La Escuela Historia e información sobre la Escuela
Información Sobre cursos, encuentros y otras actividades
Noticias Noticias y actualidad de la Escuela
 
 
XXVIII Escuela de Salud Pública de Menorca
Del 18 al 27 de septiembre de 2017
Llatzeret de Maó (Menorca)
Institut Menorquí d'Estudis - Camí des Castell, 28 - 07702 Maó (Menorca) - Tel. 971 351500 -
LOGO
       
IPAES

IPAES - Causal Inference Course

Profesores:
Dr. Felix Elwert.
Dr. Elwert is Romnes Vilas Associate Professor of Sociology and Population Health Sciences at the University of Wisconsin–Madison. He holds graduate degrees in sociology and in statistics from Harvard University. His research concerns social demography and social inequality. His methodological research concerns methods for causal inference in observational, quasi-experimental, and experimental settings. He won the first Causality in Statistics Education Award from the American Statistical Association in 2013.

Dr. Tyler Vanderweele.
Dr. VanderWeele is Professor of Epidemiology at the Harvard T.H. Chan School of Public Health. He holds degrees from the University of Oxford, University of Pennsylvania, and Harvard University in mathematics, philosophy, theology, finance and applied economics, and biostatistics. His research concerns methodology for distinguishing between association and causation in observational research including mediation, interaction and spillover effects. He won the first Causality in Statistics Education Award from the American Statistical Association in 2015.

Dr. Xavier Basagaña.
Dr. Basagaña is Associate Research Professor in biostatistics at the Barcelona Institute for Global Health. He holds a graduate degree in biostatistics from Harvard University. His research concerns methodology on statistical methods to produce valid interferences in observational studies (e.g. on measurement error and missing data), relationship between extreme temperatures and health, and relationship between air pollution and health.

Entidades colaboradoras:


Lugar:
Llatzeret

Horario:
From 18th to 22nd of September

Descripción:
Welcome to the 1st meeting of the International Programme of Advanced Epidemiology and Statistics. This 1st meeting introduces the Causal Inference course which includes three modules: Causal inference with directed graphs, Causal mediation and interaction analysis, and Methods to deal with attrition and missing data. The goal of this course is to empower epidemiologists, biomedical scientists, public health professionals, research clinicians, social scientists, and statisticians to apply causal inference with confidence. The course is taught in English by renowned international lecturers. The programme is held in the Llazeret of Maó within the Public Health School of Menorca. Participants will enjoy an intense learning experience in a unique historical environment. The Llatzeret is a small island at the port of Maó wherein all boats entering Spain during the 19th century had to stop for quarantine to get a certificate of being free of infection diseases.

Objetivos:
To empower epidemiologists, biomedical scientists, public health professionals, research clinicians, social scientists, and statisticians to apply causal inference with confidence.

Contenidos:
CAUSAL INFERENCE COURSE PROGRAM
Welcome to the course 18th of September at 9.00h.

Causal Inference with Directed Graphs. Dr. Felix Elwert
This 2-days module offers an applied introduction to directed acyclic graphs (DAGs) for causal inference from observational data. DAGs are rigorous and accessible tools for understanding and solving complicated causal problems. All causal inference relies on causal assumptions, and DAGs are a graphical notation for these causal assumptions. Analysts can use DAGs to derive the statistical implications of their causal assumptions and determine which statistical associations equal (“identify”) causal effects. We will use DAGs to study the identification of causal effects, with particular emphasis on identification by adjustment, which underlies the use of regression and matching techniques for causal inference. We will animate the material with numerous examples from the social and biomedical sciences.

Day 1 – 18th of September
9.30 – 11.00h Causal effects in the potential outcomes framework
11.00 – 11.15h Coffee break
11.15 – 13.00h Using Directed Acyclic Graphs (DAGs) to notate causal assumptions: Central elements
13.00 – 14.00h Lunch break
14.00 – 15.45h Using DAGs to infer what statistical associations: D-separation, testable implications
15.45 – 16.00h Coffee break
16.00 – 18.00h Using DAGs to determine whether an association identifies a causal effect: Adjustment criterion, backdoor criterion, and shortcuts

Day 2 – 19th of September
9.00 – 11.00h Understanding selection bias: conditioning on a collider
11.00 – 11.15h Coffee break
11.15 – 13.00h Interpreting regression coefficients
13.00 – 14.00h Lunch break
14.00 – 15.45h DAGs for linear models
15.45 – 16.00h Coffee break
16.00 – 17.00h DAGs for linear models
17.00 – 21.00h Social Evening including Course Dinner


Causal Mediation and Interaction Analysis. Dr. Tyler Vanderweele
This 1.5-day module consists in one part on causal mediation analysis and another one on causal interaction analysis. Mediation analysis concerns assessing the mechanisms and pathways by which causal effects operate. Discussion will be given as to when the standard approaches to mediation analysis are or are not valid. The no-confounding assumptions needed for these techniques will be described. The use and implementation of sensitivity analysis techniques to assess the how sensitive conclusions are to violations of assumptions will be covered, as will extensions to multiple mediators. We will also discuss interaction on additive and multiplicative scales, and their relation to statistical models. We will discuss issues of confounding for interaction analyses and how whether control has been made for only one or both of two exposures affects interpretation. We will discuss conditions under which interaction gives evidence of synergism within the sufficient cause framework, when interaction is robust to unmeasured confounding, and methods attributing effects to interaction.

Day 3 – 20th of September
9.00 – 11.00h Concepts and Methods for Mediation
11.00 – 11.15h Coffee break
11.15 – 13.00h Sensitivity Analysis
13.00 – 14.00h Lunch break
14.00 – 15.45h Mediation with Time-to-Event Outcome / Multiple Mediators
15.45 – 16.00h Coffee break
16.00 – 18.00h Multiple Mediators cont. / Unification of Mediation and Interaction

Day 4 – 21st of September
9.00 – 10.30h Concepts and Methods for interaction
10.30 – 11.00h Interaction vs. Effect Heterogeneity
11.00 – 11.15h Coffee break
11.15 – 13.00h Mechanistic Interaction
Power and Sample Size Calculations for Interaction
13.00 – 14.00h Lunch break


Methods to deal with attrition and missing data. Dr. Xavier Basagaña
This 1.5-days module will provide methodological tools to overcome potential selection biases induced by attrition, i.e. selected individuals not participating in the study or being lost to follow up, or by having missing information in some of the key study variables. Missing data are ubiquitous in observational and experimental research. An inappropriate analysis of a study with missing data can lead to incorrect inferences, both in terms of bias and in the quantification of uncertainty. We will discuss the potential problems induced by missing data and will provide an overview of two main methods that can deal with such problems, namely inverse probability weighting and multiple imputation. The module will give special focus on the practical implementation of these techniques in realistic settings.

Day 4 – 21st of September
14.00 – 15.45h Missing data: consequences, assumptions and solutions
15.45 – 16.00h Coffee break
16.00 – 18.00h Practical exercise: multiple imputation

Day 5 – 22nd of September
9.00 – 11.00h Inverse probability weighting
11.00 – 11.15h Coffee break
11.15 – 13.00h Practical exercise: inverse probability weighting
13.00 – 14.00h Lunch break
14.00 – 15.45h Advanced topics in multiple imputation
15.45 – 16.00h Coffee break
16.00 – 17.00h Concluding remarks and overview of further topics

Profesionales a los que se dirige principalmente:
The programme is aimed at applied researchers with an interest in causal inference from observational data, in particular for epidemiologists, biomedical scientists, public health professionals, research clinicians, social scientists, and statisticians. Participants should have a good working knowledge of applied regression analysis and an intermediate knowledge level of epidemiology.

Precio:
1200 euros

Documentos relacionados:

Imágenes:

Información adicional:
REGISTRATION
The registration fee includes all course materials, lunch and coffees from Monday September 18th to Friday September 22nd 2017, and the Course Dinner on Tuesday September 19th. Participants have to book their accommodation themselves. We will provide a bus service for the transfer between the hotels and the Llatzeret. Further details for registration will be announced shortly. For further information, please contact us at ipaes.emsp@gmail.com

MATERIALS
Participants will receive a bound manual containing detailed lecture notes, reading materials, and many other useful features. In the first module, which aims to strengthen your ability to think through causal problems, we will work through numerous pencil-and-paper exercises. In the second and third module participants need to bring their laptop computers to participate in the software-based exercises. Software code in SAS, STATA, and R will be provided.


Inscripciones

 
SOCIAL MEDIA
           
Organizan:

Consell Insular de Menorca GOIB UIB UIMP
Colaboran:

GOIB
Institut Menorquí d'Estudis - Camí des Castell, 28 - 07702 Maó (Menorca) - Tel. 971 351500
Gobierno de España Gobierno de España