Predicting atrial fibrillation in primary care using machine learning

• Published in: PloS one Vol. 14; no. 11; p. e0224582
• Main Authors: Hill, Nathan R, Ayoubkhani, Daniel, McEwan, Phil, Sugrue, Daniel M, Farooqui, Usman, Lister, Steven, Lumley, Matthew, Bakhai, Ameet, Cohen, Alexander T, O'Neill, Mark, Clifton, David, Gordon, Jason
• Format: Journal Article
• Language: English
• Published: United States Public Library of Science 01.11.2019; Public Library of Science (PLoS)
• Subjects: Adult; Age Factors; Aged; Analysis; Antihypertensive Agents - therapeutic use; Antihypertensives; Arrhythmia; Artificial intelligence; Artificial neural networks; Atrial fibrillation; Atrial Fibrillation - diagnosis; Atrial Fibrillation - etiology; Audio frequencies; Biology and Life Sciences; Blood Pressure; Blood pressure measurement; Body mass; Body Mass Index; Body size; Cardiac arrhythmia; Cardiovascular agents; Cardiovascular disease; Cardiovascular diseases; Cardiovascular Diseases - complications; Clinical medicine; Complications; Computer and Information Sciences; Diagnosis; Electronic health records; Evaluation; Female; Fibrillation; Forests; Health care; Health economics; Health risks; Heart; Heart failure; Humans; Learning algorithms; Machine Learning; Male; Mathematical models; Medical research; Medicine; Medicine and Health Sciences; Middle Aged; Mortality; Neural networks; Neural Networks, Computer; Novels; Patients; Performance prediction; Pharmaceuticals; Prediction models; Pressure measurement; Primary care; Primary Health Care - methods; Regression analysis; Retrospective Studies; Risk analysis; Risk assessment; Risk Assessment - methods; Risk Factors; Statistical analysis; Stroke; Studies; Support vector machines; United Kingdom; United Kingdom > UK
• ISSN: 1932-6203; 1932-6203
• DOI: 10.1371/journal.pone.0224582

Atrial fibrillation (AF) is the most common sustained heart arrhythmia. However, as many cases are asymptomatic, a large proportion of patients remain undiagnosed until serious complications arise. Efficient, cost-effective detection of the undiagnosed may be supported by risk-prediction models relating patient factors to AF risk. However, there exists a need for an implementable risk model that is contemporaneous and informed by routinely collected patient data, reflecting the real-world pathology of AF. This study sought to develop and evaluate novel and conventional statistical and machine learning models for risk-predication of AF. This was a retrospective, cohort study of adults (aged ≥30 years) without a history of AF, listed on the Clinical Practice Research Datalink, from January 2006 to December 2016. Models evaluated included published risk models (Framingham, ARIC, CHARGE-AF), machine learning models, which evaluated baseline and time-updated information (neural network, LASSO, random forests, support vector machines), and Cox regression. Analysis of 2,994,837 individuals (3.2% AF) identified time-varying neural networks as the optimal model achieving an AUROC of 0.827 vs. 0.725, with number needed to screen of 9 vs. 13 patients at 75% sensitivity, when compared with the best existing model CHARGE-AF. The optimal model confirmed known baseline risk factors (age, previous cardiovascular disease, antihypertensive medication usage) and identified additional time-varying predictors (proximity of cardiovascular events, body mass index (both levels and changes), pulse pressure, and the frequency of blood pressure measurements). The optimal time-varying machine learning model exhibited greater predictive performance than existing AF risk models and reflected known and new patient risk factors for AF.