Single-field fundus photography for diabetic retinopathy screening: a report by the American Academy of Ophthalmology

Abstract

Objective: To evaluate whether single-field fundus photography can be used as a screening tool to identify diabetic retinopathy for referral for further ophthalmic care.

Methods: A MEDLINE search of the peer-reviewed literature was conducted in June 2001 for the years 1968 to 2001 and updated in September 2003, yielding 145 articles. The search was limited to articles published in English. The Cochrane Library of clinical trials was also investigated. The authors reviewed the abstracts of these articles and selected 63 of possible clinical relevance for review by the panel. Of these 63 articles, the panel selected 32 for the panel methodologist to review and rate according to the strength of evidence.

Results: Three of the 32 articles reviewed were classified as level I evidence, and 4 were classified as level II evidence. Evidence from level I studies demonstrates that as a tool to detect vision-threatening retinopathy, single-field fundus photography interpreted by trained readers has sensitivity ranging from 61% to 90% and specificity ranging from 85% to 97% when compared with the gold standard reference of stereophotographs of 7 standard fields. When compared with dilated ophthalmoscopy by an ophthalmologist, single-field fundus photography has sensitivity ranging from 38% to 100% and specificity ranging from 75% to 100%.

Conclusions: Single-field fundus photography is not a substitute for a comprehensive ophthalmic examination, but there is level I evidence that it can serve as a screening tool for diabetic retinopathy to identify patients with retinopathy for referral for ophthalmic evaluation and management. The advantages of single-field fundus photography interpreted by trained readers are ease of use (only one photograph is required), convenience, and ability to detect retinopathy. Further studies will be required to assess the implementation of single-field photography-based programs to confirm the clinical and cost-effectiveness of these techniques in improving population visual outcomes. Future research also should include establishing standardized protocols and satisfactory performance standards for diabetic retinopathy screening programs.