TidalSense - the long road to NHS innovation
Bringing AI to COPD diagnosis.
The other day I met the young CEO of a company with some very impressive medical tech. I came away undecided about whether I should be excited that the UK can generate such innovation in this field - or exasperated at how long it takes for ideas to be put into action in the NHS.
The company was called TidalSense, the CEO was Dr Ameera Patel, and the product was a device that - if all goes well - will be used from next year to diagnose COPD (Chronic Obstructive Pulmonary Disease.) This respiratory condition affects at least 1.6 million people. “I think the actual prevalence is unknown,” says Dr Patel, “because there's so much undiagnosed COPD, but it's possibly up to a 10th of the population..”
On a video call from the fast-growing company’s Cambridge office, she showed off the N-Tidal device:
“In here is a very fast response CO2 sensor. So you breathe through this, and it's measuring pretty well every molecule of carbon dioxide that is coming out of your lungs at a very accurate level.”
Currently COPD is diagnosed using spirometry, a test developed in the 19th century. “It requires forced exhalation, “ she explains, “which is from a clinician’s point of view really difficult to get patients, particularly older patients, to do because it requires coordination of breathing. The appointment takes 30 minutes so it's not only unpleasant for patients but it takes a lot of time.”
By contrast, the N-Tidal device only requires normal breathing and patients can be tested by healthcare staff without specialist training in under five minutes. The other advantage is it makes use of AI and this is where Ameera Patel came on board. As well as training as a doctor she got a PhD in Applied Mathematics - in Germany she would be called Dr Dr - and worked for a while at Microsoft as an AI researcher.
She met TidalSense in 2019, six years after it was founded with the work so far focused on the hardware. “I joined the company because I was interested in the technology from a patient point of view because I have severe asthma.” Within two years had been made CEO.
Where AI will help is in automating, to some extent, the process of recognising in the data produced by the N-Tidal device that a patient is likely to have COPD. She showed me two waveforms generated by the device - one with a top hat shape was for a patient whose lungs were functioning normally, the other more a lopsided pyramid which indicated COPD.
Dr Patel said the algorithm was trained on vast numbers of images like this: “It's learning what the shape of that waveform looks like in someone who's got COPD, because in COPD the structure of the lungs has changed. and so the way in which CO2 moves out of the airways is different.”
For now the system is being trialled on up to 600 patients in Oxford by GPs, practice nurses and pharmacists, but it will not be able to be used as a diagnostic test until it has received regulatory approval. Rather than go to the UK’s MHRA - which other companies have told me moves extremely slowly at the moment due to a restructuring - TidalSense has applied to the EU’s medicines regulator and is hoping for approval in the first half of next year.
That will be great if it happens but I am struck that it will also mean the company’s first revenues are generated 11 years after it was founded. TidalSense was the beneficiary of government grants in its early years, and later raised some seed capital, with the latest £7.5 million funding round in May this year
Getting this kind of health technology off the ground requires huge amounts of determination from founders, an ability to negotiate your way through the thickets of NHS bureaucracy and over the regulatory hurdles - and extraordinary amounts of patience from backers who need to wait many years for a return on their investment which may never come. So if TidalSense does end up transforming the diagnosis of COPD it will be a cause for major celebrations.