"An AI system has shown its life-saving potential by notifying doctors of patients with heart test results indicating a high risk of mortality. This AI intervention led to a 31% reduction in overall fatalities among high-risk patients in a randomized clinical trial involving almost 16,000 individuals across two hospitals."

Eric Topol, from the Scripps Research Translational Institute in California, who was not involved in the research, commented, "This is truly remarkable." "It's highly unusual for any medication to achieve a 31% reduction in mortality, let alone a non-pharmaceutical intervention—simply AI monitoring individuals."

Figure 1. Training an AI to Predict Fatal Heart Issues Using Electrocardiogram Tests.

Figure 1 shows Training an AI to Predict Fatal Heart Issues Using Electrocardiogram Tests The AI, created by Chin Lin and his team at the National Défense Medical Centre in Taiwan, underwent initial training using more than 450,000 electrocardiogram (ECG) tests. These tests record the heart's electrical activity along with survival data from the ECG subjects. Each patient's mortality risk was quantified as a percentile score. The AI was trained to recognize individuals in the top 5th percentile, indicating high risk.[1]

The researchers tested the AI alarm system on 39 doctors across two hospitals. Whenever a nurse uploaded a new patient's ECG test results to a computer server, the AI system would promptly analyse the data and notify doctors if it identified the patient as high risk.

Not only did the AI system reduce the likelihood of mortality from all causes, but it also reduced the death rate from heart issues in high-risk individuals by nearly 90%. According to Lin and his team, the AI predictions could assist in focusing attention on the most at-risk group, as doctors usually responded to alerts by conducting further diagnostic testing and implementing more therapy.

In Taiwan, 14 military hospitals have already implemented the AI alert system. According to Topol, "This could be adopted by every hospital globally—it doesn't have to be costly." "When witnessing such significant benefits, it should become the norm in healthcare."[2]

Reference:

1. https://www.indianext.co.in/ai-that-assesses-mortality-risk-saves-lives-during-a-hospital-experiment/
2. https://www.newscientist.com/article/2428674-ai-that-determines-risk-of-death-helps-save-lives-in-hospital-trial/

Cite this article:

Gokila G (2024), AI Risk Assessment System Proves Life-Saving in Hospital Trial, Anatechmaz, pp.398