I have been doing some work with patient data (e.g. maternal comorbidities and complications during pregnancy) and I would like to have your opinion on how to create good baseline models for risk assessment, thus following some premisses:
(a) baseline models have to be simple (no machine learning please)
(b) risk assessment is not a classification task, i.e., is not about predicting whether or not a clinical condition will happen, but what is the risk of that happenning in the future
(c) having said "b", I have a more specific question, have your sayings on this: "given patient risk score, should it be change when the patient turns to be positive for a given condition?" . For example, "A pregnant patient is a low risk pregnancy for miscarriage, but the miscarriage suddenly happens. What about the patient risk score then, should it increase?"
- Badges
- Science topic
- Similar topics
- Linguistics
- Composition Studies
- Publication