This method assumes that the assay results to be combined are independent and are homogeneous, that is there is no significant variability in the assay results arising from different laboratory conditions. However the results are not assumed to be equally precise.
Given Relative Potency results to combine, with variances of the logarithms of the Relative Potency results, the estimate for the logarithm of the Combined Relative Potency is the mean of the logarithms of the Relative Potencies from the individual studies weighted by the inverse variance:
where the weights are
The logarithm of the confidence interval is estimated based on whether the assays are continuous or quantal.
In the case of continuous assays with degrees of freedom the logarithm of the confidence interval is estimated as
where is the 97.5% quantile for the t-distribution with degrees of freedom.
In case of quantal assays the logarithm of the confidence interval is estimated as
where is the 97.5% quantile from the normal distribution.
