Smartphone app may be able to detect stroke symptoms, according to early research  

The study involved 269 patients from 4 stroke centres in Bulgaria, all of whom had been diagnosed with ‘acute stroke’. The app was tested on these patients to see how accurate it was. Researchers found that the Fast.AI app correctly identified facial asymmetry with more than 97% accuracy, and arm weakness with more than 72% accuracy. The study authors say that there are early signs that it can also detect slurred speech, but this has yet to be tested. 

The app is still being trialled and is not yet available for the public to use. We contacted the Stroke Association to ask for their insight into the potential benefits and risks of an app such as this.  

Dr Richard Francis, Head of Research at the Stroke Association told us: “Technologies such as this app could prevent unnecessary journeys to hospitals or ensure stroke patients are admitted to the right specialist treatment units to receive life-saving treatments. 

“Telemedicine has been developing over the years for the detection and treatment of stroke, and has the support of researchers and clinicians alike”.

We don’t yet know how much of a game-changer this app could be. Dr Francis warned: “Machine learning is only as good as the data you put into it.” 

Machine learning, which is a branch of artificial intelligence, is based on supplying a computer algorithm with data for it to learn from. That means its knowledge is entirely based on the information it is supplied with. For example, there have been previous warnings that the use of AI for diagnosis could lead to women or ethnic minorities being disadvantaged, unless care is taken to avoid any biases in the data that is used to “teach” the apps.  

Dr Francis adds: “For a technology like this to be useful, it has to be tested on a representative sample of people affected, and it would be useful to see more research.”

This study has not yet been published in a scientific journal or been reviewed by other scientists, so we don’t have the full picture of the study findings or what the limitations of the research might be. One of the study authors is a co-founder of Neuronics Medical, the company which owns the app. 

Dr Francis also reminded us of the importance of fast action when someone may have had a stroke.  

“Someone has a stroke every five minutes in the UK. Stroke is when the blood flow to parts of your brain is cut off, by a blockage in an artery or by a bleed on the brain. Stroke strikes without warning and can happen to anyone, leading to devastating effects. The good news is that stroke is treatable. Stroke is a medical emergency and the quicker you get treatment, the better your recovery could be. To have the best chance of the best recovery, you need to be assessed quickly, and treated appropriately to quickly restore blood flow to the brain. 1.3 million brain cells die for every minute stroke goes untreated, so time lost is brain lost.” 

The name of the app comes from the acronym which is used to remind us how to detect warning signs of a stroke. As Dr Francis says: 

“To know the signs, use the FAST test:  

• F is for face, can the person smile? Has their mouth or eye drooped?  
• A is for arms, can the person raise both arms?  
• S is for speech, can the person speak clearly and understand what you say?  
• T is for time, time to call 999 if you see any single sign of stroke.” 

“Stroke is a medical emergency. The sooner you get to hospital, the better the outcome.” 

There are other symptoms to be aware of as well, which could potentially indicate stroke:

• Sudden weakness or numbness on one side of the body, including legs, hands or feet. 
• Difficulty finding words or speaking in clear sentences. 
• Sudden blurred vision or loss of sight in one or both eyes. 
• Sudden memory loss or confusion, and dizziness or a sudden fall. 
• A sudden, severe headache.