Chapitre 14 Exercise - Multiple biases

Folder: Exercise 6 - Multiple biases

Open up the R project for Exercise 6 (Exercise 6 - Multiple biases.Rproj).

In the Question 1 folder, you will find a partially completed R script named Question 1.R. To answer the following questions, try to complete the missing parts (they are highlighted by a #TO COMPLETE# comment). Again, we also provided complete R scripts, but try to work it out on your own first!

We will re-investigate exercise #7 from the QBA part of the course on adjusting for multiple biases, but using a Bayesian analysis. This part was described as follows:

Using the observed study data we will adjust for unmeasured confounding, misclassification of exposure, and selection bias. Use the following information to set up your bias parameters.

  • UNMEASURED CONFOUNDER:
    • Log odds ratio between confounder and disease = 1.6 (equivalent to OR of 5.0) with variance of 0.2
    • Prevalence of confounder in exposed: mode=0.60, 5th percentile=0.50
    • Prevalence of confounder in unexposed: mode=0.20 95th percentile=0.30
  • EXPOSURE MISCLASSIFICATION:
    • Sensitivity of exposure classification: Mode=0.80, 5th percentile=0.75
    • Specificity of exposure classification: Mode=0.90, 5th percentile =0.85
  • SELECTION BIAS:
    • Selection proportion in diseased exposed: Mode=0.75, 5th=0.70
    • Selection proportion in diseased unexposed: Mode=0.25, 95th=0.30
    • Selection proportion in healthy exposed: Mode=0.30, 95th=0.35
    • Selection proportion in healthy unexposed: Mode=0.10, 95th=0.15

  1. Start from the partially completed Question 1.R R script located in Question 1 folder. Using a Bayesian model compute an OR adjusted for unmeasured confounding, misclassification of exposure, and selection bias. In this case, in which order will you have to organize your biases adjustments? First use a deterministic approach (using the bias parameters values proposed, but ignoring the provided 5th and 95th percentiles or any other measures of spread of the distributions).

  2. Modify the R script that you have completed for Question 1. Now use a stochastic approach using distributions for all bias parameters.