Objective: The prevalence of abnormal in-vitro thyroid function tests in psychiatric in-patients may be as high as 24%. Thus far, however, there is only limited data addressing the underlying causes of these abnormal test results, i.e. how often they can be attributed to genuine thyroid disease. Method: We conducted an observational study of all in-patients admitted to our institution during 1 calendar year running analyses of total thyroxin (T4), free thyroxin index (FTI) and thyroid-stimulating hormone (TSH). Patients with abnormal test results were classified according to an algorithm which had been established previously. Results: In 243 of 880 patients with in-vitro thyroid function analysis, at least one concentration of either T4, FTI or TSH was found to be outside the reference range. Work-up according to the algorithm was completed in 848 patients; alterations were classified as representing thyroid dysfunction in 100 (41% of patients with abnormal test results), non-specific findings in 92 (38%), influence of ingested drugs in 18 (7%) and of severe physical disease in 1 (0.4%). As measures of T4 and/or FTI provided no essential information in 854 patients (97% of tested), we found that in most cases the determination of TSH alone was sufficient for demonstrating normal thyroid function. Conclusion: In 27.6% of newly admitted patients living in an iodine-deficient area, at least one abnormal result in either T4, FTI or TSH values was found. Genuine thyroid disease was found in slightly less than half the patients with an abnormal value.