The Harvard-Emory ECG Database

Area of Science:

• Cardiology
• Medical Informatics
• Biomedical Data Science

Background:

• The Harvard-Emory ECG Database (HEEDB) is a collaborative effort between Harvard University and Emory University.
• It comprises a vast collection of 12-lead electrocardiogram (ECG) recordings gathered from clinical settings since the early 1990s.
• The database continuously expands with new data, ensuring its relevance and comprehensiveness.

Purpose of the Study:

• To establish a large-scale, publicly accessible repository of 12-lead ECG recordings.
• To provide rich annotations including diagnoses, morphology, and rhythms for research and development.
• To facilitate advancements in cardiovascular research through comprehensive data sharing.

Main Methods:

• Collection of over 10 million 12-lead ECG recordings from Massachusetts General Hospital and Emory University Hospital.
• Standardized storage of ECGs in WFDB format, sampled at 250 and 500 Hz.
• Inclusion of 12SL annotations for a significant portion of the recordings, covering diagnoses, morphology, and rhythms.

Main Results:

• The HEEDB contains 10,608,417 ECG recordings from 1,818,247 patients (MGH) and 1,452,964 recordings from 552,481 patients (EUH).
• Detailed 12SL annotations are available for 10,471,531 MGH recordings and 1,268,277 EUH recordings.
• Future updates will incorporate demographic data and clinical codes (ICD-9/10, CPT) and medication history.

Conclusions:

• The HEEDB represents a significant resource for the medical and research community.
• Its extensive data and annotations will drive innovation in ECG interpretation and cardiovascular disease research.
• The ongoing expansion and planned data additions will further enhance its utility for diverse biomedical applications.