Document Type

Capstone Experience

Graduation Date

5-2020

Degree Name

Master of Public Health

Department

Biostatistics

First Committee Member

Kendra Schmid

Second Committee Member

Lynette Smith

Third Committee Member

Andreea Boeck

Abstract

Clinical laboratory control materials are an integral part of legally-mandated and highly regulated quality control protocols in all clinical laboratories. These controls ensure accurate performance of the laboratory testing and instrumentation used to produce medical test results for millions of patients. It is of clinical and public health interest to ensure the diagnostic test results which affect so many people are regulated by the most accurate and precise controls.

Formulation changes in control materials have the potential to impact laboratory quality control. In this study, data from two formulations of a hematology control were compared to assess equivalency of the mean results for each parameter. The Two One-Sided T-Test of equivalency was used to compare the old formulation of this hematology control to a new formulation. Equivalency testing assumes a null hypothesis that two sets of data differ, which is contrary to other statistical methods which assume a null hypothesis of non-difference.

Data for 15 hematology parameters were gathered for 6 different lots of product, on 4 different instruments, with 3 lots each for the old and new formulations. ANOVA analysis revealed significant differences in means between lots and instruments, so equivalency testing was stratified on lot and instrument to obtain specific details about any lack of equivalency. Testing was also done with non-stratified data from the mean results of the old and new formulations to obtain general conclusions about the equivalency of each parameter in each level.

The results of the TOST analysis described in this report show that all except two of the parameters in the new formulation are equivalent to the old formulation. The parameters which did not demonstrate equivalence between the two formulations were white blood cell (WBC) count and red blood cell size distribution (RDW). Although the lack of equivalency in these two parameters is notable, equivalency testing of the means alone is not sufficient to fully determine the clinical significance. Both accuracy and precision of controls must be maintained after a formulation switch. TOST equivalency testing is able to assess accuracy by determining whether target concentrations in the new formulation are equivalent with target concentrations in the old formulations but does not address any potential changes in precision as a result of altered variability. While overall equivalency of the mean results is an important step in evaluating the accuracy of the new formulation, the manufacturer should also consider evaluating the equivalence of variance to verify precision before making any decision to switch to the new formulation.

Share

COinS