Benefits of training junior physicians to detect diabetic retinopathy--the Glasgow experience.

The accuracy and appropriateness of 115 consecutive referrals by non-consultant physicians to a specialist Diabetic Retinopathy Clinic were assessed in a retrospective study. The source of the referrals was masked throughout the study. Referrals were classed as 'appropriate' or 'inappropriate' for patient management, and the referral diagnosis (where specified) was compared with the ophthalmologist's initial assessment. It was graded as 'correct', 'partly correct' and 'incorrect'. Referrals from physicians who had received 40-50 hours of outpatient training in the Diabetic Retinopathy Clinic (group A, n = 49) were compared with referrals from doctors without this special instruction (group B, n = 66). Referral was deemed 'appropriate' in 32 (65%) of group A referrals, but in only 22 (33%) of group B (chi 2 = 11.54, df = 1, P less than 0.001). Referral diagnosis (when expressed) was graded as 'correct' in 28 (67%) of group A referrals compared with only 12 (30%) of group B, being 'incorrect' in 10 (25%) of group B referrals and just two (4.5%) of group A (chi 2 = 12.9, df = 2, P less than 0.005). Regular fundoscopy with accurate assessment and appropriate action is vital to prevent loss of vision in diabetic patients. Short-term outpatient training in a Diabetic Eye Clinic leads junior physicians to more appropriate referral and more accurate referral diagnosis.