Creating a large searchable library of ECG heart traces

Electrocardiographic pattern classification for automatic repolarisation assessment

An electrocardiogram – commonly called an ECG – is a routine test that records the rhythm and electrical activity of the heart. Most healthcare professionals will be familiar with the ‘hills and valleys’ pattern of a normal ECG reading. But when people have heart conditions, or are taking certain medicines, their heart’s electrical pattern may be subtly different. It is vital that the different heart rhythm patterns are interpreted accurately and consistently, so that doctors and their patients can make the best decisions about treatment. Professor Malik has been granted unique access to a collection of ECG recordings from more than 9 million people. The collection belongs to the US Food and Drug Administration (FDA), which is the American drug approval agency. The BHF are funding Professor Malik to organise the data, and use it to build a self-learning ECG measurement system, creating a clinical reference tool for almost every possible variation in ECGs. This could benefit global heart health long into the future by providing a searchable database that can match any patient’s ECG with similar readouts that have been accurately interpreted. Such a comprehensive resource will be far better than anything that exists today, and will help healthcare professionals to better understand and act on their patients’ ECG readings.