The possible uses for artificial intelligence are almost unlimited. British researchers have now developed an AI model that can even predict death using ECG data.
Artificial intelligence is also increasingly being used in the healthcare system. British researchers at Imperial College London have now developed an AI model that can use ECG data to predict health risks.
But according to a statement from the university The AI model can not only detect diseases, but also their progression and severity. The researchers can even determine the risk of early death thanks to their AI model.
AI uses ECG data for health predictions
Researchers from Imperial College London and Imperial College Healthcare NHS Trust collaborated on the study, which was published in the journal Lancet Digital Health. They used data from millions of ECGs from international sources to train their AI model.
The aim was to be able to make accurate predictions “in which patients a new disease will develop, the disease will worsen or who will die later”. The AI system has been trained to recognize patterns in the electrical signals of the ECG data. Artificial intelligence can determine and understand more complex and subtle ECG patterns – better than even a cardiologist.
“We cardiologists use our experience and standard guidelines when we look at ECGs and sort them into ‘normal’ and ‘abnormal’ patterns to help us diagnose diseases,” explains research leader Arunashis Sau. “However, the AI model detects much more subtle details, allowing it to ‘discover’ problems in ECGs that would appear normal to us, potentially long before the disease fully develops.”
How accurate are the system’s predictions?
In their study, the researchers were able to correctly determine the risk of death in the ten years after the ECG in 78 percent of cases. In the remaining cases, there could have been an influence from unrecognizable factors, such as treatment of the patient following the ECG.
The AI system called AI-ECG risk estimator – AIRE for short – can predict future health risks such as cardiac arrhythmias, heart attacks and heart failure. But causes of death that are not heart-related could also be recognized “with a high degree of accuracy” by the AI model.
EKGs collect a variety of information from throughout the body, since diseases like diabetes, which affect organs like the kidneys or liver, also affect the heart in some way. Our analysis shows that AI can tell a lot not just about the heart, but also about what’s going on elsewhere in the body, and could be able to detect accelerated aging.
The researchers estimate that the new AI model could be used in the public health service within the next five years. However, clinical studies are carried out beforehand. According to Arunashis Sau, this is “the next important step”. The researchers want to test whether “using these models can actually improve patient outcomes.” The clinical studies are already being planned and are scheduled to start by mid-2025.
Also interesting:
- Character.ai: Did a chatbot drive a teenager to suicide?
- Letter of protest against AI training: actors and musicians take to the barricades
- With AI influencer Emma: German Tourism Center embarrasses itself
- AI creates completely new proteins – a miracle cure for medicine?
The article With ECG data: This AI should be able to predict death by Maria Gramsch first appeared on BASIC thinking. Follow us too Facebook, Twitter and Instagram.
As a Tech Industry expert, I believe that using ECG data to predict death with the help of AI can be a groundbreaking advancement in healthcare. By analyzing the intricate patterns and signals in ECG data, AI algorithms can potentially identify patterns and indicators that may suggest an increased risk of mortality.
This technology has the potential to revolutionize patient care by enabling healthcare providers to intervene earlier and potentially prevent life-threatening events. However, it is crucial to ensure that the AI models are accurate, reliable, and ethically sound. The use of sensitive health data like ECG information must also be handled with the utmost care to protect patient privacy and confidentiality.
Overall, I see great potential in using AI with ECG data to predict death, but it is important to approach this technology with caution and ensure that it is used responsibly for the benefit of patients and healthcare providers.
Credits