Foresight - sharing data to battle blindness
A big opportunity for the UK but will we grasp it?
The one resource that the UK does not lack as we strive to improve healthcare is data. But sharing the great oceans of information generated by the NHS in the form of medical records for the purposes of research has often proved tricky, with constant rows about privacy. If that is hard, imagine trying to persuade the private companies that now carry out millions of high quality eye scans to hand them over to medical researchers.
But that is what a charity called Action Against AMD is attempting as it battles to find better ways of treating age-related macular degeneration. “Presently in the UK, there's about 700,000 people with AMD, late stage, and only a fraction of them can be treated,” explains Dr Wen Hwa Lee, the leading Oxford academic who is now running the charity. That means many of them will end up blind.
Great strides have been made in diagnosing AMD, partly because high street chains of opticians such as Specsavers now have the high quality OCT scanners needed to give a full picture of what is going on inside the eye. Coupled with this, there have been big steps forward in using AI trained on these scans to provide early diagnosis not just of AMD but of other conditions including Parkinson’s.
From this combination of OCT scans and AI, a whole new field called oculomics has emerged built on the theory that the eyes are not just the windows of the soul but contain clues as to what is going wrong elsewhere in the body. Last year Pearse Keane, the Moorfields Hospital ophthalmologist and artificial intelligence expert, demonstrated what was possible when he co-authored a paper showing that eye scans could detect signs of Parkinson’s up to seven years before diagnosis.
But one thing is holding back this new discipline, a lack of data from people who are - or appear to be - perfectly healthy. For instance, Action Against AMD believes that if it can develop biomarkers that predict that a person will develop macular degeneration 10 to 15 years before it happens, then it could be possible to create drugs that would slow or stop the condition.
To close this data gap the charity has launched a programme called Foresight which aims to recruit 500,000 members of the public and build a diverse database of retinal images. The idea would be to link this - with the volunteers’ agreement - to other databanks with genetic and health information, such as the UK’s Our Future Health project. Then it would be made available to researchers around the world under carefully controlled and secure conditions.
The idea is not to access the millions of scans collected over the past five years or so - Dr Wen Hwa Lee explains that getting retrospective consent from customers of every high street optometrist to use their scans would be very hard, and then there would be other problems: “Because the equipment is operated differently, it would take much longer, way more money, just to try to clean the data for AI, and by cleaning you introduce presumptions and therefore bias.”
Instead, Foresight is now offering optometrists a contract for future scans: “They help us collect the data. We will pay them for that service, but they have to do it within the standard operating procedure that we prescribe them, and all the participants will be consenting for their data to be used for research.”
It sounds like an initiative any person with vision problems, mild or severe, would be very happy to support. But I put it to Dr Lee that how the project is marketed to the public will be vital. Ask people whether they would like to contribute to research that could save millions from going blind and they are very likely to say yes. Ask them whether they would like to help big pharmaceutical companies to discover new drugs to combat blindness and many people will be far less keen.
“Big pharma” is not a term of affection for many despite the fact that most major advances in healthcare would not have happened without the huge long-term bets made by the pharmaceuticals industry and its investors.
But Foresight is well aware of the need to win public trust. It will have a panel of experts reviewing which industries can get access to the data - to get approval from the Health Research Authority, the projects’s leaderschave had to agree to bar insurers developing algorithms to exclude from coverage people who may be likely to develop certain conditions in the future. The plan also involves the charity charging for access to its data and earning royalties from any products derived from it, all to fund its long-term mission to end AMD.
According to Dr Lee, the UK is some five to ten years ahead of the USA, where a similar project is underway but because it is being run by the federal government rather than a charity, things are moving far more slowly. “We have a window of opportunity here,’ he says, talking of switching the focus of healthcare to prevention but also of “creating a healthy ecosystem of innovation.”
But that all depends, he emphasises, on building trust. It will also need thousands of optometrists across the UK to act as evangelists for Foresight, telling their customers why it might be a good idea to hand over their scans.. But it may also need a change of heart from all of us, a willingness to see the positive benefits of sharing our data, rather than worrying so much about the dangers.
So, in a spirit of transparency, I’ve posted one of my OCT scans at the top of this post. Nearly 20 years ago my optometrist spotted something worrying in a scan and sent me to Moorfields Hospital where I was diagnosed with a malignant melanoma, a cancerous tumour behind my left eye. If my local family run optician had not invested in the latest machines, and if I had not been referred quickly to the eye hospital I might not be writing this today. So I will be cheering on any effort which unites private eyecare businesses, NHS hospitals and academic researchers in looking for better ways to mine our data riches.