Ota Go 066965

Embed Size (px)

Citation preview

  • 8/12/2019 Ota Go 066965

    1/5

    ADVANCED EPIDEMIOLOGY - 4 DAY SHORT COURSE

    CAUSATION, SYSTEMATIC ERROR,

    QUANTITATIVE BIAS ANALYSIS and CAUSAL MEDIATION

    Convenor and lecturer: Professor Tony Blakely, University of Otago

    Lecturer: Professor John Lynch, University of Adelaide

    Day 1

    Causation in Epidemiologyo Models: Causal criteria (Bradford Hill) and Sufficient Component Cause Model

    (Rothman);

    o Individual and population causesPotential Outcomes Model of Causation

    o Counterfactual or Potential Outcomes Modelo Why perfect RCTs work creating exchangeable groupso The inevitable problems with most RCTs and observational studieso Confounding, Selection, and Information Bias

    Directed Acyclic Graphs (DAGs)

    o Formalizing assumptions for a causal modelWorkshop - Whats your DAG?

    Understanding Selection Bias an application of DAGs

    o Conditioning on common effect(s) and dependent on participation by exposure andoutcome

    o St d d i ti l t i t ti

  • 8/12/2019 Ota Go 066965

    2/5

    o Study design options complete case versus imputation

    o Class exercises using ExcelProbabilistic Bias Analysis

    o Bringing it all together simultaneous assessment of selection, confounding andinformation bias.

    o Class exercises using Excelo Monte Carlo simulation

    Workshop - Dealing with difficult issues

    o Workshop group discussion of examples provided by course participants

    Day 4

    Direct and Indirect Effects

    o Estimating mediationCausal Mediation Analysis

    o Marginal Structural Models (MSM)Bringing it all together DAGs, Sensitivity Analyses and MSMs

    o Structured small-group review of paper by Nandi et al.,Epidemiology(2014)Multiple Bias Analysis

    o Examples from Rothman Ch 19.Final Quiz, Feedback and Wrap-up

    ---------------------------------------------------------------------------------------------------

    Course PrerequisitesThis course will assume knowledge of study design and analytical methods, the basic

    principles of systematic error (confounding, selection and information biases) and

    biostatistics up to multivariable regression. For example, successful completion of a Diploma

    or Masters of Public Health course in epidemiology and biostatistics (or similar) will usually

    pro ide the necessar basis to ndertake this co rse

  • 8/12/2019 Ota Go 066965

    3/5

    Readings:We have selected a number of readings that correspond to the major topicscovered in the course. The readings are available on DROPBOX

    (https://www.dropbox.com/). You will be sent an invitation to the DropBox.

    Causation and Counterfactuals

    o Kaufman, J, Poole C. Looking back on "Causal thinking in the health sciences".Annu RevPub Health. 2000;21: 101-119.

    o Coggon DIW, Martyn CN. Time and chance: the stochastic nature of disease causation.Lancet2005;365:1434-1437.

    o

    Little RJ, Rubin DB. Causal Effects in Clinical and Epidemiological Studies Via PotentialOutcomes: Concepts and Analytical Approaches.Annu Rev Public Health. 2000;21: 121-145.

    o Greenland S, Brumback B. An overview of relations among causal modeling methods.IntJ Epidemiol2002;31:1030-1037.

    DAGs

    o Hernn MA, Hernndez-Diaz S, Werler M, Mitchell AA. Causal knowledge as aprerequisite for confounding evaluation: An application to birth defects epidemiology. Am

    J Epidemiol2003; 155:176-184

    o Hernndez-Daz S, Wilcox AJ, Schisterman EF, Hernn MA. From causal diagrams tobirth weight-specific curves of infant mortality.Eur J Epidemiol. 2008; 23(3): 163166.

    o Hernan MA, Cole SR. Invited Commentary: Causal Diagrams and Measurement Bias.Am.J. Epidemiol.2009;170:959-62.

    Confounding

    o Greenland S, Robins J. Identifiability, exchangeability, and epidemiological confounding.I t J E id i l 1986 15 412 18

    https://www.dropbox.com/https://www.dropbox.com/https://www.dropbox.com/http://www.ncbi.nlm.nih.gov/entrez/eutils/elink.fcgi?dbfrom=pubmed&retmode=ref&cmd=prlinks&id=18224448http://www.ncbi.nlm.nih.gov/entrez/eutils/elink.fcgi?dbfrom=pubmed&retmode=ref&cmd=prlinks&id=18224448http://www.ncbi.nlm.nih.gov/entrez/eutils/elink.fcgi?dbfrom=pubmed&retmode=ref&cmd=prlinks&id=18224448http://www.ncbi.nlm.nih.gov/entrez/eutils/elink.fcgi?dbfrom=pubmed&retmode=ref&cmd=prlinks&id=18224448https://www.dropbox.com/
  • 8/12/2019 Ota Go 066965

    4/5

    o Glynn RJ, Schneeweis S, Sturmer T Indications for propensity scores and review of theiruse in pharmacoepidemiologyBasic Clin Pharmacol Toxicol2006; 98(3):253-9

    Instrumental Variables

    o Greenland, S. An introduction to instrumental variables for epidemiologists.Int JEpidemiol. 2000;29(4): 722-729.

    oNewhouse JP, McClellan M. Econometrics in outcomes research: The use of instrumentalvariables.Annu Rev Public Health1998;19:17-34.

    o Glymour M. Natural experiments and instrumental variable analysis in socialepidemiology. In: Oakes M and Kaufman J. (eds)Methods in Social Epidemiology. Jossey

    Bass, San Francisco: 2006, pgs 429-460.

    Marginal Structural Models

    o Hernan MA, Robins JM. Estimating causal effects from epidemiological data.J EpidemiolCommunity Health2006; 60:578-586.

    oNandi A, Glymour M, Kawachi I, VanderWeele TJ.Using Marginal Structural Models toEstimate the Direct Effect of Adverse Childhood Social Conditions on Onset of Heart

    Disease, Diabetes, and Stroke Epidemiology. 2012; 23:223-232o De Stavola BL, Daniel RM.Commentary: Marginal Structural Models: The Way Forward

    for Life-course Epidemiology?Epidemiology. 2012; 23:233-237.

    oNandi A, Glymour M, Subramanian SV. Association among socioeconomic status, healthand behaviors, and all-cause mortality in the United States. Epidemiology 2014;25:170-

    177.

    Quantitative Bias Analysis

    o M d E id i l Ch t 19 Bi A l i

    http://journals.lww.com/epidem/Abstract/2012/03000/Using_Marginal_Structural_Models_to_Estimate_the.8.aspxhttp://journals.lww.com/epidem/Abstract/2012/03000/Using_Marginal_Structural_Models_to_Estimate_the.8.aspxhttp://journals.lww.com/epidem/Abstract/2012/03000/Using_Marginal_Structural_Models_to_Estimate_the.8.aspxhttp://journals.lww.com/epidem/Fulltext/2012/03000/Commentary___Marginal_Structural_Models___The_Way.9.aspxhttp://journals.lww.com/epidem/Fulltext/2012/03000/Commentary___Marginal_Structural_Models___The_Way.9.aspxhttp://journals.lww.com/epidem/Fulltext/2012/03000/Commentary___Marginal_Structural_Models___The_Way.9.aspxhttp://journals.lww.com/epidem/Fulltext/2012/03000/Commentary___Marginal_Structural_Models___The_Way.9.aspxhttp://journals.lww.com/epidem/Abstract/2012/03000/Using_Marginal_Structural_Models_to_Estimate_the.8.aspxhttp://journals.lww.com/epidem/Abstract/2012/03000/Using_Marginal_Structural_Models_to_Estimate_the.8.aspxhttp://journals.lww.com/epidem/Abstract/2012/03000/Using_Marginal_Structural_Models_to_Estimate_the.8.aspx
  • 8/12/2019 Ota Go 066965

    5/5

    5

    TIMETABLE

    Session Day 1: Mon 28 July Day 2: Tues 29 July Day 3: Wed 30 July Day 4: Thurs 31 July

    0900 to1030

    Course Overview [TB]

    Causation [JL]

    Selection Bias QBA [TB]:

    o

    Formulaso QBA Class exercises

    Excel

    Interaction and Effect MeasureModification [TB]

    What does OR measure (if time)[TB]

    Direct and Indirect Effects [TB]Fixed effects

    MORNING BREAK

    1100 to1230

    Potential Outcomes [JL]

    DAGs [JL]

    Confounding [TB]:

    o Properties, counterfactual,DAGs

    o Approaches to regressionmodel building

    Information bias [TB]:

    o Definitions, DAGs, etc

    Information bias analysis:

    QBA Class exercises Excel

    Causal Mediation Analysis:

    Marginal Structural Models 1

    Marginal Structural Models 2

    LUNCH

    1330 to1500

    Workshop #1: Whats yourDAG? (to be handed out by JL)

    Alternative methods forconfounding [JL]:

    o Propensity scoreso Instrumental variables

    Probabilistic bias analysis [TB]:

    o Distributions, Monte Carlo,

    o Class exercise Excel

    Bringing it all together DAGs,sensitivity and MSMs

    oClass exercise on Nandi(2014) Epidemiology

    AFTERNOON BREAK

    1530 to1700

    Quantitative bias analysis (QBA)overview [TB]

    Selection bias [TB]:

    o Definitions and DAGs

    Study design options completecase vs imputation

    Quiz

    Confounding bias analysis [TB]:

    o Formulas

    o QBA Class exercises Excel

    Quiz

    Workshop #2: challengingissues [TB, LJ]

    Two students: present, classdiscussion. See emailedframework that must be followed.

    Quiz

    Multiple bias analysis[TB]:

    Modern Epi Ch 19 example

    Feedback on the course

    Final Quiz