OBJECTIVES. To investigate the predictive performance of machine learning (ML) algorithms for estimating anticoagulation control (AC) in patients with atrial fibrillation (AF), treated with vitamin K antagonists.

METHODS. This was a retrospective cohort study of adult patients (≥18 years) between 2007 and 2016 using linked primary and secondary care data (Clinical Practice Research Datalink GOLD and Hospital Episode Statistics). Various ML techniques were explored to predict poor AC, defined as time in therapeutic range (TTR) <70% based on international normalized ratio (INR) 2.0-3.0. Baseline (linear and non-linear support vector machines; random forests; stochastic gradient boosting [XGBoost]; neural networks [NN]) and time-varying data (6 week intervals up to 30 weeks (long-short term memory [LSTM] NN)) were employed. Patient records representing unique lines of warfarin therapy (LOT) were separated into training (70%) and holdout sets (30%) for model training and testing.

RESULTS. 35,479 patients were eligible for inclusion, of whom 24,684 and 10,795 were assigned to the training (32,683 unique LOTs) and holdout sets (14,218 unique LOTs). Across all models, depression was a significant driver in predicting AC. At baseline, XGBoost was the best-performing model (area under the curve [AUC]: 0.624) due to its ability to identify non-linear interactions for factors including age and weight (greater probability of poor control: <65 and >80 years and <70kg, respectively). Addition of time-varying data to the LSTM NN improved predictive performance, plateauing at AUC of 0.830 at 30 weeks.

CONCLUSIONS. ML algorithms displayed very good ability in predicting patients at greater risk of poor AC. The addition of time-varying data to the algorithm, especially prior INR measurements, improved predictive performance. These algorithms may be clinically valuable in identifying and supporting patients who may benefit from more frequent INR monitoring or switching to therapies not requiring dose adjustments.