Richard Dixon wrote:
>I have a quick query regarding handling assay results which are below the
>assay's limit of detection. For statistical purposes a value must be
>assigned to such samples. Does anyone have any suggestions on which is the
>most appropriate value to use. Those I most commonly see used are either
>the limit of detection itself, or 50 percent of the limit of detection
>(presumably assuming that of all such samples half will have a 'true' value
>between zero and 50 % LOD, and the other half between 50-100% of the LOD)
I have used a pseudo-gaussian distribution of theoretical mean m and
standard-deviation s generated by the pascal function:
x := Sqrt(-2*Ln(Random))*(Sin(2*Pi*Random))*s+m
When the value x generated by the function was less than a chosen
value y, x has been replaced by 0 to calculate the real value of
the mean and of the standard-deviation of the 100000 points.
Here are the results
no y ---> mean=2.9995 SD=0.99962
y=0 ---> mean=2.9995 SD=0.99984
y=1 ---> mean=2.9854 SD=1.03540
y=2 ---> mean=2.7660 SD=1.40340
If the limit is more than 2SD under the mean (y<=1)the error is
negligible for general biological data.
In the case of y=2, if I replace the true value x by 1 instead of 0,
that is to say half way between 0 and the limit 2, the statistics are:
y=2 ---> mean=2.92460 SD=1.10700
The error is acceptable.
In conclusion I think thant the raised problem is rather theoretical
and any solution is without consequence on the results of calculations.
---------------------------
Dr Philippe MARQUIS
Centre hospitalier regional
57000 METZ - FRANCE
E-mail: [log in to unmask]
---------------------------
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|