Predictors of Adherence to Stroke Prevention in the BALKAN-AF Study: A Machine-Learning Approach
Monika Kozieł-Siołkowska1,2, Sebastian Siołkowski1, Miroslav Mihajlovic3
1Liverpool Centre for Cardiovascular Science, University of Liverpool and Liverpool Heart & Chest Hospital, Liverpool, United Kingdom.
Insights
Machine learning identified key factors influencing stroke prevention adherence in atrial fibrillation (AF) patients. Paroxysmal AF and initial diagnosis predicted adherence, while certain comorbidities indicated lower adherence rates.
Area of Science:
- Cardiology
- Medical Informatics
- Public Health
Background:
- Guideline-adherent stroke prevention in atrial fibrillation (AF) improves patient outcomes.
- The 'Avoid Stroke' component of the ABC pathway is crucial for AF management.
- Predictors of adherence to stroke prevention strategies require further investigation.
Purpose of the Study:
- To evaluate the efficacy of machine learning (ML) algorithms in identifying predictors of adherence to the 'Avoid Stroke' criterion of the ABC pathway.
- To compare ML algorithms with logistic regression for predicting adherence.
Main Methods:
- Post-hoc analysis of the BALKAN-AF dataset.
- Application of machine learning algorithms and logistic regression for feature selection and model building.
- Definition of adherence as oral anticoagulant (OAC) use in AF patients with CHA 2 DS2 -VASc score 0 (male) or 1 (female).
Main Results:
- Machine learning identified paroxysmal AF, treatment center location (capital city), and initial AF diagnosis as predictors of adherence.
- Hypertrophic cardiomyopathy, chronic kidney disease with dialysis, and sleep apnea were associated with lower adherence.
- The random forest ML model achieved an area under the receiver-operator curve of 0.710 for predicting adherence.
Conclusions:
- Machine learning effectively identifies predictors of adherence to AF stroke prevention guidelines.
- Paroxysmal AF, central treatment locations, and early diagnosis facilitate adherence.
- Comorbidities like hypertrophic cardiomyopathy, advanced kidney disease, and sleep apnea present challenges to adherence.
Abstract:
Background Compared with usual care, guideline-adherent stroke prevention strategy, based on the ABC (Atrial fibrillation Better Care) pathway, is associated with better outcomes. Given that stroke prevention is central to atrial fibrillation (AF) management, improved efforts to determining predictors of adherence with 'A' (avoid stroke) component of the ABC pathway are needed. Purpose We tested the hypothesis that more sophisticated methodology using machine learning (ML) algorithms could do this. Methods In this post-hoc analysis of the BALKAN-AF dataset, ML algorithms and logistic regression were tested. The feature selection process identified a subset of variables that were most relevant for creating the model. Adherence with the 'A' criterion of the ABC pathway was defined as the use of oral anticoagulants (OAC) in patients with AF with a CHA 2 DS 2 -VASc score of 0 (male) or 1 (female). Results Among 2,712 enrolled patients, complete data on 'A'-adherent management were available in 2,671 individuals (mean age 66.0 ± 12.8; 44.5% female). Based on ML algorithms, independent predictors of 'A-criterion adherent management' were paroxysmal AF, center in capital city, and first-diagnosed AF. Hypertrophic cardiomyopathy, chronic kidney disease with chronic dialysis, and sleep apnea were independently associated with a lower likelihood of 'A'-criterion adherent management. ML evaluated predictors of adherence with the 'A' criterion of the ABC pathway derived an area under the receiver-operator curve of 0.710 (95%CI 0.67-0.75) for random forest with fine tuning. Conclusions Machine learning identified paroxysmal AF, treatment center in the capital city, and first-diagnosed AF as predictors of adherence to the A pathway; and hypertrophic cardiomyopathy, chronic kidney disease with chronic dialysis, and sleep apnea as predictors of non adherence.
