Abstract
BACKGROUND: Physician characteristics and perceptions and their effect on choice of therapies for patients with thyroid cancer have been well studied. Some data also exist about physician characteristics and prescribing treatment for subclinical hypothyroidism. The effect of physician characteristics on prescribing thyroid preparations for treating overt hypothyroidism is less studied. METHODS: Members of the American Thyroid Association were surveyed in 2017. Physicians were presented with 13 different theoretical patients with hypothyroidism and asked to choose among six therapeutic options, including levothyroxine, synthetic combination therapy, thyroid extract, and liothyronine monotherapy. The 13 patient scenarios incorporated parameters that potentially provide reasons for considering combination therapy (presence of symptoms, low serum triiodothyronine concentration, and documentation of deiodinase polymorphisms). Repeated-measures logistic regression analysis was performed to examine the prescribing of the various therapies. Data regarding the responding physicians were also collected. These data included number of years in practice, country of practice, and specialty. Multivariate repeated-measures logistic regression analysis of prescribing patterns was also conducted controlling for all patient and physician characteristics. RESULTS: Of the 389 survey respondents, 93% prescribed therapy for hypothyroidism. Fifty-three percent of respondents had been in practice for >20 years, and 23% had been in practice for 11-20 years. Sixty-four percent practiced in North America, and 18% practiced in Europe. Eight-six percent were endocrinologists, and 5% were surgeons. In multivariate analysis, physicians from North America were both more likely to prescribe any triiodothyronine-containing therapies (odds ratio [OR] = 1.8 [confidence interval (CI) 1.3-2.4]) and more likely to add liothyronine to levothyroxine therapy (OR = 1.9 [CI 1.2-2.9]). In addition, they were more likely to prescribe desiccated thyroid extract or liothyronine monotherapy (OR = 1.7 [CI 1.0-2.9]). CONCLUSIONS: A previous analysis of this survey showed that patient characteristics profoundly affect physician prescribing patterns. The current multivariate analysis shows that physician characteristics affect prescribing patterns. Whether this is due to impact upon physicians of patient-related experiences, media exposure, influence from pharmaceutical companies, educational activities, or other concerns cannot be determined. However, these results have potential importance for understanding physician-patient interactions at a time when the benefits and risks of triiodothyronine-containing therapies have not been fully documented.