This AI Could Save Millions Of Lives
This AI Could Save Millions Of Lives
But, it could pose a risk.
Here is a fact I bet you didn’t know. A quarter of the deaths in the UK are attributed to heart failure. In fact, this is true for almost all developed nations. Why? Well, heart failure can be caused by a multitude of different ailments, many of which can go completely undiagnosed for years, as symptoms only become apparent down the line. As such, many are diagnosed too late for effective treatment, leading to these sky-high death rates. But, an AI has just been developed to solve this troubling problem.
Let’s start with what heart failure actually is. Heart failure means that the heart is unable to pump blood around the body properly, typically because it has become too weak or stiff. As such, the heart is still technically working but requires support, such as drugs, corrective operations or pacemakers, to function correctly. As I stated, heart failure can have multiple potential causes, but the most common are high blood pressure, coronary heart disease, congenital heart disease, arrhythmia, valve damage, obesity, alcohol abuse, anaemia and muscle diseases.
All of these potential causes can be challenging to diagnose early on, but nonetheless leave tell-tale traces in a person’s medical record. For example, a patient can easily hide their alcohol addiction from their doctor, but the types of illness they become susceptible to and the length of time it takes them to recover can point to their addiction. Sadly, doctors simply aren’t equipped routinely to identify these trends.
But an AI would be perfect for this.
Hence, why, doctors and researchers have developed FIND-HF, an AI which shifts through health records and identifies patients at risk of heart failure early. The AI was trained on the medical records of 565,284 UK adults and 106,026 records from Taiwan National University Hospital.
This AI has now been tested and found to accurately predict who is at the highest risk of developing heart failure and who could be hospitalised by it in the next five years. This means that FIND-HF could bring the average diagnosis of high heart failure risk forward by two years by suggesting which patients should be brought in and tested for potential heart failure risk. These extra two years is plenty of time to implement highly effective early-stage treatment, which could dramatically reduce death rates and increase the quality of life for those at risk of heart failure.
This is a fantastic example of AI used correctly. False positives, or hallucinated trends, problems which plague even highly advanced AIs, won’t have a negative impact on people’s health. It isn’t making a diagnosis in place of a doctor. Instead, it helps to find those who most likely need additional medical examinations by professionals.
AI like this does have its risks, though. Training the AI on huge volumes of confidential data and then transferring and processing entire national databases of confidential data poses serious privacy risks. Every time you have to move data or pass it through an external program, the risk of a data leak increases. What’s more, the AI industry has a history of being sloppy with data privacy, as they have to manage such vast amounts of data that doing so correctly can be financially unfeasible. For example, Microsoft’s AI research team accidentally leaked 38 terabytes of private data that contained sensitive data.
But, surely, such a risk is worth the potential upside? FIND-HF being deployed nationwide could save, extend and improve the quality of millions of lives over time. As such, the researchers behind FIND-HF are now looking to develop it further by inviting those identified in primary care records as being at the highest risk to be assessed for heart failure. If the results of this next stage look promising, which I can’t see why they wouldn’t, the team hops to be able to deploy the system throughout the UK’s NHS.
Thanks for reading! Content like this doesn’t happen without your support. So, if you want to see more like this, don’t forget to Subscribe and follow me on BlueSky or X and help get the word out by hitting the share button below.