Artificial Intelligence Identifies Five Types of Heart Failure
An artificial intelligence (AI) study led by the UCL Institute of Health Informatics has identified five different types of heart failure that could be used to predict future risks for individual patients. The study used anonymised patient data from over 300,000 UK adults diagnosed with heart failure over 20 years. The research identified five subtypes using multiple machine learning methods and datasets: early onset, late onset, metabolic, cardiometabolic and atrial fibrillation-related. The different subtypes exhibited a range of risks of mortality within a year of diagnosis. The researchers also developed an app that clinicians could use to determine which subtype a patient has.
Improved Classification Could Lead to More Targeted Treatments
Current classification systems for heart failure do not reliably predict how the condition will progress. The study’s lead author, Professor Amitava Banerjee, said that better distinctions between types of heart failure could lead to more targeted treatments and help to think in a different way about potential therapies. The five robust subtypes identified could lead to more accurate predictions of risk and improved information for clinicians. The next step is to evaluate whether the app designed by the researchers can make a practical difference to patients and be cost-effective. The study was published in Lancet Digital Health.
Researchers Used Four Machine Learning Methods to Avoid Bias
To avoid bias from a single machine learning method, the researchers used four separate methods to group cases of heart failure and applied these methods to data from two large UK primary care datasets. These datasets were representative of the UK population and were linked to hospital admissions and death records. The research team trained the machine learning tools on segments of the data and, once they had selected the most robust subtypes, they validated these groupings using a separate dataset. The subtypes were established on the basis of 87 factors, including age, symptoms, the presence of other conditions, the medications the patient was taking, and the results of tests and assessments.
Link Found Between Particular Subtypes of Heart Failure and Polygenic Risk Scores
The study also examined genetic data from 9,573 individuals with heart failure from the UK Biobank study. Researchers found a link between particular subtypes of heart failure and higher polygenic risk scores for conditions such as hypertension and atrial fibrillation. The research team believes that the app designed could help to inform discussions with patients and potentially improve predictions of future risk. However, further research is needed to evaluate its effectiveness in routine care.
Conclusion
The study’s identification of five subtypes of heart failure that can be used to predict future risks for individual patients represents a significant step forward in the classification of this condition. The development of an app that clinicians can use to determine which subtype a patient has could lead to more accurate predictions of risk and improve information for patients. The link found between particular subtypes of heart failure and higher polygenic risk scores for conditions such as hypertension and atrial fibrillation could also help to inform treatment decisions. However, further research is needed to evaluate the app’s effectiveness in routine care.
- Heart failure diagnosis AI tools
- AI-based heart failure classification
- Heart failure prediction using AI
- Machine learning in heart failure detection
- AI technology for heart failure management
News Source : ThePrint
Source Link :AI tools discovered five types of heart failure – ThePrint –/