BACKGROUND AND OBJECTIVE: to report the 36-month experience of the Stanford University Network for Diagnosis of Retinopathy of Prematurity (SUNDROP) telemedicine initiative. PATIENTS AND METHODS: retrospective analysis of the SUNDROP archival data between December 1, 2005, and November 30, 2008, to evaluate this diagnostic technology for retinopathy of prematurity (ROP) screening. A total of 230 consecutively enrolled infants meeting ROP examination criteria were screened with the Ret-Cam II (Clarity Medical Systems, Pleasanton, CA) and evaluated by the SUNDROP reading center at Stanford University. Outcomes included referral-warranted ROP, treatment-warranted ROP, and anatomic outcomes. RESULTS: in the initial 36-month period, the SUNDROP telemedicine initiative did not miss any treatment-warranted ROP. A total of 230 infants (460 eyes) were imaged, resulting in 1,059 examinations and 10,921 unique images. Ten infants were identified with referral-warranted ROP: nine underwent laser photocoagulation and one regressed spontaneously. The sensitivity was 100% with a specificity of 99.5%. No patient progressed to retinal detachment or other adverse outcomes. CONCLUSION: the SUNDROP telemedicine screening initiative for ROP has demonstrated high reliability for identification of treatment-warranted disease. All cases of treatment-warranted disease were captured. There were no adverse outcomes.