Abstract

The quality of analytical results required in clinical practice is generally dictated by analytical rather than clinical criteria. The increasing availability of analytical technology that can be used outside the hospital laboratory challenges this situation. In an attempt to highlight this dilemma one method of glucose analysis outside the laboratory has been assessed using a pilot external quality assessment scheme (EQAS) and the results have been compared with the performance of the same machines in the laboratory. In the laboratory reliable results were obtained when the Glucometer was used by experienced laboratory personnel giving a good correlation of results (r = 0.96) when compared with an automated method. The results of the EQAS revealed that 44% of hospital ward glucose estimations and 63% of general practice estimations would be considered unsatisfactory by conventional laboratory criteria--that is, greater than +/- 2 SD of the mean laboratory result. This level of performance must, however, be considered in relation to the number of times that a clinically misleading result is obtained. No analytical system should be used in clinical practice without a continuous, objective assessment of its performance.