Data-driven strategies for carbimazole titration: exploring machine learning solutions in hyperthyroidism control

Reich T; Bakirov R; Budka D; Kelly D; Smith J; Richardson T; Budka M

doi:10.1530/ey.22.3.11

3.11

Prev Next

Section Contents Cite

ESPE Yearbook of Paediatric Endocrinology (2025) 22 3.11 | DOI: 10.1530/ey.22.3.11

ESPEYB25 3. Thyroid Auto-Immune Thyroid Disease (4 abstracts)

3.11. Data-driven strategies for carbimazole titration: exploring machine learning solutions in hyperthyroidism control

Reich T , Bakirov R , Budka D , Kelly D , Smith J , Richardson T & Budka M

J Clin Endocrinol Metab. 2025 Mar 17;110(4):1105-1114. doi: 10.1210/clinem/dgae642. PMID: 39271154

Brief Summary: This retrospective cohort study used machine learning (ML) approaches to improve carbimazole dose titration in patients with autoimmune hyperthyroidism. Using treatment data (carbimazole dose, T4 and TSH levels, attendance dates and several derived variables)from 353 patients, ML models were trained to predict optimal dosing based on prior thyroid function tests and treatment history. Several models outperformed standard physician-guided titration in simulated settings, suggesting potential to enhance biochemical control and reduce over- or undertreatment.

This is one of the many examples of how machine learning may integrate into clinical practice and may improve patient outcomes, also in Endocrinology.