Photography combined with ophthalmoscopy in retinal screening

Quality assurance in screening for sight-threatening diabetic retinopathy

AIMS

There is a need for continuous evaluation of screening services for diabetic retinopathy against agreed performance standards. We describe a quality assurance programme implemented in Newcastle in January 1999 and report on outcomes at 18 months.

METHODS

Annual retinal screening is performed using combined retinal photography and direct ophthalmoscopy in two streams. Diabetologists perform screening in the Hospital Screening Programme, which serves patients whose diabetes is managed in specialist clinics, and trained retinal screeners perform screening in the District Screening Programme, which serves patients whose diabetes is managed in the community. Reference standard examination of dilated fundoscopy with a slit-lamp and condensing lens was performed by an ophthalmologist at periodic sessions on consecutive patients attending for screening.

RESULTS

Six hundred and nine (6.4%) of 9468 patients screened underwent reference standard examination. The sensitivity and specificity of detection of sight-threatening diabetic retinopathy (STDR) was 82.5% and 98%, respectively, for the Hospital Screening Programme; 85.7% and 95.7%, respectively, for the District Screening Programme; and 83.3% and 96.8% for both services combined. One hundred and ten (18.1%) of 609 patients audited were referred to ophthalmology as a result of screening, and this led to 16 patients (2.6%) receiving laser photocoagulation for STDR. Reference standard examination identified a further four patients (0.7%) who required laser photocoagulation.

CONCLUSIONS

Preliminary data indicate that satisfactory performance standards are being achieved. The National Service Framework for Diabetes requires that all units institute quality assurance for retinal screening, and we report the practical implementation of this in one district.

Non-mydriatic fundus photography: a viable alternative to fundoscopy for identification of diabetic retinopathy in an Aboriginal population in rural Western Australia?

PURPOSE

To evaluate the Canon CR5-45NM non-mydriatic fundus camera (Canon, Kanagawa, Japan) for identifying retinopathy and the need for laser treatment in a population of Aboriginal patients with diabetes mellitus in rural Western Australia.

METHODS

Diabetic Aboriginal patients were photographed through undilated pupils using a Canon CR5-45NM non-mydriatic fundus camera, after which ophthalmoscopy was performed using indirect ophthalmoscopy through dilated pupils. The examining ophthalmologist recorded the presence of retinopathy and the need for laser treatment. A proportion of patients were rephotographed through dilated pupils. Photographs were reviewed by a second ophthalmologist who evaluated the quality of the image, the presence of retinopathy and the need for laser treatment. Results of fundus photographs and ophthalmoscopy were compared.

RESULTS

Three hundred and twenty-eight eyes in 164 Aboriginal patients were examined. The mean patient age was 48.2 years (range 16-81 years) and the mean duration of diabetes was 7.5 years (range 1-35 years). Seventy-four eyes (22.6%) were diagnosed with retinopathy using combined examination techniques, 44 (59.5%) of which were identified by ophthalmoscopy and 55 (74.3%) by photography. Thirty-five eyes were deemed to need treatment, 18 (51.4%) of which were identified by ophthalmoscopy and 30 (85.7%) by photography. Kappa coefficient measurement for agreement for presence of retinopathy and need for referral was 0.41 and 0.53, respectively. Photograph quality was significantly improved following pupil dilation.

CONCLUSIONS

The Canon CR5-45NM non-mydriatic fundus camera was relatively good at identifying diabetic retinopathy and could usefully be applied within a screening programme for treatable disease within this population.