Weekly rate of infection

[Barker_E]

Say a clinical trial was conducted of 5,000 hospitalized patients with disease X. The primary endpoint was two weeks but the maximum follow up was one year.
Suppose you wanted to look at the infection rate of those patients - how would you calculate the weekly rate of infection based on the data given?

 

[hlsmith]

Is infection the primary endpoint? Can you get infected more than once?

In general, what about visualizing it with a Kaplan Meier curve?

 

[Barker_E]

The primary endpoint isn't stated so I assume that it isn't an infection. In theory, I think you could get the infection more than once but within a year I assume it's highly unlikely.

 

[hlsmith]

So you have no raw data, but you want to calculate the infection rate at discrete intervals. Seems hard without raw data. It may help to post a link to the paper!

 

[Barker_E]

Sorry, I didn’t make it clear. It’s a hypothetical task so just looking for suggested ideas. My thoughts were incidence or logistic regression (converted to a rate) but I wasn’t sure how that would work as a weekly rate.

 

[hlsmith]

So you want to calculate it for every week for a cohort and you don't know if they can be infected more than once?

If you don't need to control for covariates, regression isn't necessarily necessary.