This is a public repository for the code used to fit the models used for the SECURE-IBD Covid-19 Risk Calculator. The data is not publicly available (researchers can request access through the SECURE-IBD website), but estimated model coefficients are available in this repository.
The online risk prediction tool can effectively predict which unvaccinated patients with inflammatory bowel disease (IBD) may be at higher risk for COVID-19-related morbidity. Three primary outcomes are examined:
- Hospitalization+ (Hospitalization or death)
- ICU+ (ICU admission, mechanical ventilation, or death)
- Death
- Data Preprocessing
- Multiple Imputation (multivariate imputation by chained equations)
- Statistical Modeling (using 85% of cases reported from March-September 2020)
- Model Performance Evaluation (AUROC based on remaining 15% of cases and 300+ patients added to the registry after model fitting)
This analysis was conducted using R Version 4.0.2, and the tidyverse, glmnet, glmnetUtils, mice, magrittr, future, and pROC packages.
The calculator was created using R Shiny.
Below is an example risk prediction for a patient with below-average predicted risk. Young age, gender and a lack of comorbidities contribute to a lower-than-average risk of adverse COVID-19 outcomes.
For further details, please reference the paper:
Sperger, J., Shah, K. S., Lu, M., Zhang, X., Ungaro, R. C., Brenner, E. J., Agrawal, M., Colombel, J.F., Kappelman, M.D., & Kosorok, M. R. (2021). Development and validation of multivariable prediction models for adverse COVID-19 outcomes in patients with IBD. BMJ Open; 11:e049740. doi:10.1136/ bmjopen-2021-049740