About HeartLab

The landmark study validating wearable ECG for atrial fibrillation detection. 419,297 participants. Positive predictive value of Apple Watch: 84%. This study proved consumer wearables can be a legitimate screening tool for AFib.

Head-to-head comparison showing 94.5% sensitivity and 95.7% specificity for AFib detection. Apple Watch correctly classified sinus rhythm in 99.3% of cases.

Deep learning models achieved cardiologist-level accuracy for 12 rhythm classes. PVC detection sensitivity: 93.2%. PAC detection: 88.7%. Single-lead wearable ECGs can detect far more than originally thought.

500 patients wore both Apple Watch and Holter for 7 days. AFib detection: 97% sensitivity (Watch, 7-day) vs 89% (Holter, 24h). PVC burden correlation: r=0.82. For AFib screening, a week of smartwatch ECGs may outperform traditional 24h Holter.

A 34-layer convolutional neural network that matches or exceeds the accuracy of board-certified cardiologists in classifying 12 different heart rhythm types from single-lead ECG data. This foundational work demonstrated that AI can reliably interpret single-lead ECG — the same signal type used by Apple Watch.

Systematic review validating that consumer wearable-derived HRV metrics (RMSSD, SDNN) show strong agreement with clinical-grade equipment (r > 0.90). Confirms that R-R interval data from wearable ECGs is reliable for HRV computation.

QTc > 500 ms carries a 2-3x increased risk of Torsades de Pointes. Drug-induced QT prolongation causes ~15,000 deaths/year in the US. Over 200 medications can prolong the QT interval. HeartLab calculates QTc using four validated formulas: Bazett (1920), Fridericia (1920), Framingham (1992), and Hodges (1983).

LQTS prevalence: 1 in 2,000 (many undiagnosed). Apple Watch QTc measurement accuracy: ±15 ms vs 12-lead. Serial QTc measurements improve diagnostic accuracy — exactly the approach HeartLab uses with longitudinal QTc trend tracking.

Comprehensive comparison of Bazett, Fridericia, Framingham, and Hodges correction formulas across 13,000+ ECGs. Fridericia showed the best rate-independence (r=0.04 vs r=0.32 for Bazett). HeartLab implements all four formulas, with Fridericia as default for its superior accuracy across heart rate ranges.

The authoritative database of medications known to prolong the QT interval, maintained by clinical pharmacologists. Over 200 drugs listed with risk categories (Known Risk, Possible Risk, Conditional Risk). HeartLab's QTc monitoring is especially valuable for patients taking medications on this registry.

PVCs occur in 75% of healthy individuals on 24h Holter monitoring. PVC burden < 10% is generally benign; burden > 15-20% carries risk of PVC-induced cardiomyopathy. HeartLab counts PVC burden per recording and tracks trends over time.

Low HRV independently predicts cardiac mortality. Post-MI patients with low HRV have 3.2x higher risk of sudden death. HeartLab derives HRV metrics (SDNN, RMSSD) directly from R-R interval analysis of each ECG recording.

Demonstrated that frequent PVCs (>24% burden) can cause reversible cardiomyopathy. Catheter ablation of the PVC focus led to normalization of left ventricular function within 6 months in 82% of patients. Underscores the clinical importance of PVC burden tracking — a core HeartLab feature.

The international standard for heart rate variability measurement and interpretation. Defines time-domain (SDNN, RMSSD, pNN50) and frequency-domain (LF, HF, LF/HF ratio) metrics. HeartLab computes HRV following these validated standards for each ECG recording.

HeartLab's R-peak detection is based on the Pan-Tompkins algorithm — the gold standard in QRS detection — enhanced with adaptive amplitude thresholds, median-based filtering, and quality-aware processing optimized for single-lead Apple Watch ECG data.

HeartLab's ML-based beat classifier uses relative heart rate information and morphological analysis to classify PVCs, PACs, and normal beats. Trained and validated against the MIT-BIH Arrhythmia Database — the standard benchmark for ECG algorithm validation.

HeartLab follows the AHA 2024 guidelines for arrhythmia classification and detection thresholds. QTc normal ranges, PVC burden risk stratification, and AFib detection criteria all align with current clinical consensus.

Framework for automated ECG signal quality assessment using multiple quality indicators. HeartLab implements a 6-factor signal quality analysis (baseline wander, noise level, lead-off detection, QRS amplitude, signal saturation, and heart rate plausibility) to ensure only reliable recordings are analyzed.

Validation of the CHA₂DS₂-VASc score for predicting stroke risk in atrial fibrillation patients. Recommended by ESC and AHA/ACC guidelines to guide anticoagulation therapy. HeartLab's free QTc Calculator and CHA₂DS₂-VASc tools implement these validated scoring systems.