Abstract:
|
Assays of biomarkers of exposure can sometimes be prohibitively expensive. For a case-control study, pooling equal aliquots of specimens before assay is an effective strategy with several important advantages. There is, however, a vexing problem with pooling: One would normally adjust individually-assayed urinary measurements for diluteness of the urine, typically by dividing by creatinine, and similarly adjust serum levels of a lipophyllic analyte by dividing by the serum total lipid level. Unfortunately, one cannot simply divide the analyte concentration measured in a pooled specimen by the creatinine (lipid) concentration found in the pool, because the ratio of means is not the same as the mean of the corresponding ratios. Our proposed remedy requires individual-specimen measurements for creatinine (lipid). One then either physically dilutes the specimens before pooling, to equalize creatinine or lipid levels across specimens in each pooling set, or pools together aliquots of algebraically-determined unequal volumes and then adjusts the pooled concentration using the individual-specimen creatinine (lipid) values. We assess performance of the approaches via simulations.
|