Sir,

In their article, Ellis et al1 explore the economical and philosophical dilemma of screening for neovascular age-related macular degeneration, or choroidal neovascular membranes (CNV). Although these may be important considerations for those implementing such screening programmes, the complexities of such considerations are of little importance to those who have the disease. Their scepticism regarding the cost effectiveness of such a programme centred purely around community optometrists is, however, valid. There are currently little published data on different CNV screening methods and clearly the ability to implement effective screening will depend on the local health care architecture. Current methods of detecting CNV include periodic fundal examination, Amsler chart distortion, the preferential hyperacuity perimeter,2, 3 and reporting of symptoms by the patient. Of these, fundal examination by a medical retina specialist +/− fluorescein angiography is the most accurate, but there are huge time constraints on such highly trained individuals working within the NHS system. With a proven effective treatment already available, and promising new treatments on the horizon, it would seem that efficient, cost effective detection of the disease is appropriate.

We recently carried out a prospective study in the Ophthalmology Department of St James's University hospital, Leeds, West Yorkshire, looking at a novel, fast-track assessment service for the refinement of suspected CNV referrals. The primary aim of the study was to determine whether referrals of suspected CNV from community optometrists could be refined by a nurse and photographer team within the department so as to detect those patients needing urgent intervention by a medical retina specialist. Based on the ophthalmic history and stereoscopic fundus photography, fluorescein angiography was performed if CNV was suspected by the presence of exudation, haemorrhage, and/or elevation of the macula. All referrals and images were subsequently reviewed by a medical retina specialist. The outcome of this review was used as a gold standard, against which the accuracy of the initial referral and of the novel fast-track assessment and refinement service was determined. In the study, 50 consecutive patients referred with suspected CNV by their optometrist, mainly using Amsler chart distortion as the marker of disease, were assessed. Of these 21 patients (42%) had neovascular AMD of whom 19 patients had CNV and two patients had retinal angiomatous proliferations. This represented a very high false-positive referral rate, which clearly could have implications both in terms of health-care economics and the timely treatment for true positive cases. Where fundal abnormalities were seen by the optometrist the specificity rose from 0 to 41%, but the sensitivity fell to 71%. Information regarding the true false-negative rate among community optometrists was not available. In contrast, the novel fast-track assessment and refinement service demonstrated a specificity and sensitivity of 96 and 90%, respectively. This was achieved with the use of presently available resources, incurring no additional costs and meant those with CNV requiring treatment were seen by a medical retina specialist in a mean of 6 days.

The above data highlight the fact that community optometrists appear to be over-reliant on an abnormal Amsler chart to diagnose neovascular ARMD and may lack the knowledge or confidence to rely on fundus examination instead. Some refinement process is necessary to ensure that patients who may benefit from treatment can be seen quickly, without overloading a medical retina service with inappropriate urgent referrals. Such a refinement process could involve specialist optometrists in the community or in hospital. Alternatively, it may involve ophthalmologists in training4 (a recently published article looking at the use of nonstereo digital fundus photographs by ophthalmic interns, for the detection of CNV, found a mean specificity and sensitivity of 85.7 and 78.8%, respectively) or other ophthalmologists. Exactly how this refinement occurs is not important, provided the process has a high sensitivity and specificity and does not induce any additional delay or require additional expenditure given the prevalence of the condition in the UK population.5 Our concern about refinement by community optometrists is the imposition of an additional step in the patient pathway before the patient is seen by a medical retina specialist. This is likely to be a source of additional delay. In the process outlined above, urgent referrals are sent directly to the medical retina specialist and then the refinement is performed using the knowledge and experience of existing staff within the department. The refinement process does not require additional resources as the assessment can usually be scheduled for a time when there is spare capacity. Patients with treatable disease can then be scheduled directly for priority treatment.