Abstract
Three surveys were made of the quality of plasma cholesterol measurements performed with a commercial desktop analyser (BCL Reflotron) in primary care. Each survey included three specimens, and results were received from 37, 61, and 69 participants. Although many participants obtained satisfactory results, 8.6% of the results differed by 1.0 mmol/l or more from the target values, and the overall between instrument dispersion of results was 1.3 times that between hospital laboratories. It was found that common sources of error were poor technique and the use of outdated reagent strips. Users of such instruments outside the laboratory need help and advice with training, and guidelines for this are provided. The main recommendations are that users should establish contact with a local clinical chemistry laboratory for training and support and should participate in external quality assessment schemes.