OHDSI Home | Forums | Wiki | Github

Comparing covariates between outcome and target cohort using FeatureExtraction

Hello,

I am using the patient level prediction package to predict endometriosis among those who come to the ER with complaint of abdominal pain. I am wanting to compare the covariates between the outcome and target cohorts to understand what covariates are important for predicting the outcome. My understanding from reading is that we should use computeStandardizedDifference to get the standardized difference in each covariate between the two cohorts. It seems that I will need to manually reconstruct the target and outcome cohort to do this. My concern is that the target contains the outcome subjects. Additionally, not every subject was included when testing the models, as some subjects in the cohort where excluded (attrition) because of time constraints, etc. Is there a way to easily do this using FeatureExtraction or should I use some other R package? In addition, is there a way to view the covariate numbers for
a specific subject in the outcome cohort? Thank you for your help!

t