Ensuring suitable quality of clinical measurements through design.

To design and deliver high quality, safe and effective products, manufacturers of in vitro diagnostic (IVD) products follow a structured, traceable process for controlling the uncertainty of results reported from their measurement systems. This process and its results however, are not often shared in detail with those outside of the manufacturing company. The objective of this paper is to facilitate discussion by describing some of the best practices used during the IVD design and development process, highlighting some design challenges manufacturers face, and to offer ideas for how IVD manufacturers and laboratories could work together to drive further improvement to public health.

A comparison of different sample matrices for evaluating functional sensitivity, imprecision and dilution linearity of the Abbott ARCHITECT i2000 TSH assay.

Important performance characteristics for any thyroid-stimulating hormone (TSH) assay include low-end sensitivity, precision across a wide dynamic range, and linear specimen dilution. Laboratories often characterize the performance of their TSH assay using many different sample matrices (e.g. native human serum, synthetic buffer, processed human serum, etc.). However, this can lead to possible confusion as the relationships between sample matrix and assay performance are often poorly understood. The purpose of our study was to evaluate the performance of the ARCHITECT i2000 TSH assay using a variety of different sample matrices. Functional sensitivity was found to be essentially equivalent in both hyperthyroid patient sera (0.0035 microIU/ml) and TSH affinity-stripped (0.0038 microIU/ml) sample matrices. Assay imprecision was also independent of sample matrix, with total imprecision < or = 5.3% for both synthetic buffer and serum matrices. Similarly, specimens were found to dilute linearly over a wide range in both serum and synthetic buffer matrices. We conclude that the i2000 TSH assay measures TSH similarly in a variety of sample matrices. This is an important design feature for this immunoassay which provides assurance that analytical performance assessed with matrices other than unprocessed human serum are representative of assay performance seen with patient specimens.