An Introduction to R for EpidemiologistsSession Chair(s): Stephen MooneyDate: 2018-06-19 Time: 8:30 am - 12:30 pmLocation:
This workshop will introduce participants to the R statistical computing platform for use in epidemiologic analysis. It is not intended to transform untested novices into R wizards in a mere half-day; rather, the goal will be to introduce the conceptual underpinnings, tools, and external resources that participants will need to overcome barriers to using R that they might encounter on their own, later. The material is designed for epidemiologists who are already familiar performing analyses using other statistical software (e.g. SAS/Stata/SPSS) but who have no first-hand experience with the R language. More specifically, the course will cover 1) basic R syntax, 2) importing data, 3) constructing, cleaning, and manipulating data objects, 4) loading and using external packages, 5) simple statistical modeling, and 6) graphics. Participants must bring a laptop with R installed; the instructor will be available by email beforehand to assist with R installation if difficulties arise.
Session Chairs:Stephen Mooney, University of Washington
An introduction to directed acyclic graphs: What you never wanted but needed to know about bias and didn’t even know to ask.Session Chair(s): Ian ShrierDate: 2018-06-19 Time: 8:30 am - 12:30 pmLocation:
This workshop will introduce participants to directed acyclic graphs (DAGs). We will review the basic principles and show how they can be used to determine appropriate sets of variables for estimating total causal effects of exposure (treatment). Participants will work through concrete examples of increasing complexity. We will also discuss more advanced applications, including natural and controlled direct and indirect effects, bias-amplification, how DAGs might be helpful in evidence synthesis, selection diagrams and SWIGs.
Session Chairs:Ian Shrier, McGill University
Confounding control for estimating causal effects: Looking under the hoodSession Chair(s): Nicolle Gatto, Ulka CampbellDate: 2018-06-19 Time: 8:30 am - 12:30 pmLocation:
This workshop introduces concepts of causal inference and confounding control for causal effect estimation. We will introduce potential outcomes, and articulate the conceptual basis and assumptions for two g-methods - standardization via g-computation and inverse probability weighting. Starting with a simple point-treatment setting we will explore how these methods estimate a causal effect, comparing them to more conventional techniques such as multivariable regression and propensity score control. We will then build to the more complex scenario of time-dependent confounding. Participants will learn how to apply these methods in SAS and R using an observational dataset with the primary goal of unpacking any “black boxes” to clarify the links among the causal effect of interest, the mechanics of these g-methods, and the programming code.
Session Chairs:Nicolle Gatto, Pfizer IncUlka Campbell, Pfizer Inc
Critical review and preparation of manuscripts reporting epidemiologic findings *(Sponsored by AJE)Session Chair(s): Moyses SzkloDate: 2018-06-19 Time: 8:30 am - 12:30 pmLocation:
In this half-day workshop, participants will critically review a paper as initially submitted to the American Journal of Epidemiology (AjE), but not yet published. The paper will be sent to participants in advance of the workshop for their critical review. During the workshop, a presentation will be made on some of the main points to be considered when preparing or reviewing a manuscript. Small group work will follow the presentation so that participants can compare their reviews, prepare a consolidated list of critical comments on the paper and a decision letter. Each group will designate a rapporteur, who will present the group’s review of the paper and decision letter to the whole group of participants. At the end of the workshop, students will receive copies of the manuscript’s AjE reviews, the initial editorial decision and the final version of the paper. Attendance will be limited to 30 participants.
Session Chairs:Moyses Szklo, Johns Hopkins School of Public Health
Implementing G Methods in Complex Longitudinal DataSession Chair(s): Ashley NaimiDate: 2018-06-19 Time: 8:30 am - 12:30 pmLocation:
Applied health scientists are increasingly dealing with complex data structures to answer questions about exposure effects and mediation. In such settings, feedback between confounders, exposures, and mediators render standard adjustment methods (regression, restriction, stratification, matching) inappropriate. G methods, which consist of the parametric g formula, g estimation of structural nested models, and inverse probability weighted marginal structural models, can be used to quantify a variety of exposure effects with both simple and complex data structures. After a brief introduction to the potential outcomes framework, I will review obstacles to effect estimation with complex longitudinal data. All three g methods will then be introduced in a series of examples using actual data and software code (SAS, Stata, R). No prior knowledge of causal modeling, counterfactuals, or g methods is required.
Session Chairs:Ashley Naimi, University of Pittsburgh
Should I attend this workshop? An introduction to (counterfactual) decision theory.Session Chair(s): Stephen ColeDate: 2018-06-19 Time: 8:30 am - 12:30 pmLocation:
Epidemiologists face hard decisions. Decisions about research questions, study designs, and analyses. Yet we receive little (if any) formal training in decision theory. Decision theory is a study of the concepts and methods used to make choices. The goal of this workshop is to introduce epidemiologists to a decision theory based on counterfactual causal inference. Rather than describing how we make choices, we focus on normative decision theory (i.e., how to optimize desired outcomes). Even informal use of normative decision theory may improve decision making. First we will review probability theory, statistical inference, and counterfactual causal inference. Next we will discuss von Neumann and Morgenstern’s utility theory. Last we will combine these ideas into a counterfactual decision theory, and discuss examples. Such decision theory provides a framework within which epidemiologists can optimize choices, and thereby optimize epidemiology to improve the healthy human lifespan.
Session Chairs:Stephen Cole, University of North Carolina, Chapel Hill
Bias analysis for unmeasured confounding and measurement errorSession Chair(s): Tyler VanderWeeleDate: 2018-06-19 Time: 1:00 pm - 5:00 pmLocation:
The first half of the workshop will concern unmeasured confounding focusing on a new metric to evaluate sensitivity to unmeasured confounding called the E-value. The E-value is the minimum strength of association, on the risk ratio scale, that an unmeasured confounder would need to have with both the exposure and the outcome, conditional on the measured covariates, to fully explain away a specific exposure-outcome association. E-value calculations for risk ratios, outcomes differences, odds ratios, and hazard ratios will be discussed. The E-value can be calculated in a straightforward way from study results and its use could become routine and help unify interpretation of robustness to unmeasured confounding. The second half of the workshop will focus on measurement error focused on methods for correction and sensitivity analysis for prevalence, risk differences, and risk ratios, regression adjustments, and techniques for both non-differential and differential measurement error.
Session Chairs:Tyler VanderWeele, Harvard University
Estimation and interpretation: introduction to parametric and semi-parametric estimators for causal inferenceSession Chair(s): Jennifer Ahern, Laura BalzerDate: 2018-06-19 Time: 1:00 pm - 5:00 pmLocation:
This workshop will introduce participants to a “causal roadmap” approach to epidemiologic questions: 1) clear statement of the scientific question, 2) definition of the causal model and parameter of interest, 3) assessment of identifiability – that is, linking the causal effect to a parameter estimable from the observed data distribution, 4) choice and implementation of estimators including parametric and semi-parametric, and 5) interpretation of findings. The focus will be on estimation with a simple substitution estimator (parametric g-formula), inverse probability of treatment weighting (IPTW), and targeted maximum likelihood estimation (TMLE) with SuperLearner. Participants will work through the roadmap using an applied example and implement these estimators in R during the workshop session.
Session Chairs:Jennifer Ahern, University of California, BerkeleyLaura Balzer, University of California, Berkeley
Introduction to Implementation ScienceSession Chair(s): Gila NetaDate: 2018-06-19 Time: 1:00 pm - 5:00 pmLocation:
Implementation science seeks to bridge the gap between research and practice by building a knowledge base about how health information and evidence-based interventions, tools, programs, and policies are disseminated and implemented. It is focused on understanding and accelerating the adoption and integration of evidence and evidence-based interventions into practice settings to improve health. Participants will learn what the field of implementation science is (and is not), the key components to design and conduct an implementation study, and how the field is relevant to epidemiologists. Participants will work on developing a specific aims page they can use to apply for funding opportunities in implementation science at the NIH, as well as understanding the key ingredients of a successful grant application. Didactic presentations, small-group work, and expert consultations will facilitate the learning process.
Session Chairs:Gila Neta, NCI, NIH
Presenters:Gila Neta, NCI, NIHWynne Norton, NCI, NIHAntinette Percy-Laurry, NCI, NIH
Multiple Bias Modeling in Causal Inference StudiesSession Chair(s): Onyebuchi ArahDate: 2018-06-19 Time: 1:00 pm - 5:00 pmLocation:
Causal inference entails strong assumptions. The assumptions include the absence of uncontrolled confounding, selection bias, and measurement error. Some combination of these sources of bias can lead to substantial bias during causal analysis. This workshop will give detailed intuitive discussions and illustrations of the separate and combined impact of unmeasured confounders, selection bias, and measurement error including misclassification on effect estimates. Modern multiple bias modeling tools including the sequential application of appropriate bias formulas for each bias source that account for the order of bias occurrence will be demonstrated. Simulation tools for joint probabilistic modeling of multiple sources of bias will be presented. We will discuss how to obtain, specify, and use bias parameters in multiple bias modeling. Interpretation and reporting will be also discussed. Numerical examples and software for implementation will be provided.
Session Chairs:Onyebuchi Arah, University of California, Los Angeles
Using Monte Carlo Simulations for Quantitative Bias AnalysisSession Chair(s): Hailey Banack, Elizabeth Rose MayedaDate: 2018-06-19 Time: 1:00 pm - 5:00 pmLocation:
"This workshop will provide a comprehensive overview of using Monte Carlo simulations as a tool for quantitative bias analysis. The content covered in the workshop is intended to strengthen participants’ understanding of the theoretical concepts of bias and technical details of running Monte Carlo simulations. To accomplish this objective we will work through three examples of bias analysis using Monte Carlo simulations: (1) an example of using Monte Carlo simulation procedures for record-level correction for exposure misclassification bias, (2) an simple example of Monte Carlo simulations for unmeasured confounding, and (3) a more complex example of Monte Carlo simulations for quantifying collider stratification bias. At the end of the workshop, participants should walk away with a greater understanding of the 'why' and 'how' of quantitative bias analysis (i.e., why bias analysis is important and how to use Monte Carlo simulation techniques to implement the bias analysis)."
Session Chairs:Hailey Banack, McGill UniversityElizabeth Rose Mayeda, University of California, Los Angeles
How to make a picture worth 1,000 words: Effectively communicating your research results through statistical graphicsSession Chair(s): Mike JacksonDate: 2018-06-19 Time: 5:30 pm - 7:30 pmLocation:
Epidemiologists can use well-designed statistical graphics to understand data and to guide us toward correct inferences. These graphics can also be powerful tools for communicating study findings. While statistical software makes it easy to produce standard figures, default options often leave much to be desired and can produce figures that distract, confuse, or even distort data. In this workshop, participants will learn the fundamentals of effective data visualization and how to apply these best practices to create their own graphics. We will begin by reviewing general principles for displaying data, enabling viewers to make comparisons and identify trends and correlations. We will then show how to put these principles into practice, such as creating charts that are appropriate for a given dataset and the effective use of color, size, and annotation. We will also walk through techniques for going beyond the default settings of some software packages to produce well-designed figures.
Session Chairs:Mike Jackson, Kaiser Permanente Washington
Regression Models For Selection Bias, Nonrandom Exposure, and Unobserved Confounding Using StataSession Chair(s): Chuck HuberDate: 2018-06-19 Time: 5:30 pm - 7:30 pmLocation:
Observational data often have issues which present challenges for the data analyst. Data are sometimes missing not at random (MNAR) which can lead to sample selection bias. The exposure of interest or treatment status is often not assigned randomly. And many statistical models for these data must account for unobserved confounding. This talk will demonstrate how to use standard maximum likelihood estimation to fit extended regression models (ERMs) that deal with all of these common issues alone or simultaneously.
Session Chairs:Chuck Huber, STATA
Research in collaborative study designs: design and analytic considerations *(JHSPH Sponsored)Session Chair(s): Keri Althoff, Bryan LauDate: 2018-06-19 Time: 5:30 pm - 7:30 pmLocation:
Individual cohort studies have formed collaborations in the fields of cardiovascular disease, cancer, diabetes, genetics, child health, kidney disease, and HIV. The goal of these collaborations is to curate “big” structured data that can be readily used for epidemiologic research. In this workshop, we will provide a mix of didactic and hands-on learning opportunities for design and analysis in disseminated and pooled individual-level data collaborative study designs, as well as tips from the leaders as to how to engage with these research enterprises. We have identified epidemiologic experts from the existing collaborative study designs (CSDs) who are are versed in the strengths and challenges of epidemiologic research in this context. As the popularity of CSDs continues to grow, the field of epidemiology must take its place at the table to harness the strength of this approach for rigorous science. Agenda: 10 mins - Why collaborations form, Bryan Lau 15 mins – Meta data visualization tools for design, Keri Althoff 5 mins – Immediate clarifying questions 15 mins – Best practices in disseminated analyses, Joe Corresh & Kunihiro Matsushita 5 mins – Immediate clarifying questions 15 mins – Best practices in individual pooled data analyses, Jessie Buckley 5 mins – Immediate clarifying questions 15 mins – The culture of collaboration, Mike Saag 5 mins – Immediate clarifying questions 10 mins – Discussant, Lisa Jacobson 20 mins – Questions (moderated by Lisa Jacobson) (This is the Johns Hopkins Department of Epidemiology sponsored workshop.)
Session Chairs:Keri Althoff, Johns Hopkins School of Public HealthBryan Lau, Johns Hopkins School of Public Health
Tools for transparent epidemiology: Evaluating time-varying covariate balance in longitudinal studiesSession Chair(s): John JacksonDate: 2018-06-19 Time: 5:30 pm - 7:30 pmLocation:
In clinical trials and observational studies, participants change treatments, are lost to follow-up, or censored when they deviate from protocol. When the future outcomes are related to time-varying covariates, selection-bias can affect intention-to-treat and per-protocol effects. In this workshop, participants will translate the concept of covariate balance with time-fixed exposures (e.g. propensity-score matching) to diagnosing measured time-varying confounding with time-varying exposures and selection-bias with informative dropout. Participants will also learn how to estimate and visualize these diagnostics using recently developed SAS macros and R functions. Using simulated data based on a comparative effectiveness study of antipsychotic medications with substantial dropout, participants will describe the measured selection-bias in the raw data, and compare the residual selection-bias after applying inverse probability of censoring weights under various modeling strategies.