What does this research mean for the field?

EHR-based machine learning models show promise for improving atrial fibrillation detection in primary care compared to standard care, but clinical applicability is currently limited by high risk of bias, variable performance, and insufficient external validation. Novelty: ClaimNovelty.SYNTHESIS. Consensus alignment: ConsensusAlignment.NEUTRAL.

October 14, 2025PLOS Digital HealthOpen Access

Are Machine Learning methods effective in detecting undiagnosed atrial fibrillation in primary care settings using electronic health records? A systematic review

View Full Paper

Ask AI

Bookmark

Why the study?

Current screening guidelines fail to detect AF effectively, and the performance of EHR-based machine learning models relative to standard care remains unclear.

Do EHR-based machine learning models improve the detection of undiagnosed atrial fibrillation in adults in primary care settings compared to standard care or validated risk scores?

Structured PICO

Do EHR-based machine learning models improve the detection of undiagnosed atrial fibrillation in adults in primary care settings compared to standard care or validated risk scores?

Population

Adults (aged >18 years) in primary care or outpatient settings managed by healthcare providers (e.g., general practitioners, family physicians, nurse practitioners). 16 studies were included in the systematic review.

Intervention

Machine learning (ML) models developed and/or applied using data from electronic health records (EHRs) or primary care databases to detect new cases of atrial fibrillation.

Comparator

Standard care practices or validated clinical risk scores for atrial fibrillation (e.g., CHARGE-AF, logistic regression models, clinician assessment).

Outcome

Incidence of new atrial fibrillation diagnoses (detection of undiagnosed AF) following the application of ML methods.

EHR-based machine learning models show promise for improving the detection of undiagnosed atrial fibrillation in primary care, but clinical integration is currently limited by high risk of bias and a lack of external validation.

Limitations

Insufficient external validation (only 4 studies)
High risk of bias (53% of included studies)
Variable performance across studies
Heterogeneity among studies precluding a meta-analysis

Implication

May aid undiagnosed AF detection in primary care; leaves open adoption without bias reduction and external validation.

Abstract

Atrial fibrillation (AF) increases the risk of stroke, heart failure and mortality. Current screening guidelines fail to detect AF effectively, and existing models have limited applicability in primary care. Electronic health records (EHRs) provide an opportunity to apply machine learning (ML) for automated AF detection; however, their performance relative to standard care remains unclear. We conducted a systematic review to evaluate the effectiveness, quality, and applicability of EHR-based ML models for detecting AF in primary care. The review is informed by Joanna Briggs Institute and Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) guidelines. We searched seven databases from inception to May 2023. Eligible studies involved adults in primary care where ML models using EHRs were compared to standard care. The primary outcome was the detection of undiagnosed AF; secondary outcomes examined impacts on patients, healthcare providers, and systems. Data were extracted using CHARMS, risk of bias and applicability were evaluated through PROBAST and MI-CLAIM checklists. This review was registered in International Prospective Register of Systematic Reviews (CRD42023390603). From 4,536 references screened, 16 studies were included. Among these, 14 (87%) were retrospective cohort studies, one (6%) was prospective, and one (6%) was a randomized controlled trial. Random forest classifiers were the most common ML model (7 studies, 43%). Only 4 studies (25%) underwent external validation, and 8 (53%) were at high risk of bias. Model discrimination (AUROC) ranged from 0.71 to 0.948, with 8 (50%) outperforming controls. Combining ML with clinical tools (3 studies, 19%) significantly improved discrimination compared to ML models alone. Reviewed models identified gout as a nontraditional predictor of AF and demonstrated that dynamic measures of BMI, blood pressure, and heart failure diagnosis were stronger predictors than static measures. EHR-based ML models show promise for improving AF detection in primary care compared to standard care. Their clinical applicability, however, is limited by insufficient external validation, high risk of bias, and variable performance. Future research should prioritize external validation, evaluation in clinical trials and the integration of predictors routinely available in primary care.