1 Week Courses
A laptop computer is essential and required for enrolled students (tablets or other similar devices are also suitable). We recognize computers to be an extension of the learning tools needed to successfully participate in our courses. Coursepacks for most courses will be available digitally via the University's Canvas course learning management system.
EPID 707 Nutritional Epidemiology
(1 credit hour) Eduardo Villamor
This course focuses on the design, conduct, analysis, and interpretation of epidemiologic studies addressing diet and health. The course will review quantitative methodological issues including dietary assessment methods, sources of variation in the diet, energy intake, measurement error, anthropometry and body composition, and biomarkers of intake. Students will advance their knowledge in nutrition research from a population perspective and gain practical experience in the analysis and interpretation of real dietary data. Prerequisites: Introductory course in epidemiology that covers study designs and measures of disease frequency and association. Introductory course in biostatistics that covers correlation and basic regression (e.g. linear and logistic regression). Knowledge of nutrition is desirable but not required. Syllabus for Epid 707
EPID 708 Machine Learning for Epidemiologic Analysis in the Era of Big Data
(1 credit hour) Alan Hubbard
The course focuses on advances in machine learning and its application to causal inference and prediction via a so-called Targeted Learning approach. These techniques allow the use of machine learning algorithms not just for prediction, but for estimating so-called causal parameters, such as average treatment effects, direct and indirect effects, dynamic treatments, optimal treatment regimes, etc. Targeted Learning provides the theoretical framework for deriving substitution estimators and rigorous statistical inference. Such techniques will become increasingly important in the era of Big Data, and the course will focus on the implementation of these techniques on existing data. The course will have a computer lab portion based on the R programming. Time permitting, we will discuss the implementation of approaches via cloud computing. This course is targeted towards more methods oriented epidemiologists and students. Prerequisites: Some background in R (can be obtained through EPID798 in this summer session) and working knowledge of regression and other standard statistical methodology common in basic epidemiological analysis. Syllabus for EPID 708
EPID 716 Clinical Epidemiology and Evidence-Based Research
(1 credit hour) Mitchell A.H. Levine
With the increasing demand for an evidence-based approach in the delivery of healthcare services and the economic pressures for a more rational and efficient use of limited healthcare resources, practitioners and administrators in the healthcare field need to develop clinical measurement and evaluative skills in order to conduct their work optimally. Clinical Epidemiology and Evidence-Based Research identifies and teaches these skills. The course will cover the basic concepts of clinical epidemiology in the context of appraising the recent medical literature pertaining to issues of causation, diagnosis, management, and economic evaluation. The format will include problem-based learning. All health professionals (clinicians and administrators) who rely on the medical literature to guide their activities are invited to attend the course. No prerequisite.
Syllabus for Epid 716
EPID 717 Design and Conduct of Clinical Trials
(1 credit hour) Stephen J. Farish
The theoretical and practical challenges to be considered in designing and conducting a randomized clinical trial will be presented. Topics to be discussed include the specification of a primary objective, adherence to accepted ethical guidelines, the role of randomization and the means of its implementation, the choice of design strategy and design strengthening features, considerations involved in sample size determination and patient recruitment and standards for reporting clinical trials. Detailed analytic issues will be considered in the complementary one-week course that follows. Prerequisite: Introductory course in epidemiology.
EPID 718 Analysis of Clinical Trials
(1 credit hour) Stephen J. Farish
Methods of analysis appropriate to various designs, such as cross-over designs, nested designs, factorial designs, and designs with repeated measures will be presented. The use of GLM techniques for analysis will also be illustrated. Topics will include estimation of survival functions, survival comparison between groups of subjects, identification of important covariates, adjustment for covariates, testing for interaction, and understanding the difference between confounding and interaction. Specific tools to be discussed include the Kaplan-Meier estimators, the log-rank (Mantel-Haenszel) statistics, and the Cox proportional hazards model. Instruction will focus on the empirical use of methodologies rather than formal algebraic knowledge. Practical applications of manual and software-based analysis will illustrate specific procedures and interpretation of results. Students receive a disk with the data and analysis programs for all examples in the course. Students are advised to bring a scientific calculator. Prerequisite: Introductory course in biostatistics.
EPID 719 Methods in Genetic and Epigenetic Epidemiology
(1 credit hour) Edward Ruiz-Narvaez
This course familiarizes students with general methods and principles of genetic and epigenetic epidemiology. The course seeks to integrate concepts in human genetics, population genetics, epidemiology, and biostatistics. The course will emphasize the practical applications of existing methods, which requires a critical evaluation of the scientific literature. Students are expected to be active participants in the course. Some of the topics to be included are population genetics, genetics of common diseases, gene-environment interaction, genetic and epigenetic association studies, and social epigenomics. Prerequisites: Introductory level course in epidemiology that covers study designs and measures of disease frequency and association. Introductory level course in biostatistics that covers correlation and basic regression (e.g. linear and logistic regression) Syllabus for Epid 719
EPID 722 Medical Product Epidemiology and Global Regulation
(1 credit hour) Judith J.K. Jones
Almost all medical products- pharmaceuticals, biologics, vaccines, devices and other medical products, are highly regulated worldwide. This has stimulated the need for data and varied studies on very large populations to establish the safety of the products. This course will cover the epidemiologic methods to study the use and effects of medical products within the context of regulatory requirements for safety and appropriate use. In particular, it will cover spontaneous reporting systems, ad hoc epidemiologic studies and the growing use of large automated healthcare databases (big data). Emphasis will be placed on the need to quantify the frequency of events in the different types of products and the risk factors that predispose to them. We will also address the more recent implementation of pharmaceutical risk management plans, and the implementation of different safety monitoring systems by regulators such as the FDA and the European Medicines Agency (EMA). Other topics include differences in epidemiology of the different medical products and methods to measure the frequency of their use. The courses will include lectures, group exercises to understand the stages from signal/hypothesis to study protocol development, and will finish with participants' presentations to a hypothetical regulator. Prerequisite: Introductory level course in epidemiology. Syllabus for Epid 722
EPID 742 Introduction to Generalized Linear Models
(1 credit hour) Hyungjin Myra Kim
The vast majority of epidemiologic data involve either binary or count data. Logistic (binary data) and Poisson (count data) regression analyses are two important analytic approaches that frequently provide valuable insights into collected data. Both techniques will be presented in a practical and an applied fashion. The discussed material begins with the simplest case with the goal of understanding the fundamental properties of each model. Once these properties are established, more advanced topics such as collinearity, variable selection, non-linear explanatory variables and goodness-of-fit will be described and applied to several multivariable epidemiologic data sets. These two analytic approaches not only provide simple and effective ways to explore complex relationships but illustrate the general process of using a linear model to draw conclusions from the analysis of epidemiologic data. Prerequisites: Introductory level courses in epidemiology and biostatistics.
EPID 747 Successful Scientific Writing
(1 credit hour) Paul Z. Siegel, John Iskander
This course takes an active, participatory approach to help public health and health care professionals learn how to communicate the findings of their research and investigations more effectively and expedite publication of their manuscripts. Working in small groups, students spend much of their class time critiquing actual published and unpublished manuscripts, including their own and solving a wide range of exercises that exemplify the real-world challenges that authors face. Free-form in-class discussions make it possible for class members to learn from one another's experiences. Major components of the course include the following: basic sections of a scientific article: the purpose, elements and organization of each section; principles of style for writing in public health and epidemiology; systematic approaches to the process of writing and publishing an article in a peer-review journal; and effective strategies for dealing with requests of journal editors and reviewers. No prerequisite. Syllabus for EPID 747
EPID 757 Systematic Reviews and Meta-analyses
(1 credit hour) Joel Gagnier
Systematic reviews and meta-analyses are useful for evidence-based clinical and public health practice. The widespread and growing application of systematic review methods for the synthesis of evidence on important or pressing research and clinical questions underscore the need for health-care professionals to understand and critique this research design. This course will provide a detailed description of the systematic review process, discuss the strengths and limitations of the method, and provide step-by-step guidance on how to actually perform a systematic review and meta-analysis. Specific topics to be covered include formulation of the review question, searching of literature, quality assessment of studies, data extraction, meta-analytic methods, assessment of heterogeneity and report writing. The course will also cover statistical issues such as selection of statistical models for meta-analysis, practical examples of fixed and random effects models, best evidence syntheses (qualitative systematic reviews) as well as examples of methods to evaluate heterogeneity and publication bias. STATA statistical software will be used to perform meta-analysis during the computer lab, along with tutorials on how to effectively use tools such as PubMed for conducting reviews. Prerequisites: Introductory level courses in epidemiology and biostatistics are recommended.
EPID 761 Social Determinants of Population Health
(1 credit hour) Carlos Mendes de Leon
This course will provide an introduction to current concepts and research in the social determinants of population health. We will develop a basic understanding of how key social factors shape the distribution of health and disease in the general population, with a focus on the social gradients in health, health disparities across racial and ethnic groups, and the role of the social environment in population health outcomes. In addition, we will review a few key psychological and biological processes that are thought to mediate or interact with social determinants in producing health differences within and between populations. Finally, we will discuss the implications of the social determinants of population health with regard to health and health care policies. Prerequisite: Introductory level course in epidemiology.
EPID 762 Using STATA (or SAS) with Complex Survey Data to Examine Associations and Time Trends
(1 credit hour) Donna Brogan
Many health surveys in the U.S. (e.g. NHANES, NHIS, BRFSS) and globally (e.g. DHS) publicly release their data. Analysts who wish to make inference back to the target population must use specialized statistical software that recognizes common and nonstandard characteristics of these complex survey data.Course participants will conduct descriptive and association analyses, trend analyses over time, and logistic regression analyses with the statistical software package STATA using two complex survey datasets: the 2013 NHIS (National Health Interview Survey) and six years of BRFSS (Behavioral Risk Factor Surveillance System) data for one state.For those with no or limited STATA experience, a Sunday session will be held the day before the course begins to review selected commands in STATA. A complimentary version of STATA can be downloaded to each participant's computer for use during the course. Although instruction will focus on STATA's capabilities for analysis of complex survey data, participants already familiar with the SAS survey procedures may wish to conduct some of the course analyses also in SAS.
Prerequisites: Introductory courses or experience in epidemiology, logistic regression, and use of a statistical software package such as STATA, SAS or R for data management and data analysis. Background in the theory and/or practice of selecting probability samples is helpful but not required.Many health surveys in the U.S. (e.g. NHANES, NHIS, BRFSS) and globally (e.g. DHS) publicly release their data. Analysts who wish to make inference back to the target population must use specialized statistical software that recognizes common and nonstandard characteristics of these complex survey data. Syllabus for Epid 762
EPID 766 Analysis of Longitudinal Data from Epidemiologic Studies
(1 credit hour) Daowen Zhang
It has been popular in epidemiology to conduct longitudinal studies where study participants are followed over time and repeated measurements of interest are obtained. Compared to traditional cross-sectional or case-control studies, longitudinal studies can be more efficient to detect a difference of interest, offer more evidence for possible causal inference, etc. However, longitudinal data are likely to be correlated, which presents a substantial challenge in analyzing such data. This course will address 1) epidemiologic methods for the design and interpretation of longitudinal studies involving repeated measures and 2) statistical methods appropriate for longitudinal data including generalized estimating equations (GEEs), linear mixed models and generalized linear mixed models. A series of studies will be used to illustrate the major design issues and statistical approaches. Relevant procedures in statistical package SAS will be introduced and appropriate interpretation of results will be emphasized. Prerequisites: Students are expected to have one or two graduate biostatistics courses on (simple and multiple) linear regression models, categorical data analysis such as logistic regression models and experience of conducting data analysis using statistical software SAS.
EPID 777 Geographic Information Systems for Epidemiology
(1 credit hour) Shannon J. Brines
Geographic Information Systems (GIS) are used for displaying and analyzing spatial data. Data from a variety of sources may be compared utilizing overlay analysis and spatial statistics. Modern tools permit novice GIS users to perform spatial analysis without extensive training. This course will introduce students to ArcGIS, the world's leading GIS analysis package. Examples of epidemiological applications will give students the opportunity to see and use this powerful tool. Some of the topics to be covered are data import/export, layering, data table management, classification, labeling, spatial and attribute queries, and buffer analysis. No prerequisite. Syllabus for EPID 777
EPID 778 Spatial Statistics for Epidemiological Data
(1 credit hour) Veronica Berrocal
With the increasing availability of geographic information systems, spatial data have become more frequent in many disciplines, including public health and epidemiology. This course aims to provide an introduction to spatial statistical methods for epidemiological data, covering modeling approaches for the two different types of spatial data: point-referenced data, where the geographical coordinates of the observations have been recorded; and areal-averaged data, where summary statistics (e.g., number of disease cases by county, zip code, etc.) are reported for each areal unit. Topics covered include exploratory analysis for spatial data, covariance functions, kriging, spatial regression; disease mapping, spatial smoothing; point processes, assessment of clustering, and cluster detection. Each lecture will feature a lab component, during which spatial analyses of datasets, made available to the participants, will be performed using the publically available R statistical software (downloadable to your laptops at www.r-project.org). Although previous experience with R is preferred, it is not required. Prerequisite: Course in basic statistics (e.g., EPID 701).
EPID 780 Applied Epidemiologic Analysis for Causal Inference
(1 credit hour) Jay Kaufman
This course focuses on regression models of potential outcomes for the estimation of causal parameters in epidemiologic research. One day of theoretical background will be followed by 4 days of lectures on a range of techniques along with a computer laboratory portion with hands-on exercises using the Stata statistical software package (V13 or higher). Techniques covered include propensity score matching, inverse probability weighting for confounder control and for censoring, marginal structural models, mediation, the parametric g-formula, and some introduction to econometric techniques such as differences-in-differences. Emphasis is on understanding the causal models, generating analysis with software code, and interpreting the resulting estimates. Prerequisites: Students should have at least one basic epidemiology course and some background in Stata, along with a working knowledge of regression and other standard statistical methodology common in basic epidemiological analysis.
EPID 783 Methods in Community-Based Participatory Research for Health
(1 credit hour) Barbara A. Israel, Wilma Brakefield-Caldwell, J. Ricardo Guzman, Edie Kieffer, Toby Lewis, Gloria Palmisano, Angela Reyes, Zachary Rowe, Amy Schulz
(Co-taught by other faculty and community partners involved in the Detroit Community-Academic Urban Research Center (http://www.detroiturc org/) and its affiliated CBPR projects) There is increasing recognition and support for more comprehensive and participatory approaches to research and interventions in order to address the complex set of determinants associated with public health problems that affect populations generally, as well as those factors associated with racial and ethnic disparities in health more specifically. Community-based participatory research (CBPR) is one such partnership approach that equitably involves all partners in all aspects of the research and intervention process, aimed at both increasing knowledge and understanding and linking the knowledge gained with interventions and policy change to enhance the health and quality of life of community members.
This course will provide an introduction to some of the core principles, concepts and methods involved in using a CBPR approach. Organized along the phases of CBPR, this course will focus on describing and understanding partnership formation, maintenance, and evaluation; the use of quantitative and qualitative methods (e.g., survey, focus group interview, observational checklist) for the purposes of community assessment, examining basic research questions, and developing and evaluating interventions; and feedback, interpretation, dissemination and application of research results. The course will examine the rationale for, benefits of and challenges associated with using a community-academic partnership approach to research and interventions. Class format includes lectures, discussions, case studies, and small group exercises. No prerequisite. Syllabus for EPID 783
EPID 784 Survival Analysis Applied to Epidemiologic and Medical Data
(1 credit hour) Chris Andrews
The primary objective of this course is to provide participants with the background required to understand commonly used survival analysis methods and to apply such methods using standard statistical software. The course material relies heavily on examples and intuitive explanations of concepts. The mathematical level is completely accessible with knowledge of high school algebra, one semester of calculus, and a one-year course in basic statistical methods. Examples will be chosen from various epidemiologic and medical applications. The topics will include: an introduction to survival analysis; right censoring and left truncation; life tables, non-parametric estimators (e.g., Kaplan-Meier, Nelson-Aalen); two- and k-sample tests (e.g., log-rank, Wilcoxon); parametric methods for analyzing survival data (e.g., exponential model); semi-parametric methods (e.g., Cox proportional hazards model). The statistical techniques will be illustrated using various medical and epidemiological studies. Students will carry out some applied (pencil-and-paper) problems to illustrate the main ideas of survival analysis and to solidify the concepts. There will also be a number of data analysis exercises that will utilize statistical software. Prerequisite: Introductory level course in statistics (including an introduction to regression methods).
EPID 787 An Introduction to Multilevel Analysis in Public Health
(1 credit hour) Katherine M. Keyes
Multilevel analysis is an essential analytic tool in epidemiology and public health that allows the simultaneous investigation of the effects of factors defined at multiple levels on individual-level outcomes. This short course will review the rationale for multilevel analysis in public health research, build the theory and practice of these models from the fundamentals of the statistical approach and demonstrate a variety of different forms that the models can take. Fitting and interpreting models will demonstrate using Stata 13 statistical software. Special emphasis will be placed on the strengths and limitations of multi-level analysis in investigating social and group-level determinants of health. Prerequisites: Knowledge of basic epidemiology and linear and logistic regression. Syllabus for EPID 787
EPID 793 Complex Systems Modeling for Public Health Research – FULL DAY COURSE
(2 credit hours) Kristen Hassmiller Lich
This course will provide an introduction to three major complex systems science modeling techniques with wide applicability to public health. We will cover an introductory overview of each technique, examples of applications, brief discussions of best practices, and some initial hands-on lab experience. At the completion of the course, the student will be able to explain current and potential future roles of complex systems science in public health and describe each of the three approaches and their respective advantages/disadvantages. Students will be well positioned to further explore the incorporation of systems science methods into their own research or participation in interdisciplinary teams using the modeling techniques. Prerequisite: Introductory level course in epidemiology. Syllabus for EPID 793
EPID 798 Epidemiologic Data Analysis using R
(1 credit hour) Sung Kyun Park
This course will introduce the R statistical programming language for epidemiologic data analysis. R is freely available, versatile, and a powerful program for statistical computing and graphics. This course will focus on core basics of organizing, managing, and manipulating data; basic graphics in R; descriptive methods and regression models widely used in epidemiology. Prerequisites: Introductory level courses in epidemiology and biostatistics are required. Prior experience of R is not required. Syllabus for EPID 798