AI software to improve hospital patient turnaround times
Kavitha Vimalesvaran is a cardiology registrar and a doctoral candidate at Imperial College London, where she is developing AI-assisted clinical decision support systems for cardiac MRI. Vimalesvaran is also a Clinical AI Fellow at the AI Centre for Value Based Healthcare at Guy’s and St Thomas’ Hospital, where she is leading a national clinical trial for the implementation of an AI-based software that can diagnose critical brain abnormalities from Head CT scans.
“The goal is to assess whether tools like these can improve patient turnaround times, particularly in the Emergency Department,” she says.
Here, she tells us more.
Hi Kavitha! What has your career journey been like in the healthcare sector?
“My career in healthcare has been a real adventure with lots of unexpected turns. I did my Masters in Cardiovascular Research at King’s College London and then moved to specialise in Cardiology in Northwest London. I got particularly interested in Cardiac Imaging which led me to a PhD with the UKRI Centre for Doctoral Training in AI for Healthcare at Imperial. I was programming naïve when I started but now, I get to design and test AI tools to improve Cardiac MRI scanning, which has been a fantastic experience.
“When I saw the Clinical AI Fellowships advertisement, I just knew it was the perfect match for what I had been working on in my PhD. Though I never thought, as a cardiologist, I would be leading trials involving head CTs for patients presenting to the ED!
“Nevertheless, the fellowship has helped me gain invaluable insights about AI in healthcare, managing different stakeholders and navigating the complexities of the NHS systems. I must also say that I have met the most incredible and inspiring people along the way who continue to shape my career.”
Tell us about your work with AI.
“My work with AI has been unexpected and challenging but hugely fulfilling. Through my PhD I have had the opportunity to design, train and test novel AI models in cardiac imaging, where the fewest images are acquired to detect pathology. In my role as a clinical AI fellow on the other hand, our clinical trial – ACCEPT-AI – aims to evaluate the impact of a commercial AI tool which can detect multiple brain abnormalities on Head CTs. It also produces a dummy report whilst also assigning a priority level in the radiology reporting system. Ultimately, I would like to close the gap between AI research and actually using it in everyday healthcare.”
What is AIDE?
“AIDE or the AI Deployment Engine is like the Apple “App Store” but for healthcare providers. It is a platform that lets them use AI models safely and efficiently in the clinical workflows. So, you can think of it like a unified infrastructure that lets you run several algorithms simultaneously through bespoke Application Programming Interfaces (APIs). It helps NHS Trusts to use multiple AI solutions on a single platform for everyday patient care. It is already live in several NHS trusts including King’s College Hospital NHS Foundation Trust and we hope to have it up and running in all major London hospitals soon.”
What do the next 12 months hold for you?
“Well, I think the next year is going to be quite packed for me. I’m on the last stretch of my PhD, hoping to finish my thesis in the next six months and of course, there’s the ACCEPT-AI clinical trial that’s going to keep me busy. We’ve just got the green light from ethics and will start recruiting patients soon. I suspect I’ll be diving deeper in the health economics analysis and dealing with any challenges that pop up from the trial. Amid all of this, I am really excited to get back to clinical training early next year and apply some of the techniques I’ve developed. The goal has always been to make a difference in patient care. So yes, lots to look forward to!”