Normal Reference Range Vs Pre-defined site normal values [Design Issues]

posted by ElMaestro  – Denmark, 2019-12-18 00:21 (1821 d 15:05 ago) – Posting: # 20988
Views: 10,743

Hi all,

❝ rare, exceptional basis.


I have been involved in such discussions with WHO. A lot.
I pointed out to them that the reference range (which I would not myself define in a different manner than a normal range, but we can discuss that separately) usually involves 5% being outside even if normal. And an event which occurs with 5% frequency is not rare or exceptional, but "common", I think, when we talk the SPC definitions used for e.g. adverse events. So, there is some degree of word clutter here.

❝ To cite an example

❝ A. Laboratory Reference range for Haemoglobin - 13-18 g/dL

❝ B. Pr-defined Site Normal range for Haemoglobin - > 12.5 g/dL


It is not uncommon that labs have a reference range or a normal range or whatever you want to call it, and that the PI can decide to enroll someone who is outside if she/he is deemed healthy (if the protoco needs hevo's, etc). This is all at the PI's discretion or his/her delegate. However, if the phenomenon occurs frequently (let us say more than 5%, and no I am not goiing to comment on statistical testing for exactly that), then you are right in the middle of the trouble because thewn you are truly recruiting from a subject pool which does not reflect what's normal or what's reference and this is what the essence of WHO's publication is about, as far as I know.

As a results of WHO's publication some years ago some lab redefined their ref. range for hemoglobin, to get around the issue. Which is exactly what they were not supposed to do.
So even if WHO's choice of words was not optimal, the way it has been handled at CRO's is not ideal either.

Bottom line:
  1. The PI can deem a subject outside the ref. range enrollable.
  2. But if the PI does so too often, then it spells trouble.
  3. Recruit from a population which is truly reflecting what has been submitted to IEC/IRB and NCA, not a pool of subjects with lower hemoglobin or whatever range.
  4. Do not speculate too heavily into stats testing or multiplicity or correlation (there are usually 20-40 reference ranges in play in BE studies; there is always someone who has something which is outside; if HB is out then it is likely that e.g. Hematocrit is too).

Are you confused?
So am I.

Pass or fail!
ElMaestro

Complete thread:

UA Flag
Activity
 Admin contact
23,336 posts in 4,902 threads, 1,669 registered users;
21 visitors (0 registered, 21 guests [including 6 identified bots]).
Forum time: 15:27 CET (Europe/Vienna)

Biostatistician. One who has neither the intellect for mathematics
nor the commitment for medicine but likes to dabble in both.    Stephen Senn

The Bioequivalence and Bioavailability Forum is hosted by
BEBAC Ing. Helmut Schütz
HTML5