Image source: Adobe Stock/TuMeggy

News • Enhancing diagnostics

Al and robots could help detect urinary tract infections early

Scientists are developing artificial intelligence (AI) and talking robots to help to detect urinary tract infections (UTIs) in vulnerable people early.

The £1.1 million FEATHER project, led by the School of Informatics, aims to increase an individual’s wellbeing and reduce the number of serious outcomes from incorrect or late diagnoses of urinary tract infections. Researchers hope that problem solving techniques – combined with AI and robotics – will detect infections early and lessen the over-prescription of antibiotics for the condition. 

A team of experts from the Universities of Edinburgh and Heriot-Watt are working with Blackwood Homes and Care and Leuchie House to develop AI computer programmes and interactive robots to alert clinicians of a potential UTI in residents early. Sensors, placed in residents’ rooms will gather data about their daily activities. The technology will be able to spot changes in a person’s behaviour or activity levels, which will then trigger an interaction with the robot. Behaviour changes could include variations in walking pace and movement, altered sleeping patterns and increased visits to the bathroom. The FEATHER platform will combine and analyse this data to flag potential infection signs, before an individual or carer is even aware there is a problem. 

Studies show that there is a significant association between delirium and UTI in older adults and, while it is possible that carers will pick up these signs, we should not be relying on observations alone

Lynne Baillie

UTIs are one of the most common types of infection, affecting around 150 million people worldwide annually. Early signs of a UTI can also be challenging to recognise because the symptoms – including pain, temperature, frequency of urination, changes in sleep patterns and tremors – vary according to a person’s age and existing health conditions. Diagnosis can be difficult, with lab analysis, a process taking up to 48 hours, providing the only definitive result. By raising the alert early, the project could enable a GP to have the time to wait for lab results from urine tests, to ensure the appropriate prescription of antibiotics, reducing the cost to the NHS and improving outcomes for the patient. 

Early diagnosis should also reduce the number of patients taken to A&E and reduce the number of cases developing to sepsis, kidney failure and even loss of life. The AI and implementation aspects of the project will be led by Professor Kia Nazarpour, Dr Nigel Goddard and Dr Lynda Webb from Edinburgh’s School of Informatics. The interaction between humans and robots will be led by Professor Lynne Baillie, assisted by Dr. Mauro Dragone, from Heriot-Watt University. The technology will be developed and tested in the new Assisted Living Lab at the National Robotarium, where the team will run extensive trials in a realistic social care setting. 

The research has received funding from UK Government by the Engineering and Physical Sciences Research Council, part of UK Research and Innovation, and the National Institute for Health and Care Research. 

"This unique data platform will help individuals, carers and clinicians to recognise the signs of potential urinary tract infections far earlier, helping to prompt the investigations and medical tests needed," says Professor Nazarpour. "Earlier detection makes timely treatment possible, improving outcomes for patients, lowering the number of people presenting at A&E, and reducing costs to the NHS." Professor Baillie adds: "We hope this work will create an additional structured support mechanism for people who live independently. Studies show that there is a significant association between delirium and UTI in older adults and, while it is possible that carers will pick up these signs, we should not be relying on observations alone. We are working with stakeholders to co-design the robot interaction and data collection for the machine learning methods to better support longer and healthier independent living."

Kitty Walker, a care receiver and regular guest at Leuchie House, who has experiened the condition herself, says: "The impact of having a UTI can be far more serious than a lot of people may realise. Commonly, my speech becomes affected which can make it difficult to communicate with people like I normally would. More seriously, I’ve been hospitalised in the past after the late diagnosis of a UTI led to me having a seizure and I required mouth-to-mouth resuscitation. Being able to spot the early indicators that I have a UTI would save any anxiety I might feel when I know there is a problem and help reduce the number of different antibiotics I need to take." 

"Understanding how socially assistive AI can be used to better detect UTIs has the potential to improve the health & wellbeing of our customers," says Colin Foskett, Head of Innovation & Research at Blackwood Homes and Care. "Early UTI detection could prevent hospital admissions, associated decline and ensure people can continue to live independently for longer." Mark Bevan, CEO at Leuchie House, adds: The personal, health and financial cost of urinary infections are massive, costing the NHS at least £500M last year, devastating people’s lives and adding great complexity to the provision of increasingly complex adult care." 


Source: University of Edinburgh

11.12.2022

Related articles

Photo

News • Study shows lack of reliability

Can an AI predict Covid-19 from cough sounds? Not really, study finds

Predicting a Covid-19 infection from the sound of a cough? Researchers found that technology using Machine Learning performed no better than simply asking people to report their symptoms.

Photo

News • Experts propose combination strategy

Kidney cancer: early detection through 'piggyback' screening

Kidney cancer is relatively rare, so dedicated screening programmes would not be cost-effective. Combined with lung cancer screening, however, this would be much more feasible, a new study shows.

Photo

News • Sensor solution

Wearable lung patch uses deep learning to detect asthma and COPD

Early detection and management of asthma and COPD is critical. US researchers have developed a deep learning model paired with a wearable sensor patch to automatically detect wheezing sounds.

Related products

Subscribe to Newsletter