A rational request behavior: The development of prediction instruments regarding thyroid function tests in primary care

A prospective study was performed to encourage a rational thyroid-stimulati ng hormone (TSH) test request behavior of physicians, using prediction inst ruments, The latter give direct feedback about request adequacy on the basi s of pretest probabilities of hyperthyroidism using patients' signs and sym ptoms. For instrument design, stepwise logistic regression was used on diag nostic data acquired through questionnaires, answered by 80 physicians and 668 patients for whom physicians requested a TSH test. Instruments were des igned for clinical and subclinical hyperthyroidism and for clinical hyperth yroidism alone. Use of the instrument for clinical or subclinical hyperthyr oidism on the selected group, at a 5% probability threshold, can result in a 37% reduction of unnecessary TSH test requests. With the instrument for c linical hyperthyroidism at a 5% probability threshold, the number of unnece ssary test requests can be reduced by 57%. Therefore, it can be concluded t hat the instruments can determine TSH test request adequacy and encourage a rational TSH test request behavior of physicians at low pretest probabilit y thresholds.