Search

In general, blood gas analysers can also determine the value of haematocrite by measuring the blood's conductivity. The question to ask is whether this value is reliable. In this study, hematocrit obtained via conductivity from 6 different pieces of equipment were compared with those measured using the gold standard method, which is microcentrifugation. By interpreting the results of 320 arterial blood samples taken in the intensive care unit DAR "B" we can see that the reliability between two measurements on the same piece of equipment is very good, in general > 0.95 whatever the equipment. The reliability between the means of the two measurements and the gold standard is slightly lower but remains very satisfactory, most often between 0.8 and 0.9. The Gem Premier 3000 (IL) analyser and the Roche OMNI S gave the best reliability compared with centrifugation. The Spearman coefficients between the mean values of the analysers and those of centrifugation were high, with the exception of the Rapidpoint 405. They are all statistically different from zero (p<0.0001).

An evaluation process directed by the biochemistry department of the UHC of Montpellier about acquisition of blood gas analysers for a point of care testing in 10 different medical care services. The values obtained by the different measurement machines have been compared taking into account the intra and inter measures variabilities. In this work, the results of 4 machines tested: Rapid-point 405, GEM Premier 3000, i-STAT 1 (série 300) and ABL 77 are compared with one i-STAT (série 200) and one OMNI S already available in the laboratory. The study lasted one month (november 2004). Each week, the two machines of the laboratory was compared with a new analyser randomly chosen between the four machines tested. In this study, including 21000 values, the statistical tests have been applied under the hypothesis that none of the measurement machine could be a "gold standard". Beside the classical research of correlation between the different values, we added the research of concordance between machines and the application of imperfect gold standard method. A high level of statistic concordance was observed between the different analysers. So the equipment has been selected using biological, analytic and computing criteria.