Environmental Exposure Assessment - ENVR 8940-10



Quantitative Methods for Exposure Science: ENVR 769

(original title “Advanced methods of exposure assessment”)

Spring semester, 2010. (3 credit hours)

Tuesday/Thursday 11:00 – 12:15 pm

Instructors: Joachim D. Pleil, Ph.D., Research Physical Scientist, U.S. EPA,

Adjunct Associate Professor, ESE, SPH, UNC

Jon R. Sobus, Ph.D., Research Physical Scientist, U.S. EPA,





Prerequisites: BIOS 600, BIOS 662 (or equivalent); BIOS 511 (or familiarity with SAS proc reg, proc glm, proc mixed); any two of the following: ENVR 430, ENVR 470, ENVR 707, ENVR 740, ENVR 770, ENVR 890-003, ENVR 890-006, ENVR 890-010.

Permission of instructors will override class prerequisites for qualified students.

It is expected that students have access to the UNC on-line library system and SAS software (available for free from the University).

Course description: Accurate and complete exposure data provide the independent variables against which adverse health outcomes are measured. Time and spatial trends in exposure measures provide the means to test the effect of implemented prevention strategies. Regrettably, the development and interpretation of such data is also considered the most difficult and most often overlooked component of environmental health research.

The course material develops the mathematical approaches for assessing environmental and/or occupational exposures to chemicals in human populations. There are three fundamental tools for interpreting exposure data: stochastic (group) statistics, regression analysis and modeling, and pharmacokinetic modeling. Each is valuable and provides specific information useful for developing mitigation strategies, assessing risk, and linking exposure to health outcome. We will demonstrate these techniques using real-world data sets drawn from ongoing EPA and academic exposure science research. Measurements will include both external (air, water, soil, etc.) concentrations as well as internal (blood, breath, urine, etc.)

This course is intended to develop rigorous mathematical approaches for environmental and occupational exposure assessment. The target audience includes ENVR and EPID graduate students, as well as students pursuing MPH, MSPH, and other public health, medical, or health related graduate degrees, who wish to develop quantitative skills in assessing exposure measurement and biomarker data.

The textbook is “Quantitative Exposure Assessment” by S. M. Rappaport. Additional readings will be assigned from the peer reviewed literature to serve as a basis for discussions. There will be one “take home” mid-term exam and one “take home” final exam that will provide 40% of the grade (each), the remaining 20% will be based on class participation and occasional short assignments.

All work is expected to be independent; there are no group efforts or collaborations allowed. Class attendance and participation are highly recommended for successful performance because tests will rely heavily on concepts developed during lectures.

Class lecture topics: The following list is subject to change to accommodate potential guest lectures, current environmental events, and late breaking exposure assessment research.

1. Introduction to quantitative exposure science class: logistics, expectations, tests, attendance, homework, and grades. Course Overview: sampling exposures; exposure distributions; determinants of exposure levels; exposure variability within and between persons; mixed models of exposure; biomarker of exposure; statistical models for exposure-biomarker relationships; exposure, dose, and damage; classical toxicokinetic models for biomarkers; physiologically-based toxicokinetic models for biomarkers; exposure reconstruction methods.

2. Embark on general concepts of exposure science. Introduce swing diagram: sources, pathways, and routes of human exposure to environmental chemicals. Discuss occupational vs. environmental exposures and acute vs. chronic exposures. Discuss samples of data; stationary samples, personal samples, random samples, repeated samples. Assign readings.

3. Discussions on sampling and analysis: typical sampling systems for environmental media, instrumentation for organics: GC-MS and LC-MS. Applications to various environmental analytes (e.g., volatile organic compounds [VOCs] and polycyclic aromatic hydrocarbons [PAHs]). Discuss sensitivity, specificity, LOQ, LOD, QA, calibration, regression curves, confidence bands, internal/external standards interference and artifacts. Compare assay variability with exposure variability. Assign readings.

4. Discussion of exposure distributions. Introduce the lognormal distribution. Review of statistical concepts (i.e., mean, median, mode, variance, standard deviation, correlation coefficients) and underlying assumptions for summary statistics. Introduce geometric mean and geometric standard deviation. Discuss ANOVA, linear regression, confidence intervals, prediction bands. Homework assignment using excel for summary statistics.

5 Introduction and demonstration of SAS software: importing data from and exporting data to excel; manipulating data in SAS; evaluating distributions using SAS; calculating descriptive statistics using SAS; generating preliminary plots, tables, and figures using SAS. Homework assignment using SAS for summary statistics.

6 Using SAS to perform simple regression analysis and multivariable regression analysis (Proc GLM). Discuss the interpretation of parameter coefficients, parameter confidence estimates, and significance values.

7 Introduce data set from exposure chamber experiments with diesel exhaust. Interpret environmental measures, patterns and correlations. Build multivariable regression models for environmental measures using covariates. Assign SAS homework using data set.

8 Introduction of exposure variability within and between persons. Discuss the concepts of repeated measures, independence, and autocorrelation. Discuss the importance of between-person variability. Introduce the one-way random effects model. Discuss assumptions of random effects models. Discuss relative measures of variability. Discuss sources and ranges of variability in environmental and occupational exposures. Homework assignment: use SAS and excel to generate ANOVA estimates of within- and between-person variance components from a random-effects model.

9 Mixed models of exposure. Introduction to mixed models using matrix notation. Discuss assumptions of mixed-effects models. Introduction to Proc MIXED in SAS. Examination of covariance and correlation matrices. Discuss the use of profile plots to examine autocorrelation. Compare ANOVA estimates from SAS and excel to REML estimates from proc MIXED.

10 Fitting linear mixed-effects models: estimating variance components across groups and estimating model parameters. Methods of constructing models with multiple covariates and interaction terms. Examining model fit using diagnostic procedures; AIC, BIC, normality of predicted random effects, normality of residuals. Introduce dataset and assign homework.

11 Review for Midterm exam. Assign exam at end of class.

12 Midterm exam help session - general explanations and discussions

13 Midterm exam turned in beginning of class. Answer general questions.

14 Field trip to NHEERL environmental chambers.

15 Introduce concept of exposure-biomarker linkages (revisit the swing diagram). Use general linear models (proc GLM) to understand relationships between exposure measurements and biomarkers (non-repeated measures). Use PAH examples of single biomarker measurements (Occ. Envir. Med.). Assign homework.

16 Using mixed models to understand exposure-biomarker relationships when repeated measurements are made. Introduce road-pavers’ study. Using time plots and covariance matrices to evaluate the effects of autocorrelation. Discuss model development and interpretation. Assign homework.

17 Reintroduce the concepts of exposure, dose, and damage. Discuss the compartmental concept (ADME). Develop simple one and two compartment classical pharmacokinetic models. Evaluate these models using empirical data. Assign homework.

18 Develop classical toxicokinetic models for temporal biomarker data, generalize to complex exposure scenarios, exploit differential and difference equations, establish iterative solutions methods. Review previous homework, assign new homework.

19 Introduce physiologically-based toxicokinetic (PBTK) models. Discuss the history and purpose of these models. Draw box diagrams, explain parameters of the models (i.e., blow flow, tissue volumes, partition coefficients, metabolic parameters), and write differential equations. Introduce modeling software for PBTK modeling.

20 Introduce dataset for PBTK models. Give students model parameters and have them write code. Introduce simulated exposure data. Have students program the exposure data with timing commands. Have students optimize model parameters and perform a sensitivity analysis. Assign homework.

21. Introduce concepts, methods, and models for exposure reconstruction. Demonstrate with incremental models using MTBE, TCE and isoflurane. Assign homework.

22. Demonstrate mathematical tools for complete and practical data development sequence: sampling, calibration, analysis, data reduction, data interpretation of blood and urinary biomarkers.

23. Continue practical applications – develop methodology for assessing urinary biomarkers, creatinine correction, excretion rates, metabolic rates, etc. Assign homework.

24. Continue practical applications – develop methodology for assessing linkage between native compounds in environment and blood. Calculate biologically available dose. Assign homework.

25. Continue practical applications – develop methodology for incorporating measured biomarkers (metabolites) from blood and urine into PK and stochastic models, estimate risk factors.

26. Continue practical applications – develop mitigation/regulatory strategy based on practical application.

27. Review for Final exam. Assign exam at end of class.

28. Final exam help session

29. Exam due at beginning of class; answer general questions; class evaluations.

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download