SPH Syllabus Template 2017-2018 Revised: June 2017; ACAP approved: July 2017; Executive Committee approved: August 2017 Fall 2017 - Biostatistics 523: Statistical Methods for Epidemiology Tuesdays and Thursdays, 1-3pm, 1690 SPH1A (Main Lecture), Mon/Tue/Wed/Fri, 5-6pm, SPH2 G442A (Lab Sections) Professor: Graduate Student Instructors: Zhenke Wu zhenkewu@umich.edu 4623 SPH-I Office Hours: Tuesdays 8:30am-9:30am Chen Liang ccliang@umich.edu Office Hours: Mondays 1:30-2:30pm, M1123 Wednesdays 1:30-2:20pm, M1170 Wed/Fri Lab Instructor Pedro Orozco del Pino porozco@umich.edu Office Hours: Mondays 4-5pm, M1122 Thursdays 11-noon, M4639 SPH I Mon/Tue Lab Instructor Course Description: This course introduces and illustrates the statistical methods commonly used in epidemiologic studies. Emphasis will be placed on organizing data into analyzable forms, choosing appropriate statistical methods and subsequent interpretation for precise scientific communications. The selected topics include probability, measures of association and risk, sample size and power calculations, meta-analyses, matched-design analysis, logistic regression, Poisson regression, survival analysis and regression techniques for correlated outcomes. SAS programs will be used to demonstrate the statistical procedures for analyzing real data. Course Materials: Required: Course Notes Will be posted on Canvas by instructor. Students are responsible for printing out copies as needed. The course notes are the most important resource for learning the material. The materials for the lab sections, including code, data and example outputs will also 1
be accessible via Canvas (https://umich.instructure.com/). Eric Vittinghoff, David V. Glidden, Stephen C. Shiboski, Charles E. McCulloch, Regression Methods in Biostatistics: Linear, Logistic, Survival, and Repeated Measures Models (Statistics for Biology and Health), 2nd Edition, Springer, 2012. Free version available electronically through UM library Springerlink. Additional sources: Harrell Jr., F. E. Regression Modeling Strategies with Applications to Linear Models, Logistic and Ordinal Regression, and Survival Analysis (Springer Series in Statistics), 2nd Edition. 2015. Free version available electronically through UM library Springerlink. David Kleinbaum and Mitchel Klein. Survival Analysis - A Self Learning Text. (Springer Series in Statistics for Biology and Health), 2nd Edition. 2005. Free version available electronically through UM library Springerlink. D.W. Hosmer, S. Lemeshow and S. May. Applied Survival Analysis: Regression Modeling of Time to Event Data (Wiley Series in Probability and Statistics), 2nd Edition. 2008. Kindle version available. D.W. Hosmer, S. Lemeshow and R. X. Sturdivant. Applied logistic regression (Wiley Series in probability and statistics), 3rd Edition. 2013. Free version available electronically through UM library http://onlinelibrary.wiley.com.proxy.lib.umich.edu/book/10.1002/9781118548387?globalmessa ge=0 H. Brown and R. Prescott. Applied Mixed Models in Medicine. Wiley, 3rd Edition. 2015. N.E.Breslow and N.E.Day. Statistical Methods in Cancer Research, Volume 1. 1980. International Agency for Research on Cancer. Available for free download at http://www.iarc.fr/en/publications/pdfsonline/stat/ N.E.Breslow and N.E.Day. Statistical Methods in Cancer Research, Volume 2. 1986. International Agency for Research on Cancer. Available for free download at http://www.iarc.fr/en/publications/pdfsonline/stat/ M. Woodward. Epidemiology : Study Design and Data Analysis (Chapman & Hall/CRC Texts in Statistical Science), 2nd Edition. 2004. Steve Selvin, Practical Biostatistical Methods. Duxbury, 1995. J.J. Schlesselman Case Control Studies. 1982. Oxford. D.G. Kleinbaum, L.L Kupper, and H. Morgenstern Epidemiologic Research. 1982. Van Nostrand Reinhold Company, New York J.L. Fleiss Statistical Methods for Rates and Proportions, 2nd Edition. 1981. Wiley. D.E. Matthews and V.T. Farewell: Using and Understanding Medical Statistics, 2nd ed. 1988. Karger. T. Colton: Statistics in Medicine. 1974. Little, Brown and Co. 2
Pre-requisites: Students are responsible for knowledge of introductory materials on the following topics from Biostat 521 and Biostat 522: Probability (conditional probability, Bayes theorem) Discrete and continuous distributions (binomial, Poisson, normal, chi-square), Statistical estimation (point estimates, confidence intervals), Hypothesis testing (significance level, type I and II errors, power, p-value), Multiple linear regression (model and assumptions, categorical and binary variables, inference about parameters, interactions, transformations, and model diagnostics), and Logistic regression (model and assumptions, categorical and binary variables, inference about parameters, interactions and model diagnostics). Course Goals: After completing the course, the student can ordinarily expect to be able to: 1. Use simple statistical techniques for analyzing matched and unmatched data in contingency tables, such as odds ratio estimation in a single 2 2 table and a series of 2 2 tables, the Mantel-Haenszel test, odds ratio testing in 2 K tables including tests for heterogeneity and tests for trend, and the McNemar test for matched pairs. 2. Understand the role of meta-analysis in combining analysis across different studies. Learn appropriate software for this purpose. 3. Understand diagnostic test inference, including ROC curve analysis. 4. Fit appropriate logistic regression models to data from epidemiologic cohort studies and case-control studies using SAS, interpret regression coefficients in these models, test hypotheses about them, and evaluate the fit of these logistic regression models. 5. Understand when ordinary logistic regression methods should be replaced by conditional logistic regression models, be able to fit conditional logistic regression models using SAS, interpret regression coefficients from these models, test hypotheses about them and evaluate model fit. 6. Fit appropriate Poisson regression models to grouped or ungrouped data from epidemiological cohort studies using SAS, interpret regression coefficients from these models, test hypotheses about them and evaluate the fit of these regression models. Understand the use of offset terms in modeling rates as opposed to counts for this data structure. 7. Understand when it is appropriate to use a negative binomial model for count (or rate) data, fit these models in SAS, interpret regression coefficients from these models, test hypotheses about them and evaluate the fit of these regression models. 8. Understand when it is appropriate to use zero-inflated models for counts (rates), fit these models in SAS, interpret regression coefficients from these models, test hypotheses about them and evaluate the fit of these regression models. 3
9. Use simple statistical methods for analyzing censored survival data, including the Kaplan- Meier estimator and the logrank test. 10. Fit appropriate Cox (proportional hazards) regression models to continuous time data from epidemiological cohort studies using SAS, interpret regression coefficients from these models, test hypotheses about them and evaluate the fit of these regression models. 11. Understand competing risks and common strategies for analyzing this data structure, including cumulative incidence. 12. Understand when and how to model correlated normally distributed outcome data using Mixed Models, fit these models in SAS, interpret regression coefficients from these models, test hypotheses about them and evaluate the fit of these regression models. 13. Understand when it is appropriate to use GEE models for correlated outcome data, fit these models in SAS, interpret regression coefficients from these models, test hypotheses about them and evaluate the fit of these regression models. 14. Fit appropriate multivariate Cox (proportional hazards) regression models to correlated continuous time data in SAS, including recurrent events data (Andersen-Gill model) and clustered data (Lee-Wei-Amato model). The Wei-Lin and Weissfeld marginal model will also be demonstrated for modeling recurrent events data and contrasted with the Andersen-Gill method. 15. Understand how to choose appropriate analyses for an epidemiological case-control study or cohort study and be able to implement them in SAS. 16. Present results of analyses used throughout the course to readers who are familiar with these methods. Competencies: This course strengthens core public health learning experiences in 1. Describing the role biostatistics serves in the discipline of public health. 2. Describing basic concepts of probability, random variation and commonly used statistical probability distributions. 3. Describing preferred methodological alternatives to commonly used statistical methods when assumptions are not met. 4. Distinguishing among the different measurement scales and the implications for selection of statistical methods to be used based on these distinctions. 5. Applying descriptive techniques commonly used to summarize public health data. 6. Applying common statistical methods for inference. 7. Applying descriptive and inferential methodologies according to the type of study design for answering a particular research question. 8. Interpreting results of statistical analyses found in public health studies. 9. Developing written and oral presentations based on statistical analyses for both public health professionals and educated lay audiences. 10. Describing a public health problem in terms of magnitude, person, time and place. 11. Apply the basic terminology and definitions of epidemiology. 12. Calculating basic epidemiology measures. 4
13. Communicating epidemiologic information to lay and professional audiences. 14. Drawing appropriate inferences from epidemiologic data. 15. Evaluating the strengths and limitations of epidemiologic reports. 16. Demonstrating effective written and oral skills for communicating with different audiences in the context of professional public health activities. 17. Applying evidence-based principles and the scientific knowledge base to critical evaluation and decision-making in public health. The following lists the new CEPH competencies that are applicable to this course: Foundational Learning Objectives Profession & Science of Public Health 1. Explain the role of quantitative and qualitative methods and sciences in describing and assessing a population s health 2. Explain the critical importance of evidence in advancing public health knowledge Foundational Competencies Evidence-based Approaches to Public Health 1. Apply epidemiological methods to the breadth of settings and situations in public health practice 2. Select quantitative and qualitative data collection methods appropriate for a given public health context 3. Analyze quantitative and qualitative data using biostatistics, informatics, computerbased programming and software, as appropriate 4. Interpret results of data analysis for public health research, policy or practice Planning & Management to Promote Health 5. Select methods to evaluate public health programs Course Requirements: Evaluations: Software: SAS in the main lectures and labs. There is a free SAS university version for students to download and use. Alternatively, online remote access is available through Virtual Sites (http://virtualsites.umich.edu/) with log-in. Calculator: you should have access to a calculator that can perform basic arithmetic, square roots, logarithms, y x, and e x. We do not have calculators to loan. Lab: Once per week on Monday, Tuesday, Wednesday or Friday. Labs start on the second week of the term. You must register for a specific lab section and should only attend during your scheduled time. Students auditing the course can only use lab space after all registered students have been accommodated. 5
Homework: Students need to submit six homework assignments via Canvas website. The GSIs will grade the submitted homework on the website. Late work will not be accepted. Solutions for problems will be posted on the course website. Submitted homework papers must be written independently. Copying some or all of a homework assignment from someone else or from the web, or allowing your assignments to be copied by someone else, is cheating. If you have questions about this policy please discuss with the instructor. Examination and homework grading and re-grading: Any requests for re-grades should be submitted in writing within one week after the examination or homework is returned. Should you wish to request a regrade, submit your graded material and a written description of the issues to the instructor via email. During any regrade, the entire paper is subject to regrade, so if errors made in your favor in other problems are noted, these may also be corrected. The midterm and final exams will be in-person. Please see detailed time and room below. Conflicts with the times of exams should be reported to the instructor by the end of the first week. Final exam times are fixed by the University registrar and cannot be altered (see http://www.ro.umich.edu/exams/ for the university policy). All the exams are closed book. Two-sided review sheet of your own making is permitted (one for mid-term, two for the final exam). You can bring a calculator, but not a laptop/computer/smartphone or other devices connected to the internet. Midterm Exam: 30% (Thursday October 19th, 1-3pm, Last Name Starting A-P Room 1655 Last Name Starting Q-Z Room 1690 SPH1A) Final Exam: 40%. (Wednesday, December 20th, 4-6pm, Last Name Starting A-P Room 1690 SPH1A Last Name Starting Q-Z Room 1655) Homework: 30% Six homework assignments of equal weights. Total: 100% Classroom Expectations/Etiquette: The faculty of the School of Public Health believes that the conduct of a student registered or taking courses in the School should be consistent with that of a professional person. Courtesy, honesty and respect should be shown by students toward faculty members, guest lecturers, administrative support staff and fellow students. Similarly, students should expect faculty to treat them fairly, showing respect for their ideas and opinions and striving to help them achieve maximum benefits from their experience in the School. Student academic misconduct refers to behavior that may 6
include plagiarism, cheating, fabrication, falsification of records or official documents, intentional misuse of equipment or materials (including library materials), and aiding and abetting the perpetration of such acts. The preparation of reports, papers, and examinations, assigned on an individual basis, must represent each student s own effort. Reference sources should be indicated clearly. The use of assistance from other students or aids of any kind during a written examination, except when the use of aids such as electronic devices, books or notes has been approved by an instructor, is a violation of the standard of academic conduct. Diversity, Equity, and Inclusion: The University of Michigan School of Public Health. The University of Michigan School of Public Health seeks to create and disseminate knowledge with the aim of preventing disease and promoting the health of populations worldwide. We recognize the histories of social discrimination globally, and seek to promote and extend opportunities for members of all groups that historically have been marginalized. We commit to developing the institutional mechanisms and norms necessary to promote the values of diversity, equity, and inclusion, both inside and outside our classrooms. To this end, SPH upholds the expectations that all courses will (1) be inclusive, (2) promote brave discussions, (3) follow multicultural ground rules and (4) abide by UM policies and procedures. 1) Inclusive courses, are those in which teachers and learners co-create and co-sustain environments that support and encourage all members to participate equitably. See http://crlt.umich.edu/multicultural-teaching/inclusive-teaching-strategies for more resources. 2) Brave (rather than safe) discussions promote diversity and social justice learning by acknowledging dynamics of oppression and privilege both inside and outside the classroom. Read more at http://ssw.umich.edu/sites/default/files/documents/events/colc/from-safespaces-to-brave-spaces.pdf. 3) Multicultural ground rules acknowledge diverse experiences in the classroom and offer strategies for holding one another appropriately accountable. See examples from the UM Program on Intergroup Relations and others at http://ncdd.org/rc/item/1505. 4) UM policies and procedures can be found at http://diversity.umich.edu with additional resources and instructions for reporting discrimination at https://sph.umich.edu/diversityequity-inclusion/resources.html. Academic Integrity: The faculty and staff of the School of Public Health believe that the conduct of a student registered or taking courses in the School should be consistent with that of a professional person. Courtesy, honesty, and respect should be shown by students toward faculty members, guest lecturers, administrative support staff, community partners, and fellow students. Similarly, students should expect faculty to treat them fairly, showing respect for their ideas and 7
opinions and striving to help them achieve maximum benefits from their experience in the School. Student academic misconduct refers to behavior that may include plagiarism, cheating, fabrication, falsification of records or official documents, intentional misuse of equipment or materials (including library materials), and aiding and abetting the perpetration of such acts. Please visit https://sph.umich.edu/student-resources/mph-mhsa.html for the full Policy on Student Academic Conduct Standards and Procedures. SPH Writing Lab: The SPH Writing Lab is located in 5025 SPH II and offers writing support to all SPH students for course papers, manuscripts, grant proposals, dissertations, personal statements, and all other academic writing tasks. The Lab can also help answer questions on academic integrity. To learn more or make an appointment, please visit the SPH writing lab website. Student Well-Being: SPH faculty and staff believe it is important to support the physical and emotional well-being of our students. If you have a physical or mental health issue that is affecting your performance or participation in any course, and/or if you need help connecting with University services, please contact the instructor or the SPH Office for Student Engagement and Practice. Please visit https://sph.umich.edu/student-life/wellness.html for information on wellness resources available to you. Student Accommodations: Students should speak with their instructors before or during the first week of classes regarding any special needs. Students can also visit the SPH Office for Student Engagement and Practice for assistance in coordinating communications around accommodations. Students seeking academic accommodations should register with Services for Students with Disabilities (SSD). SSD arranges reasonable and appropriate academic accommodations for students with disabilities. Please visit https://ssd.umich.edu/topic/our-services for more information on student accommodations. Students who expect to miss classes, examinations, or other assignments as a consequence of their religious observance shall be provided with a reasonable alternative opportunity to complete such academic responsibilities. It is the obligation of students to provide faculty with reasonable notice of the dates of religious holidays on which they will be absent. Please visit http://www.provost.umich.edu/calendar/ for the complete University policy. Course Topics/Reading List: Upon entering the course students will be responsible for knowledge of introductory materials on the following topics: probability (conditional probability, Bayes theorem), discrete and continuous distributions (binomial, Poisson, normal, chi-square), statistical estimation (point estimates, confidence intervals), hypothesis testing (significance level, type I and II errors, power, p-value), multiple linear regression (model and assumptions, categories and dummy variables, inference about parameters, interactions, transformations, and model diagnostics) 8
and logistic regression (model and assumptions, categories and dummy variables, inference about parameters, interactions and model diagnostics). Date Outline of Topics Lecture hours Sept 5. Review of Fundamental Concepts, Analysis of Normally Distributed Outcomes 2 Sept 7. Review of Proportions, Contingency Table Analysis, Trend Tests, Meta-Analysis 4 Sept 14. Diagnostic Tests, ROC Curves 2 Sept 19. Matched Outcome Analysis 2 Sept 21. Review of Logistic Regression for Modeling Binary Outcomes, Model Building Techniques 5 Sept 28. Missing Covariate Data, PROC MI, PROC MIANALYZE 1 Oct 3. Modeling Counts (Poisson Regression) 2 Oct 5. Modeling Overdispersed or Zero-inflated Counts 2 Oct 10. Introduction to Time-to-Event Data 2 Oct 12. Review for Midterm Exam, Exam 1 Materials 2 Oct 17. Break Oct 19. Midterm Exam 2 Oct 24. Kaplan-Meier Survival Estimation, Two-sample Tests for Survival Data (& Review) 4 Oct 31. Introduction to Cox Proportional Hazards Model 2 Nov 2. Cox Model (Model Selection, Diagnostics) 6 Nov 14. Introduction to Dependent Outcomes (Correlation 9
and Covariance Structures, Random Intercept) 2 Nov 16. Dependent Outcomes, Cont. (Cross-over Trials) 2 Nov 21. Dependent Outcomes, Cont. (Longitudinal Trends) 4 Nov 23. Break Nov 30. Modeling Dependent Count Outcomes via GEE 2 Dec 5. Modeling Dependent Binary Outcomes, GEE, Conditional Logistic Regression) 2 Dec 7. Modeling Dependent Survival Outcomes (Overview) 2 Dec 12. Review for Final Exam 2 NO CLASS: Tuesday October 17th Thursday November 23th (Fall Study Break) (Thanksgiving Break) Summary of Important Dates: Date Details Mon Sep 25, 2017 Homework 1 due by 11:59pm Fri Oct 6, 2017 Homework 2 due by 11:59pm Wed Oct 18, 2017 Homework 3 due by 11:59pm Thu Oct 19, 2017 BIOSTAT523 Mid-term 1pm to 3pm Wed Nov 8, 2017 Homework 4 due by 11:59pm Wed Nov 22, 2017 Homework 5 due by 11:59pm Mon Dec 11, 2017 Homework 6 due by 11:59pm Wed Dec 20, 2017 BIOSTAT523 Final Exam 4pm to 6pm 10