Georg Uttam wrote:
>We got the following data (measured TIBC on X-axis;
>calculated TIBC on Y-axis).
>
>slope 1.22
>Intercept -19.62
>RSQ 0.922
>r 0.962
>TIBC Range 205-484 ug/dL
>
>Calculated values are ~20% higher than measure values (to get TIBC in
>ug/dL, values of trasferrin in mg/dL were multiplied by 1.43).
>
>How is the experience of other people in switching measured TIBC to
>calculated TIBC from transferrin? Thanks for comments/suggestions.
What was the intention with that correlation? If you want to predict
TIBC from transferrin you should plot transferrin on the X-axis. The
parameters of a correlation is not sufficient in order to judge the
reliability. A good correlation can be obtained if choosing a great
range of values. The distribution of values is essential for the
judgement of the predictivity etc. At least you must mention the
number of observations!
>From a correlation it is possible to calculate the confidence interval
for Y given a certain value for X. By choosing critical values for
X, e.g. the reference limits, you can decide if the correlation is
good enough in order to predict Y at those critical limits.
In general when comparing specific and unspecific methods such a calculation
will effectively show that the specific method (despite its higher cost)
has a considerably better clinical predictivity than the unspecific test.
One such example that has been discussed in this mailbase is total vs. ionized
calcium, where the analytical and physiological errors in total (or albumin
corrected) calcium frequently will cause hyperparathyroidism to be undetected
and healthy persons to be judged as hypercalcaemic.
In this case transferrin is a specific method and TIBC is more unspecific. On
the other side, iron may be bound not only to transferrin but also to other
proteins in plasma. So what was the intention with the correlation?
Mr Sten Öhman, PhD
[log in to unmask]
-----------------------------------------------------
From: Mr Sten Öhman, PhD
Postal address: p.o. Box 133, S-590 70 Ljungsbro, Sweden
E-Mail address: [log in to unmask]
Phone: int: +46 13 219020, nat: 013-219020
Fax: int: +46 13 219021, nat: 013-219021
Personal home page: http://hem1.passagen.se/stoh7971/
Company home page: http://www.elfinilab.se/
------------------------------------------------------
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|