Interval Eye Screen

With 2013 now upon us, and the Mayas looking sheepish, perhaps the biggest threat to the future of the NHS Diabetic Eye Screening Programme is our failure to restructure our screening intervals. Back at the dawn of civilisation (circa 2003), it was decided that diabetic retinopathy screening in the UK should adopt a 'one size fits all' approach, which is unfortunate, as it's a phrase which strikes fear into the hearts of a lot of type 2 diabetics with body image issues. As a result, patients found to have sight-threatening DR are referred to the hospital eye service, while everyone else is screened annually, regardless of potential risk.

As patient numbers increase, and budgets are squeezed, the need to look again at this policy has become ever more important. On the one hand we have those diet-controlled octogenarians who claim they were only diagnosed with diabetes because they ate a Chocolate Hob-nob ten minutes before having a blood test - and then prove it by coming back year after year with no retinopathy - and on the other, we have the young type 1s with bad control and a bit of background, who can go downhill faster than Rik Waller on a log flume. Yet we continue to screen them all annually as if there were no distinction.

The less money we have, the more crucial it is that those scarce resources are targeted appropriately, and it's long been suggested that to screen patients with no retinopathy on an annual basis is a waste of our time and money. Conversely, we need to ensure that those patients who have the potential to progress on to sight-threatening DR are not left too long between screenings.

The solution, it would seem, is to identify those patients at low risk, screen them less frequently, and channel those extra resources into more frequent screening for those at higher risk. It's a simple and sensible solution, makes sound financial sense, and seems a lot better for all concerned. Assuming you know who they are.

The problem is identifying which patients fall into which group. And that was the aim of a recent study carried out by Irene M Stratton (the M is for Maculopathy), senior statistician of the Gloucestershire Diabetic Retinopathy Research Group. I've a lot of respect for Irene. She's a vastly experienced number-cruncher with some head-turning vital statistics, and she's produced more funnel plots than a conspiracy theorist. I also like the way she seems to be turning the word 'Stratification' into an eponym.

The important thing is that Irene has a very attractive figure. And that figure is 14,554. It's the number of patient outcomes she's studied in order to develop a simple algorithm to assess the risk of a patient screened annually for diabetic retinopathy developing sight-threatening DR within that twelve month period. You can read more about Irene's stratifications by clicking here, but it's the conclusions which are interesting, and more than a little surprising.

The aim was to estimate a patient's future risk by looking at the results of two consecutive eye screenings. And here's what she found:

Of 7,246 with no DR at either screening, 120 progressed to sight-threatening diabetic retinopathy (STDR), equivalent to an annual rate of 0.7%.

Of 1,778 with no DR in either eye at first screening and in one eye at second screening, 80 progressed to STDR, equivalent to an annual rate of 1.9%

Of 1,159 with background DR in both eyes at both screenings, 299 progressed to STDR, equivalent to an annual rate of 11%.

A lot of attention has been focused on that final statistic, but I find the first one just as shocking. I'd like to know what went on in the lives of those 120 people who had R0 for two years running, and then went straight to M1 (or possibly R2?). And more to the point, who was responsible for their care? Or lack thereof.

The implication is that patients who have no retinopathy for two consecutive screenings can probably be left for a couple of years, although I wouldn't like to be the one who has to explain that policy to an anxious individual who reads the figures above, and points out that one such patient in every 150 will go blind before their next visit.

That aside, it would appear that anyone found to have background retinopathy in both eyes on two consecutive screenings has a better than 1 in 10 chance of developing STDR within twelve months, and therefore needs to be screened more often. But is that the full picture?

My suggestion is no. I think this is:

That eye has received an official, undisputed, and quality assured grade of R1M0, meaning it has background retinopathy with no maculopathy.

And so has this one:

That's not a doctored photo, it's a genuine case of R1. There's a faint microaneurysm less than two disc diameters from the fovea, on the temporal side.

Does it seem likely that the first patient could progress on to sight-threatening DR within the next twelve months? Absolutely. But what about the second?

The problem is that just as the current system makes no distinction between a low risk type 2 with no retinopathy, and a high risk type 1 with background, a system based purely on consecutive grades of R1 would make no distinction between the two patients above. If the second patient came back a year later with the same microaneurysm, they'd be classed as high risk, and called back in six months. Which, on an instinctive level at least, doesn't seem like the best use of our resources.

I'm all for simplicity, but perhaps screening grades alone are unreliable. Perhaps we need to include factors such as whether the patient is type 1 or type 2, whether they're diet controlled or on medication, and how long they've been diabetic. Maybe we even include their latest HbA1c and take their blood pressure.

Or maybe we just redefine R1. A grade which covers both the patients above, and everyone in between, seems far too wide to be meaningful. Especially if we're considering using it to judge future screening intervals. But in these times of austerity and increasing workloads, there's no doubt the system needs reform. Perhaps desperate times call for DESP E-Rate measures.