Evaluation of clinical chemistry laboratory performance in twenty Veterans Administration hospitals.

Twenty highly automated Veterans Administration (VA) Hospital laboratories participated in a study to evaluate their performances of routine chemistry tests. Ten of the laboratories had been identified as "superior" on the basis of previous College of American Pathologists Surveys. Several aspects of proficiency surveys were simultaneously examined: use of masked vs. unmasked survey specimens, methods of ranking, and comparison of medically useful performance criteria. The laboratories were found in general to be performing at the level of the state of the art and, more importantly, realistically meeting the criteria for medical needs for most routine tests. There was no difference between results obtained with masked and unmasked specimens. The ten "superior" laboratories showed slightly greater precision than the others, but were not more accurate, and two of them performed very poorly in the present study. The sum of squared SDI's was found to be a helpful technic for ranking and identifying outstanding and poor performances.