I have devices N1- N 20, each device could be used in a patient (once, twice, thrice or four times.). A single patient may have more that one device used on him (each device could have been coming from single use, twice, etc). Now patients' outcome is measured according to the use of several devices (may be one, twice..etc) used on the patient, and the outcome I try to measure one of the following (no event, mild event, severe event, and death).

What is the best statistical test to use here, tying number of device use to outcome?

Similar questions and discussions