IMPROVED LABORATORY TEST SELECTION AND ENHANCED PERCEPTION OF TEST-RESULTS AS TOOLS FOR COST-EFFECTIVE MEDICINE

Inconsistencies in the way physicians perceive and handle identical la boratory results have untoward effects on morbidity, mortality and cos t of medical care. In this context, the selection of suitable tests to answer definite clinical questions, and the manner in which laborator y results are presented have great impact on the action taken by the c linician. This review addresses preferred methods to improve laborator y test selection, and examines methods that more effectively convey la boratory results to clinicians. It is anticipated that refined selecti on of tests, and presentation of the test results in a configuration t hat is easily perceived by the clinician, will facilitate interpretati on of laboratory reports. Furthermore, any measures that promote the a pplication of laboratory information in medical practice improve econo mics at the laboratory-clinical interface. The presently described met hods to optimize test selection and interpretation are: likelihood rat ios to provide estimates of the ability of a test to identify a clinic al condition; consensus- and discriminant function-analysis to estimat e the performance of tests in diagnosing a particular disease or condi tion; receiver operating characteristic (ROC) curves to assess discrim ination capabilities. The methods which improve test result perception are expression of results as multiples of the upper normal limit, uti lizing signal strength to provide prognostic probabilities, and presen tation of results in graphic forms that display mutually interrelated functions, with a specific cluster of results being highly suggestive of a given condition. In addition, we discuss application of expert sy stems to provide rules based on knowledge and experience to analyze re sults of tests and suggest diagnosis and action, including additional tests when required. It is anticipated that judicious utilization of l aboratory services by application of the reviewed methodologies will h elp to achieve medically justified responses at a lower cost and help to achieve a proper balance between cost of tests and their clinical u sefulness.