I have a public EHR dataset which contains info on

a) lab tests b) diagnosis c) surgical procedures d) drugs prescribed etc

Now, using the above data elements, I would like to predict the below

a) When will be the next 3 visits of the patient to the hospital?

b) What will be the reason for his next 3 visits?

For example - Patient A's next 3 visit will be on Jan 21st 2017 (for disease A) , Dec 15th 2017 (for disease B) and Mar 16th 2018 (due to disease C).

I think what I am trying to do is multi step forecasting but not just using a plain time series but with a data which has chronology of events occurred for a patient in hospital.

While I understand date level predictions may be too detailed, any other approach on how this can be done and how to formulate this problem would really be helpful.

Can experts here provide me the list of steps and techniques to follow? So, I can learn methods that you list and be aware of the tasks involved in this pipeline?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.