Karolinska Institutet and National University of Singapore
Beyond Classic Designs and Analysis for Health Data
online course, 27 June to 08 July 2022 (6 half-day sessions)
Here is the schedule
and a summary of the structure and learning outcomes
Prior to the course, participants will be invited to describe an investigation from their own work,
that they will be encouraged to use in the practical exercises wherever possible
Before the course, participants should prepare their computer and software environment using the following advice for
R users here
Stata users here
(ADO and HLP files for Session 5) here
These documents also explain where to store the data sets which should be downloaded from here and unzipped.
The lectures and exercises will be provided here as PDF files, organised by day
and solutions to the exercises will be provided in advance as ZIP files
Key papers are provided below, sorted within day in the order in which the topics are covered in lectures
Session 1, June 27 2022
- Lecture 1.1 Disease occurrence and risk (review)
- Lecture 1.2 Sources of bias
- Lecture 1.3 Review of epidemiological tables
- Exercise 1.1 Description of own work in terms of design and sampling scheme
- Exercise 1.2 Pencil & paper, openepi.com, Stata/R. solutions are here
Session 2, June 29 2022
- Lecture 2.1 From tables to logistic regression models
- Lecture 2.2 The role of matching in cohort and case-control studies
- Lecture 2.3 Analysis of matched data.
- Exercise 2.1 Logistic regression. solutions are here
- Exercise 2.3 Exercise 2.3: Conditional logistic regression. solutions are here
Session 3, July 1 2022
- Lecture 3.1 Introducing time: inclusive, exclusive and concurrent sampling
- Lecture 3.2 Survival, hazard, Cox regression.
- Lecture 3.3 Matching on time, HR from conditional logistic regression.
- Exercise 3.1 Inclusive, exclusive and concurrent sampling Illustration in an Excel sheet and in-class demonstration
- Exercise 3.2 Cox, logistic and conditional logistic regression. solutions are here
Session 4, July 4 2022
- Lecture 4.1 Missing data, 2-stage studies, inverse probability weighting
- Lecture 4.2 Optimal sampling of 2-stage data
- Lecture 4.3 Two-stage designs for time-to-event data
- Exercise 4.1: Sampling and analysis of 2-stage data. solutions are here
- for Stata users Exercise 4.2: Optimal sampling in Stata. solutions are here
- for R users Exercise 4.2: Power calculations in R. solutions are here
Session 5, July 6 2022
- Lecture 5.1 RR from case-control data by doubling the cases
- Lecture 5.2 Breaking the time matching; reconstructing the study-base for NCC designs
- Lecture 5.3 Weighted Cox regression of NCC data
- Exercise 5.1: RR from doubling the cases solutions here
- Exercise 5.2: pencil and paper and computation of weights to reconstruct a cohort solutions here and in-class demonstration
Session 6, July 8 2022
- Lecture 6.1 Extension to nested case-control sampling: case-cohort design
- Lecture 6.2 Exposure- enriched controls: counter-matching
- Lecture 6.3 Other controlled designs
Key papers arranged in order of topic within day