Abstract
BACKGROUND: Over the past four decades, there has been a substantial increase in the incidence of thyroid cancer with studies suggesting that greater use of thyroid ultrasound contributes to the rise in incidence. However, little is known about physician reported practice patterns on ultrasound use.
Methods: Patients diagnosed with differentiated thyroid cancer in 2014–15 from the Surveillance, Epidemiology and End Results registries of Georgia and Los Angeles were surveyed and asked to identify the surgeon who performed their thyroid surgery, and the endocrinologist and other doctors most involved in their thyroid cancer treatment decision making. We surveyed all physicians identified by more than one patient, and a random sample of physicians identified by one surveyed patient (N=610; 65% response rate). Surveyed physicians were asked to identify the clinical scenarios in which they would schedule a thyroid or neck ultrasound. We generated descriptive statistics for all categorical variables and used multivariable logistic regression to identify factors associated with thyroid ultrasound misuse.
Results: The cohort consisted of primary care physicians (PCPs; N=162), endocrinologists (N=176), otolaryngologists (N=130), and general surgeons (N=134). In addition to physicians reporting ultrasound use for accepted reasons such as palpable nodule on exam (98%), large goiter (92%), and nodule seen on other imaging test (88%), a substantial number of physicians endorsed ultrasound use for clinically unsupported reasons: patient request (33%); abnormal thyroid function tests (28%); and positive thyroid antibodies (22%). In multivariable analysis, compared to PCPs, endocrinologists, otolaryngologists, and general surgeons were significantly more likely to schedule an ultrasound in response to patient request (odds ratio (OR) 2.52, 95% confidence interval (CI) 1.27–5.11; OR 2.98, 95% CI 1.57–5.79; OR 2.14, 95% CI 1.17–3.97, respectively). Physicians in private practice were more likely to schedule an ultrasound for abnormal thyroid function tests (OR 2.44, 95% CI 1.33–4.73) and positive thyroid antibodies (OR 2.47, 95% CI 1.27–5.21) compared to those in academic medical centers. Physicians who managed ten patients or less, compared to more than 50 patients, with thyroid nodules in the past 12 months were less likely to schedule an ultrasound for positive thyroid antibodies (OR 0.43, 95% CI 0.19–0.95).
Conclusion: Physicians report scheduling thyroid ultrasound for reasons not supported by clinical guidelines and in conflict with the Choosing Wisely recommendations. Understanding why physicians use thyroid ultrasound and factors that correlate with clinically unsupported reasons is essential to creating targeted educational interventions to improve physician adherence to guidelines, reduce unnecessary imaging, and curb the overdiagnosis of low-risk thyroid cancer.
Causal associations of thyroid function and dysfunction with overall, breast and thyroid cancer: A two‐sample Mendelian randomization study