Validation studies
Validation studies were performed by manufacturer representatives and/or laboratory staff at each institution. These studies included precision, linearity/analytical measurement range (AMR), and method comparison correlating each institution's DxC800 and AU5822 analyzers using EP Evaluator® software (Build 10.0.0.517, Data Innovations, LLC, South Burlington, VT). Precision studies were performed according to the CLSI EP05 protocol [5], using at least two levels of quality control (QC) material. Precision was evaluated by analyzing 20 replicates of each level of the quality control material and compiling the data for analysis. To determine the linearity and AMR for each analyte, we performed studies according to the CLSI EP06 protocol [6]. Depending on the analyte being evaluated, 4 to 7 concentrations of linearity standards material were used to evaluate linearity over the AMR. Triplicate measurements were taken at each concentration spanning the analytical range of the analyte as specified by the manufacturer. Method comparison studies were performed according to the CLSI EP09-A2 protocol [7]. For most analytes, a minimum of 20 to 40 patient specimens were included at each institution for each analyte, as required by the protocol. The range was from 12 (CRP by NMH) to 111 (ALP by IU). Additionally, we consolidated method comparison data points from multiple institutions, and recalculated method comparison statistics based on the consolidated data set. For the 66 analytes studied, the total number of specimens tested ranged from 40 for IgM to 342 for ALP with 58 assays (87.9%) having more than 100 specimens and 27 assays (40.9%) having more than 200 specimens in the merged data set. Specimens were analyzed on each of the two chemistry systems, the DxC800 and AU5822. Deming regression was used to calculate slope, intercept and correlation coefficient (R) for each analyte at each site independently and also for the larger merged data set.