Using ECG Data to Predict Overall Health


A Mayo Clinic study examining artificial intelligence to analyze ECG data found readings can be used as a predictor of overall physiological health.

Suraj Kapa, MD

Suraj Kapa, MD

As technology in medicine continues to advance, physicians are constantly looking for way to effectively incorporate that technology into practice in a way that benefits their patients.

A recent Mayo Clinic-funded study has found applying artificial intelligence to standard 12-lead electrocardiogram (ECG) data could be a way to measure the overall health status of a patient.

“Our standard diagnostic tools may have far more information behind them then we’ve come to expect through our standard approaches to diagnostic interpretation,” explained Suraj Kapa, MD, study investigator and cardiac electrophysiologist at the Mayo Clinic in Minnesota.

In order to determine the effectiveness of artificial intelligence in predicting a person’s chronological age and sex using only 12-lead ECG signals, investigators developed a convolutional neural network (CNN) to analyze the data. The CNN developed by investigators incorporated ECG data from 499,727 patients to predict sex and age. After, the CNN was tested on a separate cohort that contained 275,056 patients. Investigators performed additional analyses on 100 randomly selected patients with multiple ECGs to assess within-individual accuracy of CNN age estimations.

Of the 275,056 patient cohort, 52% were male and the mean age of the group was 58.6 years. When identifying patient sex, the CNN achieved 90.4% accuracy with an area under the curve of 0.97.

Investigators noted that age was estimated as a continuous variable with an average error of 6.9±5.6 years (R-squared =0.7). Among a smaller group of 100 patients with multiple ECGs over the course of at least 2 decades, 51% of patients had an average error between real age and CNN-predicted age less than 7 years.

In an interview with MD Magazine, Kapa pointed out that while the ECG appears to be effective in predicting age and sex, it may be the incorrect readings that provide physicians the most information on a patient’s overall health.

“Being able to give an alert that there may be something going on that is making them “age faster” as per the AI derived age, we may be able to alert individuals to be further assessed for subclinical diseases that haven’t driven them to seek medical care yet,” Kapa said. “Earlier intervention targeted towards those who most need it, in turn, may improve health at a broader scale.”

Among patients whose CNN-predicted ages exceeded chronological age by more than 7 years, investigators noted an increase in low ejection fraction, hypertension, and coronary disease (P<0.01). In the 27% of patients where correlation was greater than 0.8 between CNN-predicted and chronologic age, investigators found no incidents occurred during the follow-up period (33 years).

This study, “Age and Sex Estimation Using Artificial Intelligence From Standard 12-Lead ECGs,” was published online in Circulation: Arrhythmia and Electrophysiology, a journal of the American Heart Association.

Related Videos
Tailoring Chest Pain Diagnostics to Patients, with Kyle Fortman, PA-C, MBA
Solutions to Prevent Climate Change-Related Illness, with Janelle Bludhorn, PA-C
Kyle Fortman, PA-C, MBA: Troponin and Heart Injury Risk Screening Recommendations
What Should the American Academy of Physician Associates Focus on in 2025?
The Rising Rate of Heat-Related Illness, with Janelle Bludhorn, PA-C
Video 6 - "Evaluating Safety of Novel LDL Management Mechanism"
Video 5 - "Optimizing PCSK9 Inhibitors and Analyzing Plaque Reduction Data"
Video 4 - "Innovations in Small Interfering RNA (siRNA) Therapy"
Video 3 - "Ongoing Lp(a) Trials and Clinical Approaches to Treatment"
Roger S. McIntyre, MD: GLP-1 Agonists for Psychiatry?
© 2024 MJH Life Sciences

All rights reserved.