Point-of-Care Testing (POCT) Networks

Abstract

Background Certification rules of point-of-care testing (POCT) require periodic evaluation of performance of POCT devices using control material different from that used daily. The aim of this report was to verify the reliability and the labor required for a quality assessment program operated by the laboratory. Methods Two pools (L1 and L2) of serum were obtained by collecting residual materials from inpatients. The glucose value of the pools was established by a central laboratory method (ADVIA 2400, Siemens). The values for the 2 pools were L1 = 70.3 ± 0.02 and L2 = 296 ± 0.05 mg/dL). The pools were also tested for viral serologies. Both samples were analyzed on 2 different days (at least 1 month apart) on 50 POCT devices (Accucheck Inform2, Roche Diagnostics). The overall results were used for analytical performance evaluation. The labor required for the program was recorded. Results Mean values for all of the POCT instruments were: L1 = 80.23 ± 1.85 and L2 = 308.3 ± 8.3. Bias between the laboratory and the POC devices were L1 = 6.59% and L2 = 2.03%. No total error more than 15% was observed; 1 data point showed a total error greater than 10%; and 7 samples showed a total error greater than 5%. Labor required was 15 minutes for laboratory staff and 7 ± 5 minutes for nurses. Discussion Overall comparison confirms a good correlation between POC and the laboratory, although a bias greater than 5% may be expected for low values. Interdevice variability and the number of samples with standard errors greater than 5% results were very low. An evaluation of the accuracy of POCT instruments can be performed by the central laboratory. The main difficulties are related to sample selections and delivery. A total of 11:40 labor hours for nurses and 25 hours for the laboratory operators was required for monitoring 50 instruments.