Dr. David Mendez is Core Lead for the Career Enhancement Core (CEC) and Project Lead for Research Project 2. He is also a member of the CAsToR Steering Committee. Dr. Mendez is an Associate Professor in the Department of Health Management and Policy at the University of Michigan. His research focuses on modeling trends of cigarette smoking cessation or switching to e-cigarettes. Dr. Mendez’s research also investigates the financial implications of these trends, with a specific focus on tobacco control in the United States.

This module will:
1) Expose you to different approaches to time series modeling
2) Help you understand the underlying assumptions behind different time series modeling

Huann-Sheng Chen, PhD, is a mathematical statistician and program director in the Statistical Research and Applications Branch (SRAB) of the Surveillance Research Program, Division of Cancer Control and Population Sciences at the National Cancer Institute (NCI). Dr. Chen's research interests include biostatistics, statistical genetics and genetic epidemiology. At SRAB, he has been involved in research on genetics simulation, reporting delay modeling of cancer incidence rates and real time reporting, age-period-cohort model for cancer trends, and cancer incidence and mortality projections in national and local populations. In addition, he has been working on further developments of the Joinpoint model methodology. Dr. Chen received a Doctor of Philosophy degree in statistics from the University of Illinois at Urbana-Champaign and a bachelor’s degree in mathematics from the National Taiwan University.

This module will teach you:
1) Understanding the use of Joinpoint model to analyze trends
2) Selection of joinpoints, estimation, and interpretation
3) Annual percent change, average annual percent change and their confidence intervals
4) Model selection procedures
5) Clustering of similar trends
6) Jump model for data with coding change

Dr. Jihyoun Jeon is Project Lead for the Data Analysis and Dissemination Core (DAD). She is an Assistant Research Scientist in the Department of Epidemiology at the University of Michigan School of Public Health and a member of the Rogel Cancer Center’s Cancer Epidemiology and Prevention Program. Dr. Jeon has been an instrumental member of several consortia focused on developing modeling tools to characterize cancer risk, such the Lung Cancer group of the NCI consortium ‘Cancer Intervention and Surveillance Modeling Network (CISNET)’ and the Colorectal Transdisciplinary (CORECT) Study in the Genetic Associations and Mechanisms in Oncology (GAME-ON). Her modeling expertise incorporates risk prediction and cancer etiology to evaluate screening methods to effectively reduce cancer incidence.

The hands-on part of the module will teach you how to:
1) Perform a trend analysis using the Joinpoint Trend Analysis software developed by the NCI
2) Identify trend change points which break the trend into distinct periods
3) Compare trends between different subpopulations

Note:
The Joinpoint analysis that we will run during this workshop is quite straightforward, so there are no required prerequisites or tutorials for the Joinpoint portion of this workshop.

Dr. Theodore Holford is Project Lead of Data Analysis and Dissemination Core (DAD), Co-Investigator for Research Project 1, and Co-Investigator for Research Project 2, as well as a member of the CAsToR Steering Committee. He is the Susan Dwight Bliss Professor of Public Health at the Yale University School of Medicine, as well as a member of the Yale Cancer Center and Yale Institute for Global Health. Dr. Holford has worked extensively in methodology development, specifically to evaluate trends using age-period-cohort models and inform intervention strategies at the population level.

This module will teach you how to:
1) Express temporal trends from an age, period, and cohort perspective.
2) Parameterize time in a statistical model.
3) Partition temporal trends into linear and curvature effects.
4) Model trends of ever smoking prevalence.
5) Model trends in smoking initiation and cessation probabilities.
6) Combine the contributions of initiation and cessation to quantify smoking exposure.
7) Quantify the effect of smoking on population health.

Training: PhD, Sociology, University of Toronto; M.A. Criminology, University of Toronto; MA Sociology (specialization program and policy evaluation, University of Western Ontario; BA Sociology and History, University of Western Ontario

Research Focus: I am broadly interested in the intersection between criminology and public health, and social problems related to mental health, substance abuse, their overlap, and other risky and illicit behaviours. Three overarching themes guide my current research agenda: (1) the study of child and adolescent mental health using clinical and population-based data, (2) the study of hard-to-reach and at-risk populations, and (3) the use of population-based data to understand the prevalence and risk factors associated with risky and dangerous behaviours. I have an emerging interest in criminological and public health problems related to technological innovations and social media use, and am keenly interested in examining the spatial and temporal processes associated with these problems.

The hands-on part of the module will teach you how to:
1) create censor and duration variables;
2) restructure data to create an unbalanced person-period data set;
3) estimate discrete time models;
4) incorporate replicate weights.

Required software:
STATA will be installed on your workstation during the workshop.

Note:
It is not required that workshop participants have a background in Stata, as it will be relatively simple to take the logic of Dr. Cook’s practical example and then to apply it to the programming software of your choice. However, if you are interested in learning more about Stata, Dr. Cook has recommended several resources.

Dr. David Mendez is Core Lead for the Career Enhancement Core (CEC) and Project Lead for Research Project 2. He is also a member of the CAsToR Steering Committee. Dr. Mendez is an Associate Professor in the Department of Health Management and Policy at the University of Michigan. His research focuses on modeling trends of cigarette smoking cessation or switching to e-cigarettes. Dr. Mendez’s research also investigates the financial implications of these trends, with a specific focus on tobacco control in the United States.

This module will teach you how to:
1) Understand the basic concepts behind compartmental models and their advantages and disadvantages over individual-based (micro) models.
2) Understand a basic taxonomy and key properties of macro-dynamic models.

Andrew Brouwer is mathematical epidemiologist and modeler. He is currently an Assistant Research Scientist in the Department of Epidemiology at the University of Michigan. Andrew uses mathematical and statistical modeling to address public health problems in infectious disease, cancer, and tobacco control. He is a faculty affiliate in the Center for Study of Complex Systems. Andrew holds graduate degrees in mathematics, statistics, and environmental science and engineering.

This module will teach you how to:
1) Understand why transitions are important to many research questions for tobacco control
2) Specify, implement, and interpret a multistate transition applied to longitudinal data of tobacco product use
3) Graphically summarize transition probabilities and hazard ratios
4) Apply a multistate transition model to the Population Assessment of Tobacco and Health (PATH) Study

Required software:
R (v4.0 or later) and RStudio. Packages: msm, minqa, expm, dplyr, numDeriv, ggplot2, reshape2

Note:
Includes a 15-minute break between conceptual and hands-on.

Andrew Brouwer is mathematical epidemiologist and modeler. He is currently an Assistant Research Scientist in the Department of Epidemiology at the University of Michigan. Andrew uses mathematical and statistical modeling to address public health problems in infectious disease, cancer, and tobacco control. He is a faculty affiliate in the Center for Study of Complex Systems. Andrew holds graduate degrees in mathematics, statistics, and environmental science and engineering.

This module will teach you how to:
1) Understand why transitions are important to many research questions for tobacco control
2) Specify, implement, and interpret a multistate transition applied to longitudinal data of tobacco product use
3) Graphically summarize transition probabilities and hazard ratios
4) Apply a multistate transition model to the Population Assessment of Tobacco and Health (PATH) Study

Required software:
R (v4.0 or later) and RStudio. Packages: msm, minqa, expm, dplyr, numDeriv, ggplot2, reshape2

Dr. Ritesh Mistry is Core Lead for the Career Enhancement Core (CEC) and a member of the CAsToR Steering Committee. Dr. Mistry is an Associate Professor of Health Behavior and Health Education at the University of Michigan. His research focuses on evaluating health disparities in smoking trends, as well as adolescent tobacco use and secondhand smoke exposures in global settings.

Learning objectives:
1) Understand the conceptual underpinnings of latent transition analysis and its value to tobacco control research.
2) Describe the key analysis steps involved in conducting latent transition analysis.
3) Interpret the parameters and results generated from latent transition analysis.

Dr. David Mendez is Core Lead for the Career Enhancement Core (CEC) and Project Lead for Research Project 2. He is also a member of the CAsToR Steering Committee. Dr. Mendez is an Associate Professor in the Department of Health Management and Policy at the University of Michigan. His research focuses on modeling trends of cigarette smoking cessation or switching to e-cigarettes. Dr. Mendez’s research also investigates the financial implications of these trends, with a specific focus on tobacco control in the United States.

This module will teach you how to:
1) Understand the basic principles of Bayesian estimation.
2) Understand how Bayesian estimation relates to structural time series modeling.
3) Understand the mechanics of the Kalman filter and its applications to structural time series modeling.

Dr. Le is an Assistant Research Scientist, the Department of Health Management and Policy, School of Public Health, University of Michigan. Dr. Le received her Ph.D. in Applied Mathematics, from the University of Padua, Italy. Dr. Le's research interest includes Dynamic Systems, Mathematical Modeling, and Parameter Estimation.

The hands-on part of the module will teach you how to:
1) Apply the Kalman filter to estimate the trajectory of the smoking cessation rate from smoking prevalence observations over time.
2) Learn how to implement and interpret the results of a Kalman filter estimation in R.

Required software:
RStudio