[Validation of the electronic mailing of retinographs of diabetic patients in order to detect retinopathy in primary care].

Objective: Evaluate concordance in biomicroscopy evaluation of diabetic retinopathy degree among ophthalmologists. Validation of e-mail transmission of digital fundus photographs of type 2 diabetes patients as a method of diabetic retinopathy detection.

Design: Descriptive study.

Setting: Urban, primary health centre, and hospital.

Participants: Type 2 diabetic patients selected of consecutive form when going to the primary health center (n=352).

Main Measurements: Parallel observer-blind evaluation of degree of retinopathy through biomicroscopy performed by ophthalmologists, against digital photographic images sent by e-mail taken by the family doctors. Concordance in the biomicroscopy among ophthalmologists was previously tested.

Results: Retinopathy was revealed in 25.70% of the retinographs; 10.44% was mild, 12.05% moderate, and 3.21% severe. Weighted kappa was 0.876 for biomicroscopy concordance. Sensitivity in detecting retinopathy is 76.6% and specificity 95.2%; 92.7% and 99.5% for ophthalmologist-derivable retinopathy. Sensitivity was 66.7% for non-mydriatic retinograph without dilation, 76.9% with elective dilation, and 85% with the mydriatic.

Conclusions: Concordance in evaluation of retinopathy degree through biomicroscopy was "very good." This allows using a single ophthalmologist's exploration as a reference model. E-mail transmission of the photograph of the back of the eye in type 2 diabetic patients as a retinopathy detection method is feasible. Regardless of the type of retinograph used, the photographs should be taken on the dilated eye, as this significantly improves sensitivity.