How to replace the old panels of conflicted experts
Diagnostic and treatment indication criteria can be based on evidence
as opposed to 'panels of conflicted experts'. Data from trials can be
used to decide on cut-off points for starting treatments and choosing the
best tests e.g. by comparing albumin-creatinine ratios and albumin
excretion rates (AER).
In a trial to compare irbesartan with placebo, about 196 patients
with an AER between 20 and 40 mcg/min needed to be treated (NNT) to stop
one diabetic patient developing nephropathy within 2 years. However, in
patients with an AER of 41 to 160 the NNT was about 9. If the AER was
over 160 one more patient developed nephropathy for about every 21 treated
[1].
This analysis of data from the IRMA2 trial allowed patients to be
'triaged' into those who were (1) too mild to treat (where the chance of
adverse effects would outweigh the chance of benefit) (2) those worth
considering for treatment and (3) those too severe to benefit.
It is possible to plot the estimated proportion benefitting against
the AER using histograms or Gaussian kernel spline functions [1]. The
proportion benefiting can then be estimated for the individual patient's
particular AER result (or mean of many results) and used in Decision
Analysis if wished [2].
Diagnoses are titles to theories to explain why some actions benefit
patients with groups of findings [3]. In this sense, a 'triaged' AER
between 20 to 40mcg/min can be regarded as a 'sufficient diagnostic
criterion' for 'mild microalbuminuria'. Those with an AER between between
41 and 160 mcg/min can be labelled 'moderate microalbuminuria' and those
with an AER over 160 mcg/min labelled 'severe microalbuminuria'.
The same approach can be applied to 'scores' based on symptoms, test
results etc and applied to any treatment and any diagnosis.
References
1. Llewelyn D E H, Garcia-Puig, J. How different urinary albumin
excretion rates can predict progression to nephropathy and the effect of
treatment in hypertensive diabetics. JRAAS, 5; 141-5, 2004.
2. Llewelyn H, Hopkins A (eds). Analysing how we reach clinical
decisions. Royal College of Physicians of London, 1993.
3. Llewelyn H, Ang AH, Lewis K, Al-Abdullah A. The Oxford Handbook of
Clinical Diagnosis, 2nd edition. Oxford University Press, Oxford. 2009.
Rapid Response:
How to replace the old panels of conflicted experts
Diagnostic and treatment indication criteria can be based on evidence
as opposed to 'panels of conflicted experts'. Data from trials can be
used to decide on cut-off points for starting treatments and choosing the
best tests e.g. by comparing albumin-creatinine ratios and albumin
excretion rates (AER).
In a trial to compare irbesartan with placebo, about 196 patients
with an AER between 20 and 40 mcg/min needed to be treated (NNT) to stop
one diabetic patient developing nephropathy within 2 years. However, in
patients with an AER of 41 to 160 the NNT was about 9. If the AER was
over 160 one more patient developed nephropathy for about every 21 treated
[1].
This analysis of data from the IRMA2 trial allowed patients to be
'triaged' into those who were (1) too mild to treat (where the chance of
adverse effects would outweigh the chance of benefit) (2) those worth
considering for treatment and (3) those too severe to benefit.
It is possible to plot the estimated proportion benefitting against
the AER using histograms or Gaussian kernel spline functions [1]. The
proportion benefiting can then be estimated for the individual patient's
particular AER result (or mean of many results) and used in Decision
Analysis if wished [2].
Diagnoses are titles to theories to explain why some actions benefit
patients with groups of findings [3]. In this sense, a 'triaged' AER
between 20 to 40mcg/min can be regarded as a 'sufficient diagnostic
criterion' for 'mild microalbuminuria'. Those with an AER between between
41 and 160 mcg/min can be labelled 'moderate microalbuminuria' and those
with an AER over 160 mcg/min labelled 'severe microalbuminuria'.
The same approach can be applied to 'scores' based on symptoms, test
results etc and applied to any treatment and any diagnosis.
References
1. Llewelyn D E H, Garcia-Puig, J. How different urinary albumin
excretion rates can predict progression to nephropathy and the effect of
treatment in hypertensive diabetics. JRAAS, 5; 141-5, 2004.
2. Llewelyn H, Hopkins A (eds). Analysing how we reach clinical
decisions. Royal College of Physicians of London, 1993.
3. Llewelyn H, Ang AH, Lewis K, Al-Abdullah A. The Oxford Handbook of
Clinical Diagnosis, 2nd edition. Oxford University Press, Oxford. 2009.
Competing interests: No competing interests