From guidelines to hospital practice: reducing inappropriate ordering of thyroid hormone and antibody tests.

OBJECTIVE Because of major technical improvements and conscious care about cost effectiveness, limiting the inadequate use of thyroid biological tests appears to be a major issue. DESIGN To (i) estimate the ordering prevalence of each thyroid test, (ii) assess the prevalence of relevant thyroid tests, and (iii) evaluate the impact of expressing justification for tests during a 2-month intervention period on these prevalences. METHODS During a prospective 2-month survey (June-July 1997), all the request forms were divided into four groups of prescription: (1) investigation of thyroid function, (2) taking drugs affecting the thyroid, (3) monitoring of nodule and cancer, and (4) investigation of thyroid autoimmunity. Their appropriateness was thus determined according to consensus in our hospital and previously published recommendations. Results were compared with those of retrospective similar 2-month periods in 1996 and 1998. Combinations of thyroid function tests and thyroid antibodies were analyzed during the 1996, 1997 and 1998 periods. RESULTS The overall estimated rate of appropriate ordering between 1996 and 1997 increased from 42.5% to 72.4% (P<10(-4)), with a significant improvement in each group of main diagnosis referral, except in group 3 where suitability was always over 85%. However, in group 4, appropriateness remained low (36%). Combinations of thyroid tests revealed an increase in single TSH order forms and single autoantibodies to thyroperoxidase (TPOAb) ones, while TSH+free thyroxine+free tri-iodothyronine and TPOAb+ autoantibodies to thyroglobulin ones decreased significantly. Interestingly, all these changes were maintained 1 year later (June-July 1998) even though physicians were not aware of this new study. CONCLUSIONS Persistent change in medical practice was thus assessed.