RISK PREDICTION ALGORITHM

More cost-effective screening programmes are urgently needed to deal with the expected increase in diabetes worldwide over the coming decades, Einar Stefánsson MD, PhD told delegates attending the 14th EURETINA Congress in London.

“We need smarter screening in order to reduce costs and to allocate health care resources according to patient risk. If we are going to deal with the coming global epidemic of diabetes and screen 400 to 500 million people and treat them prophylactically in a timely manner, we have to develop better and more cost-effective methods to do so,” said Dr Stefánsson.

With this in mind, Dr Stefánsson and co-workers have developed a risk prediction algorithm for sight-threatening diabetic retinopathy that has already proven effective in validation trials in Denmark, the Netherlands and the UK.

“We currently screen high- and low-risk patients as if they were identical, which is clearly not the case. So we decided to introduce more risk factors in the algorithm to allow for individualised risk assessment and move away from one-size-fits-all screening towards a more personalised approach to medicine. The multifactorial risk assessment that we have developed takes into account six risk factors: duration of diabetes, state of retinopathy, blood glucose levels (HbA1c), blood pressure, type of diabetes and the gender of the patient,” he said. The program is available on www.retinarisk.com.

The introduction of national diabetic retinopathy screening programmes from the 1970s onwards has dramatically reduced the risk of blindness stemming from macular oedema and/or proliferative diabetic retinopathy, said Dr Stefánsson.

“Combining laser photocoagulation and diabetic eye screening turned out to be a marriage made in heaven because the screening would catch the diabetic patients at a time where the laser treatment was optimal at the very beginning of sight-threatening characteristics, be they proliferative or oedematous,” he said. This still holds true for newer treatments in use today such as intravitreal drugs and vitreous surgery.

Timing of treatment

While screening optimises the timing of treatment and has been enormously beneficial, it is only a very small percentage of patients who actually need treatment for sight-threatening retinopathy, said Dr Stefánsson.

“If you screen all diabetic patients once a year, only three out of 100 will actually need treatment for sight-threatening retinopathy. So the other 97 that come in for annual screening are simply told to come back again the following year. Screening is enormously useful and is a great public health tool but it could also be made much more cost-effective,” he said.

More discriminating screening methods have already been shown to be safe and effective, said Dr Stefánsson, citing a study by his group published in 2007 which showed that every-other year screening for diabetics without retinopathy reduced the number of screening visits by more than 25 per cent without any loss in safety.

The new algorithm, which draws on data from several landmark epidemiological studies in diabetes, will lead to further improvements, said Dr Stefánsson.

“Using these risk factors in an algorithm we can predict 80 per cent of the risk of developing sight-threatening retinopathy. We can use the individual risk measurement to control screening intervals and frequency, with more screening in high-risk patients and less frequent screening in low-risk patients,” he said.

The algorithm has already been validated and tested in diabetic cohorts in Denmark, Netherlands and the UK, predicting 76 per cent to 80 per cent of the risk of developing sight-threatening diabetic retinopathy.

“Using this algorithm means about 60 per cent reduced screening frequency than using an annual screening programme. This cuts the costs by more than half but without compromising safety,” he said.

Dr Stefánsson said that promoting diabetic eye screening has received the official backing of EURETINA.

“While the details are still being worked out, I think the initiative should include pan-European advocacy to use EURETINA’s influence to spread the word and induce ophthalmologists to get involved in introducing this screening method in their respective countries. The organisation is also interested in supporting model systems for screening in eastern Europe, with the first one scheduled in Szeged, Hungary,” he said.

Einar Stefánsson: einarste@landspitali.is