Fantastic! Now I can just refer to this paper instead of making my own hard-to-read rant about colliders and table 2 fallacy when co-authors want to do this type of analysis. Love the section debunking the normal defences of this practice, that makes it even more convincing.

Together, the low incidence of very abnormal results and context surrounding them suggest that a rethinking of toxicity monitoring may be in order. We suggest performing monitoring on indication and with reduced frequency in those without indication rather than every few months for all patients.

We also looked at actions following a very abnormal result, the most common of which was no action at all in 36%, further indicating that at least a sizable proportion of these tests were of limited value.

We manually chart-reviewed the circumstances of newly occurring very abnormal results. We found that almost always, patients had a clear reason for higher pre-test chances of finding abnormalities (see table), suggesting testing could be performed on indication or omitted (for known abnormalities).

We looked at all routine safety labs (consisting of ALAT, eGFR, Hb, and Leukocyte and platelet counts) for DMARD-using RA patients (starting 6 months after DMARD initiation) in our clinic from 2008 to 2020.

After determining the cumulative incidence of very abnormal results was quite low, ....

FYI, I can only tick one option here

That honestly sounds like it should be more than enough 'health and disease' knowledge, so I don't know why you're told none of them count. You could try asking another program/trainer.

But AFAIK, the only thing you need the title for is to supervise training of other epi's so not a big miss.

You do not need to study medicine, adjacent studies such as biomedical science are also fine. If you didn't get a health related BSc, you can still become an epi with 8 credits on subjects of 'health and disease', compared to 60 credits required on epi methods (+PhD, + ≥4 epi-related papers)