week average ozone concentrations from 1999 to 2013 in models of each of the six MESA Air metropolitan regions. Ozone models perform well in both spatial and temporal dimensions at the agency monitoring sites in terms of prediction accuracy. City-specific leave-one (site)-out cross-validation R(2) accounting for temporal and spatial variability ranged from 0.65 to 0.88 in the six regions. For predictions at the home sites, the R(2) is between 0.60 and 0.91 for cross-validation that left out 10% of home sites in turn. The predicted ozone concentrations vary substantially over space and time in all the metropolitan regions. Using the available data, our spatiotemporal models are able to accurately predict long-term ozone concentrations at fine spatial scales in multiple regions. The model predictions will allow for investigation of the long-term health effects of ambient ozone concentrations in future epidemiological studies.
Research Interests:
Research Interests:
Studies of pediatric lead exposure and IQ suggest that the association between measured blood lead concentrations and IQ is nonlinear, with the decline in IQ greater at lower levels of exposure (Bellinger and Needleman, 2003; Canfield et... more
Studies of pediatric lead exposure and IQ suggest that the association between measured blood lead concentrations and IQ is nonlinear, with the decline in IQ greater at lower levels of exposure (Bellinger and Needleman, 2003; Canfield et al., 2003; Jusko et al., 2005; Lanphear et al., ...
Research Interests:
Research Interests:
Research Interests:
The purpose of this study was to evaluate the long-term stability of orthodontically induced changes in maxillary and mandibular arch form. Dental casts were evaluated before treatment, after treatment, and a minimum of 10 years after... more
The purpose of this study was to evaluate the long-term stability of orthodontically induced changes in maxillary and mandibular arch form. Dental casts were evaluated before treatment, after treatment, and a minimum of 10 years after retention for 45 patients with Class I and 42 Class II, Division 1 malocclusions who received four first premolar extraction treatment. Computer generated arch forms were used to assess changes in arch shape over time. Buccal cusp tips of first molars, premolars, and canines plus mesial, distal, and central incisal aspects of incisors were marked, photocopied, and digitized in a standardized manner. An algorithm was used to fit conic sections to the digitized points. The shape of the fitted conics at each time period was described by calculating the parameter eccentricity; a small value represented a more rounded shape and a larger value represented a more tapered shape. Findings demonstrated a rounding of arch form during treatment followed by a change to more tapered. Arch form tended to return toward the pretreatment shape after retention. The greater the treatment change, the greater the tendency for postretention change. However, individual variation was considerable. The patient's pretreatment arch form appeared to be the best guide to future arch form stability, but minimizing treatment change was no guarantee of postretention stability.
Research Interests:
This study was part of a multidisciplinary investigation of the effects of gestational ethanol exposure in nonhuman primates. Thirty-one pregnant Macaca nemestrina were exposed to weekly ethanol doses of 0.0, 0.3, 0.6, 1.2, 1.8, 2.5, 3.3,... more
This study was part of a multidisciplinary investigation of the effects of gestational ethanol exposure in nonhuman primates. Thirty-one pregnant Macaca nemestrina were exposed to weekly ethanol doses of 0.0, 0.3, 0.6, 1.2, 1.8, 2.5, 3.3, or 4.1 g/kg maternal weight. Dose cohorts 0.0 through 1.8 were exposed to the initial ethanol dose within 10 days postconception. Dose cohorts 2.5 through 4.1 received their initial dose after the fifth week of gestation. Morphometric analyses performed on cranial radiographs showed that animals exposed to high doses of gestational ethanol had, on average, slightly smaller, distorted crania than control animals. A dysmorphic, flat face characteristic of fetal alcohol syndrome was recognized in one animal of the 1.8 g/kg cohort. The animal that received the highest doses of gestational ethanol was microcephalic. Similar malformations were not seen with low ethanol exposures or in controls. These data suggest a pattern of cranial distortion that may be recognizable and characteristic of ethanol teratogenesis.
Research Interests:
This study was designed to assess the limits of alcohol-related facial dysmorphogenesis. Standard full face and lateral facial photographs were obtained on 21 7-year-old children who had been exposed gestationally to known, heavy... more
This study was designed to assess the limits of alcohol-related facial dysmorphogenesis. Standard full face and lateral facial photographs were obtained on 21 7-year-old children who had been exposed gestationally to known, heavy quantities of ethanol. Only two of these children had been previously considered to have definite fetal alcohol syndrome (FAS). Similar photographs of 21 other 7-year-old children with negligible gestational ethanol exposure were obtained for control purposes. Copies of the 42 photographs were given to each of seven expert clinicians who were asked to select any child with an FAS-related facial appearance. Six of seven judges were accurate in identifying children with high levels of alcohol exposure as having a fetal alcohol-affected face. A set of homologous points on the photographs were then digitized and analyzed by newly developed morphometric methods to determine the facial shape characteristics that distinguish the selected photographs of highly exposed children. The analysis confirmed that the judges specifically identified children with facial changes consistent with those previously published as defining the face of the FAS: short palpebral fissures, a relatively long and flat midface, and a retrusive mandible. This methodology may be useful in more accurately delineating the facial phenotype in other conditions diagnosed primarily on the basis of subjective clinical criteria.
Research Interests:
We critique published incidences for fetal alcohol syndrome (FAS) and present new estimates of the incidence of FAS and the prevalence of alcohol-related neurodevelopmental disorder (ARND). We first review ...
Research Interests:
Research Interests:
This paper presents the most recent methodological developments for an approachto modelling nonstationary spatial covariance structure through deformations ofthe geographic coordinate system that was first introduced in a technical report... more
This paper presents the most recent methodological developments for an approachto modelling nonstationary spatial covariance structure through deformations ofthe geographic coordinate system that was first introduced in a technical report 10years ago (Sampson, 1986).We address primarily the problem of estimating the spatial covariance structurein levels of an environmental process at arbitrary locations (both monitored andunmonitored), based on records from N point monitoring sites x 1 ;...
Research Interests:
Research Interests: Cognitive Science, Neuropsychology, Behavior, Drug interactions, Memory, and 16 moreNeurotoxicology, Prospective studies, Pregnancy, Humans, Child, Female, Partial Least Squares, Male, Ethanol, Arousal, Educational Status, Intelligence tests, Neurosciences, Partial Least Square, Psychomotor Performance, and Neuropsychological Tests
Research Interests:
Research Interests:
Research Interests:
OBJECTIVE: To examine the relative importance of prenatal alcohol exposure and family history of alcoholism for the prediction of adolescent alcohol problems.METHOD: In 1974-75, a population-based, longitudinal prospective study of... more
OBJECTIVE: To examine the relative importance of prenatal alcohol exposure and family history of alcoholism for the prediction of adolescent alcohol problems.METHOD: In 1974-75, a population-based, longitudinal prospective study of alcohol and pregnancy began with self-report of alcohol use by pregnant women. In a 14-year follow-up, 439 parents provided information on the family history of alcohol problems for these adolescent offspring. The 14-year-old adolescents provided information on the frequency and quantity of their own alcohol consumption within the past month, on the consequences of their drinking over the past 3 years, and on their age at first intoxication. Additional covariates were assessed prenatally and at follow-up.RESULTS: Prenatal alcohol exposure was more predictive of adolescent alcohol use and its negative consequences than was family history of alcohol problems. Prenatal exposure retained a significant predictive effect even after adjustment for family history and other prenatal and environmental covariates. By contrast, the nominally significant correlation of family history with adolescent drinking is weaker after adjustment for prenatal alcohol exposure and disappears entirely after adjustment for other relevant covariates. We observed no evidence for an interactive effect of fetal exposure and family history in predicting adolescent alcohol use.CONCLUSIONS: Fetal alcohol exposure is a risk factor for adolescent alcohol involvement and alcohol-related problems and may account for variance in prediction of problems otherwise attributed to family history of alcoholism. Studies of alcoholism etiology and family history need to include consideration of even modest levels of fetal alcohol exposure.
Research Interests:
Research Interests: Pathology, Brain Imaging, Sex, Teratology, Statistical Analysis, and 24 moreMagnetic Resonance Imaging, Morphometry, Adolescent, Fetal Alcohol Syndrome, Brain, Pregnancy, Humans, Corpus Callosum, Toxicity, Cerebral Cortex, Female, Alcohol Drinking, Male, Ethanol, Fetal Alcohol Spectrum Disorders, Spectrum, Agenesis of the Corpus Callosum, Theoretical Models, Adult, Sex Factors, Nuclear Magnetic Resonance Imaging, Brain Damage, Brain stem, and Magnetic resonance image
Patients with Fetal Alcohol Syndrome (FAS) and Fetal Alcohol Effects (FAE) often have difficulty functioning appropriately in everyday life and seem to employ poor problem-solving strategies. Tests of executive function are relevant for... more
Patients with Fetal Alcohol Syndrome (FAS) and Fetal Alcohol Effects (FAE) often have difficulty functioning appropriately in everyday life and seem to employ poor problem-solving strategies. Tests of executive function are relevant for quantifying the functional deficits and underlying real-life problems associated with prenatal alcohol exposure. This study considers two pathways for the effects of prenatal alcohol on executive function: a direct effect and an indirect effect through prenatal alcohol's effect on IQ. We compared 30 men who had been diagnosed with FAS or FAE with young adults participating in a longitudinal prospective study (n = 419) and 15 control participants that comprised a comparison group. This study is unique in its analysis of the same battery of assessments of executive function in both a large low dose longitudinal study sample and a clinically diagnosed group. Participants were evaluated on 9 tests (including 58 scores) of executive function. For some but not all of the tests in this executive function battery, the decrement in the alcohol exposure group is greater than would be predicted from their IQ scores. We found that 3 of 6 Stroop scores, 2 of 4 Trails scores, 9 of 16 Wisconsin Card Sorting scores, 1 of 2 Ruff's Figural Fluency scores, and 2 of 4 Consonant Trigrams scores appear to be particularly sensitive to the direct effects of prenatal alcohol damage for patients with FAS and FAE. The findings suggest that these executive function tests would be particularly useful in clinical evaluations of persons suspected of fetal alcohol damage because they would not simply reflect deficits in IQ or facial stigmata.
Research Interests: Psychology, Cognitive Science, Intelligence, Cognition, Prospective studies, and 15 morePregnancy, Humans, Female, Male, Ethanol, Fetal Alcohol Spectrum Disorders, Developmental neuropsychology, Newborn Infant, Facies, Adult, Developmental Neuropsychology, Neurosciences, Predictive value of tests, Case Control Studies, and Neuropsychological Tests
Research Interests: Psychometrics, Behavior, Statistical Analysis, Memory, Toxicity, and 3 moreEthanol, Gestation, and Nino
Research Interests:
Heavy maternal drinking during pregnancy has consistently been linked to decreased intrauterine growth, but the effect of smaller amounts of alcohol is less clear. In this study, the relationship between fetal growth and... more
Heavy maternal drinking during pregnancy has consistently been linked to decreased intrauterine growth, but the effect of smaller amounts of alcohol is less clear. In this study, the relationship between fetal growth and "moderate" drinking by low-risk, nonsmoking prenatal patients is explored. The sample consists of 144 women seen for the first time at the prenatal clinic of University College Obstetrics Hospital, London, England, between July 1979 and May 1980 and meeting the following criteria: white, aged 19-35 years, 8-16 weeks gestation at first prenatal visit, nonsmoker, nonalcoholic, lower middle class or higher, and in general good health. Average daily consumption of 10 g of ethanol (about one drink) in the week prior to recognition of pregnancy is related to a decrease in infant birth weight of 225 g, after adjustment for gestational age, sex of child, and maternal age, weight, height, pregnancy weight gain, social class, gravidity, and parity. In addition, consumption of this amount in the week before first prenatal visit is related to a comparable decrease in birth weight for male but not for female infants. These findings suggest that risk of decreased intrauterine growth begins very early in pregnancy, and that fetal response to later alcohol use may vary with sex of the child.