Routine thyroid function testing (TFT) in hospitalised patients is not recommended; however, delayed management of thyroid dysfunction may have significant consequences. We quantified the rate of TFT in hospitalised patients and identified factors influencing the likelihood of finding significantly abnormal TSH (<0.01 or >10 mUI). TFTs were performed during 26 937/280 000 (9.6%) admissions between 2007 and 2011. 57.6% of those tested were female and mean age was 63.8 (±19.3 S.D.) years. 75% were emergency admissions and the median hospital stay was 7 days. The primary reason for admission included circulatory (24.2%), respiratory (17.3%), digestive (12.3%), cancer (9.2%), and endocrine disorders (4.4%). 15.0% of those tested were undergoing active treatment for hyper-or hypothyroidism. Multivariate regression analysis identified current thyroid treatment (AOR=3.72), female gender (AOR=1.41), longer hospital stay (AOR=1.48 24 days; AOR=2.22 510 days; AOR=5.00 >10 days vs 1day), older age (AOR=1.40, 1.68, 1.66, and 2.12 per quintile) and emergency admission (AOR=1.18 vs elective) as independent factors associated with increased probability of testing. Those with endocrine disorders were more likely (AOR=2.36) and those with neoplasms (AOR=0.43) or digestive disorders (AOR=0.65) less likely to undergo thyroid function testing compared with subjects with circulatory diseases. Significant TSH abnormalities were found in 1481 patients (5.9% of those tested). Subjects with a primary cancer or an endocrine diagnosis (AOR=3.79 and AOR=2.19 vs circulatory disorders), females (AOR=1.93), those not undergoing active thyroid treatment (AOR=4.0 vs active treatment) and those admitted for elective procedures (AOR=1.4 vs emergency) all had a higher likelihood of having significantly abnormal TSH concentrations.
Conclusions: Within a very large cohort, fewer than 10% of hospitalised patients had thyroid function tested during their in-patient stay; abnormal TSH concentrations were evident in only a small proportion. Further studies are required to identify appropriate drivers for thyroid function testing during hospitalisation to best identify those with thyroid dysfunction who would benefit from further investigation and treatment.