From Film to AI: The Success of Diabetic Eye Screening in England

Prof Peter Scanlon, FRCP FRCOphth FHEA DCH DRCOG DO


October 05, 2018

Good afternoon. My name is Professor Peter Scanlon. I'm clinical director for the NHS Diabetic Eye Screening Programme in England. I'm also a consultant eye specialist at Gloucestershire Hospitals and the Oxford University Hospitals.

I’ve been invited to the European Association for the Study of Diabetes to speak on the success of the English NHS Diabetic Eye Screening Programme.

From Film to Digital

I was involved in the early stages of the programme, because in 1996 I was asked in Gloucestershire to suggest how we should take forward eye screaming for people with diabetes in the county of Gloucestershire. At that time, there were screening programmes that were successful in England that were based on 35 millimetre film and Polaroid photography and I was keen that we should use digital photography and the first digital cameras had come onto the market in the English NHS shortly before that time.

And we started our screening programme in Gloucestershire in 1998 and it took 2 years for the first round of screening but it was shown to be successful, and a national screening programme was announced in England in 2003, and I was asked to be clinical director for the English screening programme at that time and I've remained as clinical director since.

We rolled out the programme in England between 2003 and 2008. I must acknowledge that there were numerous other people who were very interested in diabetic eye screening at the time, namely people like Roy Taylor, who's professor of diabetes in Newcastle, Simon Harding in Liverpool, Richard Greenwood in Norwich, and Professor David Owens in Wales.

As we rolled out the programme we have noticed over the years that there's been a gradual reduction in the areas that we are living in the number of registrations that we have given for blindness. And there's also been a reduction in the number of surgeries that have been undertaken for late diabetic disease. And so by detecting the condition early we've managed to reduce the registrations. In my own county of Gloucestershire, the number of registrations in 2003-2008 was about five per year for the seriously sight impaired or blindness, and about nine per year for the sight impaired or partially sighted. And we've reduced that to one a year now for blindness or seriously sight impaired, and four per year for sight impaired or partially sighted.

Dramatic Reduction in Diabetic Eye Disease

So there's been a dramatic reduction. And this was also shown at the national level in 2013, when it was showing that diabetic retinopathy, for the first time in 30 years, was no longer the leading cause of blindness in the working age group.

It had been taken off as the leading cause of blindness, and was now inherited eye disease.

We offer screening at the present time to everybody in England over the age of 12 with diabetes. We offer them a screening appointment where they have dilated photography with the digital cameras, and these images are graded. And then if significant abnormalities are found on the images, then the person is referred to have an examination by an ophthalmologist.

Better Attendance and AI

We have certain improvements that we feel would be beneficial for the programme.

Our current attendance rate is 80% of people invited, which is very good, but we're looking at the non-attenders, and the non-attenders are often the people who are most at risk, and the age group between 19 and 34 are a challenge to get to attend, and we're looking to improve the attendance in that age group.

We're also looking at new technology to see if any of the new types of cameras on the market rather than the digital cameras might be of benefit to us. They are more expensive, and so we also have to look at the cost effectiveness of the new cameras.

We also need to look at automated analysis to see if computerised algorithms can read the images for us. And the technology at this present time seems to be beneficial, and is being used in Scotland to take out images that are normal. So we feel that by using this technology for removing the number of gradings that we need to do on normal images, then we might be able to use it successfully in our programmes.

There are a number of cautions that we have and so we're going through the process of trying to work out how it can best be used.

The main message really for us today at this meeting is to demonstrate that in a large population, we have managed to introduce screening in a way that has actually reduced blindness. So that early detection and treatment of the sight threatening retinopathy that's found, combined with good control of diabetes, can prevent blindness in the population.

Thank you very much.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: