OBJECTIVE: To estimate the prevalence and causes of blindness in Chiapas, Mexico, and to assess the feasibility of using the Rapid Assessment of Avoidable Blindness framework to estimate diabetic retinopathy (DR) prevalence. DESIGN: A cross-sectional population-based survey. PARTICIPANTS: Sixty-six clusters of 50 people 50 years of age or older were selected by probability proportionate to size sampling. Households within clusters were selected through compact segment sampling. METHODS: Participants underwent visual acuity (VA) screening and diagnosis of cause of visual impairment by an ophthalmologist. Participants were classed as having diabetes if they had a previous diagnosis of diabetes, were receiving treatment for glucose control, or had a random blood glucose level of more than 200 mg/dl. Participants with diabetes were assessed for DR using dilated clinical examination (direct and indirect ophthalmoscope) and 1 dilated digital fundus photograph per eye (graded by an ophthalmologist during the survey and regraded by a retinal specialist-"reference standard") following the Scottish DR grading protocol. MAIN OUTCOME MEASURES: Prevalence of blindness (VA <20/400 in the best eye with available correction) and DR. RESULTS: Three thousand three hundred subjects were selected, of whom 2864 (87%) were examined. The estimated prevalence of bilateral blindness was 2.3% (95% confidence interval [CI], 1.7%-2.9%). Cataract was the leading cause of bilateral blindness (63%), followed by posterior segment diseases (24%), which included DR (8% of blindness). The prevalence of diabetes was 21% (19.5%-23.1%). Among participants with diabetes, the prevalence of DR (in at least 1 eye) was 38.9% (95% CI, 33.7%-44.1%). The prevalence of sight-threatening DR (STDR; defined as proliferative DR, referable maculopathy, or both) was 21.0% (95% CI, 16.7%-25.3%). Agreement with the reference standard was good for any retinopathy and STDR for the clinical examination (κ = 0.80 and 0.79, respectively) and the photograph graded during the survey (κ = 0.80 and 0.82, respectively). CONCLUSIONS: The prevalence of diabetes and DR in Chiapas was high. Including the DR component was possible, but added considerably to the cost and complexity of the survey, and so would be warranted only if a high prevalence of diabetes is expected and if resources and time permit.