Prior to this, the Artificial Intelligence (AI) has proven to be highly efficient and accurate in predicting various diseases and is capable of even predicting the mortality risk in intensive care patients.
How Artificial intelligence predicts deaths in heart patients?
Scientists at the Francis Crick Institute in London, England, developed a new algorithm which is able to predict the mortality risk in patients with heart disease more accurately than the models that were developed by specialists.
Andrew J. Steele, the lead author of the study, revealed in a statement that more than 80,000 patients electronic data was used to develop the algorithm, and his research team was supported by the Farr Institute for Health Informatics Research and the NHS Foundation Trust at University College London.
Specifically, the AI model was based on cardiovascular diseases, which are currently the leading cause of death in the UK. Steele and his colleagues trained the system to make predictions based on at least 600 variables such as physical complaints, gender, and age.
Steele’s main conclusion that reached through his study is that his new AI has achieved a better algorithm than any previous model that tries to predict mortality and has also been able to create its own variables, such as home visits by doctors.
Even though the model was only developed for use in research among competing experts, Steele did not rule out the possibility that it could be used in the future as a clinical practice tool.
The scientific community in this country is not the only one working to prevent deaths from heart disease. In early August, we reported on a partnership between the National Health Service (NHS), the British Heart Foundation (BHF) and Microsoft to map the location of defibrillators to save thousands of patients from heart failure.
Reference: Machine learning models in electronic health records can exceed traditional survival models for predicting patient mortality in coronary heart disease. PLOS One, Released: August 31, 2018, DOI: https://doi.org/10.1371/journal.pone.0202344