Contents
A well-designed epidemiologic research does the best to produce a precise and valid estimate of association between a particular exposure and outcomes of interest. Nonetheless, even the association obtained from a well-designed study may inevitably be susceptible to both random errors and systematic errors.
The conventional epidemiologic approach to data analysis yields a quantitative assessment of random errors by estimating 95% confidence intervals for point estimate. It also adjusts the association for confounding variables in regression model. However, it is rare to quantify systematic errors induced by selection bias, measurement error, confounding by unmeasured confounders, or residual confounding by measured confounders that are poorly specified or poorly measured. Quantitative bias analysis addresses these shortcomings in the conventional approaches handling epidemiologic data analyses.
Currently, quantitative assessments of systematic error in most of published epidemiologic research are absence. However, recent development in leading epidemiology have called for routine training in bias modeling for epidemiology students, so demand for implementation and presentation of bias modeling in research articles will certainly grow in the near future.
Aim
The course aims at providing the background, implementation, and interpretation of quantitative bias analyses concerning epidemiological studies.
Prerequisites
- Basic knowledge on epidemiological concept of selection bias, confounding bias, and information bias.
- Knowledge on logistic regression is a requited.
- Stata will be used during the whole process.
Max number of participants
20 PhD students
Course fee
The course is free of charge for PhD students enrolled in Universities that have joined the "Open market agreement" or NorDoc.
For other participants there is a course fee of:
DKK 2400
EUR 321