Practical and theoretical implications of target values in clinical chemistry: a QC approach

摘要

The automation of test validation procedures in a routine clinical laboratory by integrating artificial intelligence with the lab information system ( LIS) is the focus of this study. Test validation consists of three "layers": (1) acceptance of patient results based on technical criteria ("technical validation"), (2) acceptance of patient results based on internal quality control ( QC) results ("QC validation") and (3) medical validation of the patient protocol. The emphasis here is put on QC validation. General concepts regarding the setting of performance limits for internal QC results are briefly reviewed. As a practical example, we describe how the different layers of test validation were integrated with Molis, a LIS (Sysmex, Barchon, Belgium) used in a routine clinical biochemistry lab. PGP5 software (PGP, Brussels, Belgium) is used for technical validation and is installed on a Linx work-cell (Thermo Clinical Labsystems, Vantaa, Finland). The latter integrates four different Vitros instruments (OCD, Beerse, Belgium). QC validation and medical validation are performed with QC-Today (IL, Zaventem, Belgium) and VALAB (Erim, Toulouse, France), respectively. Both programs are bi-directionally coupled with the LIS. After 2 years of experience with this integrated system, it is clear that automation of validation rules and procedures has enhanced efficiency and overall quality of patient results.