Real-world data: the art of modelling (workshop at annual WEON conference)

Covid-19 is a disease with a high mortality and morbidity rates, for which hospitalization is not uncommon. After hospitalization, patients may suffer from long-term complaints, including e.g. fatigue. You are asked to develop a prediction model that gives the probability of return to work within 6 months after hospitalization.

You are provided with a dataset ( download data ) and a data dictionary (below). This dataset contains a set of potential predictor variables, available at the time of hospital discharge, as well as the time to return to work (with a maximum follow-up of 1.5 years).

Variable name Description Values
ID Patient ID
age Age (years) Numeric
sex sex 0 = male, 1 = female
ICU Was admitted to ICU during hospitalization 0 = no, 1 = yes
oxygen Leaves hospital while using additional oxygen 0 = no, 1 = yes
Tiff Tiffeneau index numeric
BMI Body mass index numeric
vacc Corona vaccination status 0 = unvaccinated, 1 = vaccinated
ThrombEvent Thrombotic event during hospitalization 0 = no, 1 = yes
LOS Hospital length of stay Numeric
education Educational level 0 = low, 1 = high
DM Diabetes mellitus 0 = no, 1 = yes
MRI Consolidations present on MR images at the day of hospital discharge 0 = no, 1 = yes
Time Time since discharge (days) Numeric
Event Return to work 0 = no, 1 = yes