AI-powered study backs anticoagulants for heart failure patients without AF
A new study has found that blood thinning drugs improve outcomes for all heart failure patients, regardless of whether they already have atrial fibrillation (AF).
Healthcare AI company Sensyne Health analysed anonymised, routinely collected data from nearly half a million NHS patients as part of the study, then compared the data over three years of heart failure patients with no record of anticoagulant prescription with that of patients prescribed either warfarin or NOACs – the two most commonly used classes of anticoagulants.
The benefit of anticoagulants in patients with atrial fibrillation is well documented, but the data in the study also suggested a small but statistically significant survival benefit for heart failure patients without atrial fibrillation on anticoagulants.
However the company said that further analysis is required to confirm if the difference is clinically relevant, and if so which subgroups of patients would benefit the most.
The analysis also suggests that there may be relevant differences, not only between the use of warfarin compared to NOACs, but also between the NOACs currently in use.
Heart failure is a highly heterogeneous disease which affects more than one million UK adults, is one of the most common causes of hospital admission and is responsible for 10,000 deaths a year. It severely compromises a patient’s quality of life, and treating it comes with significant costs, amounting to 1-2% of the NHS’ annual budget (c. £625 million).
AI-powered study
Notably, the study used both real world data and Sensyne’s machine learning patient stratification algorithm, which identifies different subgroups of heart failure patients. This allowed the company to look at a more diverse set of patients than might traditionally be possible in a randomised controlled trial.
Explaining the benefits of their approach, Sensyne said: “In silico analysis of the real-world data in electronic patient records offers the enticing potential of a faster, more representative and cost-effective alternative.
“These findings are interesting in their own right, but the combination of these standard, statistical approaches with the application of Sensyne Health’s machine learning algorithm... is where the opportunities lie for enhanced analysis, insights and clinical understanding.”
Sensyne has a unique partnership with NHS Trusts that enables it to analyse ethically sourced, anonymised patient data to undertake such research.
Sir Bruce Keogh, interim chairman of Sensyne Health and former national medical director of the NHS Commissioning Board, said that the partnership “allows rapid and cost-effective analysis of therapeutic efficacy outside the confines of randomised clinical trials in a way that reflects the reality of routine clinical practice” and “offers the prospect of new insights leading to iterative improvements in healthcare”.
