This repository uses the simulated cohort data from the public repository https://github.com/MethodsForReproducibleHealthResearch/Assignment2.
Variables in the dataset (my interpretation):
- smoke: binary variable indicating that the patient is an ever smoker (1) or non-smoker (0)
- female: binary variable indicating that the patient is female (1) or male (0)
- age: continuous variable indicating the patient's age in years
- cardiac: binary variable indicating any cardiac event (1) and no event (0)
- cost: continuous variable indicating the cost of medical treatment in dollars
Below is a description of the repository structure:
1-data: contains raw cohort data in csv format
Quarto document for analysis that generates the reproducible report
2-output: contains tables and figures generated by the Quarto document
references.bib is the bibliography file used to create citations
Assignment_4.html is the reproducible report
Methods:
Aim 1: We use a generalized linear model with a binomial family to evaluate the association between being a smoker and having any cardiac event. The model was adjusted for potential confounders sex and age.
Aim2: We use a generalized linear model with a binomial family and an interaction term between cost and smoking status to evaluate effect modification (by cost) of the relationship between being a smoker and having any cardiac event. The model is adjusted for sex, age, and the main effect of cost.
Summary of results:
- Table 1 contains descriptive statistics for variables grouped by cardiac event status. There was no missing data.
- Figure 1 shows that participants with cardiac events have a slightly higher median age compared to controls
- Figure 2 shows that on average, participants who experienced any cardiac event have higher medical treatment costs compared to those who did not
- Table 2 contains the adjusted odds ratios for aim 1
- Table 3 contains adjusted odds ratios for aim 2
Reproducibility:
- The reproducible report uses pacman to load packages so that any missing packages can be automatically installed.
- Clicking on Session Information at the end of the reproducible report (before references) will provide R version and package version numbers to reproduce this analysis.
Statement: I did not use generative AI technology (e.g., ChatGPT) to complete any portion of the work.