0
$\begingroup$

I have a data set of individual level data for the outcome in question. I also have aggregated data (at a national level) from which I'm obtaining the denominator. Using this information, I've determined the annual incidence of the condition (after a certain exposure) over a 20 year period. Each individual can only get the disease once and then no longer contributes to the person-time.

I am then removing a subset of individuals whom I believe are being over-diagnosed. My intent was to show the difference in the incidence rates over time between when these individuals are included in the data set vs when they are excluded.

Now, my question is -- how do I compare these two annual incidence rates over time --- especially given that they are not two distinct populations.

$\endgroup$
0
$\begingroup$

I would not approach this problem this way.

Rather, I'd use the whole data set and then predict the disease using the variables you have. One variable could be your opinion on whether they are in the 'over diagnosed' group and you could then use logistic regression.

But be careful that your opinion is not based on results. In fact, if you've already seen the results, then I wouldn't do this at all.

I am not accusing you of anything, but there's a reason for blinding.

$\endgroup$

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.