Data-Driven Strategies for Carbimazole Titration: Exploring Machine Learning Solutions in Hyperthyroidism Control
Authors' conclusion
To simulate a use-case, the accumulation of the prediction error over time was assessed. It was determined that an improvement in accuracy is expected if this model was to be deployed in practice.
Study context
- Population
- Patients with hyperthyroidism (primarily autoimmune) treated with carbimazole titration at University Hospitals Dorset (UHD)421 extracted
- Participants
- 353 patients
- Exposure
- Machine learning classification algorithms for carbimazole dose determination
- Outcome
- Model performance: weighted F1 scores and Brier scores
Methodological detailssingle-centre, Different ML classification algorithms tested under several data processing regimes; iterative model selection followed by feature selection
- Single-centre
- Different ML classification algorithms tested under several data processing regimes; iterative model selection followed by feature selection
Main finding
As reported in the abstract
Prediction of carbimazole dose (model performance)
| Group | Value | Selected model F1 | Selected model Brier score |
|---|---|---|---|
| Random forest (RF) | average F1 score 0.731 | 0.751 | 0.38 |
Random forest (RF)
Valueaverage F1 score 0.731
Selected model F10.751
Selected model Brier score0.38
Observational data: associations, not effects.
Original publication
Data-Driven Strategies for Carbimazole Titration: Exploring Machine Learning Solutions in Hyperthyroidism Control. J Clin Endocrinol Metab. Published 13 September 2024. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →