Table 2.

Comparison of pooled value marker with actual average of in individual cases contributing to the pools

Pelvic disease group
Menopause
Pre
Post
Post
Pre/post
Pre
Post
Stage
III/IV
III/IV
I/II
I/II/III/IV
Benign
Benign
HistologyNonmucinousNonmucinousNonmucinousMucinousEndometriosisOvarian serous
CA 125
    Pool2,349.02,246.0374.5114.0113.187.1
    Sample average2,265.62,145.8337.8111.1109.340.6
    Standard bias*0.020.030.070.020.040.86
CA 19.9
    Pool30.029.0199.92557.034.717.7
    Sample average29.328.1231.9766.733.618.9
    Standard bias*0.020.03−0.080.500.02−0.06
CA 15.3
    Pool216.8187.946.626.821.219.8
    Sample average216.3187.346.727.620.819.0
    Standard bias*≈0≈0≈0−0.040.040.09
CEA
    Pool1.81.92.37.81.62.2
    Sample average1.61.62.27.41.42.0
    Standard bias*0.140.180.110.050.380.23
  • NOTE: The formula for standard bias follows from the sequential application of two principles. The first is that pooling of samples averages the individual concentrations for the cases by linearity of concentrations in mixing equal volumes. Therefore, the first calculation is ”mean (individual cases),“ and this value is compared with ”pool value.” The second is that normal theory should be applied on a scale for which the distribution is closest to normality. Hence, the next step is to apply the transform, in this case ”log,” to all measurements in the calculation, then determine the difference and the SD on the scale that is closest to normality.

  • * Standard bias = [log(pool value) − log(mean(individual cases))] / SD(log(individual cases)).