purpose: non-penetrating deep sclerectomy (npds) has emerged as a viable option in the surgical management of open-angle glaucoma. our aim is to assess the cost-effectiveness of npds and to compare it to maximum medical treatment in a 5-year follow-up. methods: a decision analysis model was built. surgical (npds) arm of the decision tree was observational (consecutive retrospective case series) and maximum medical treatment arm was hypothetical. maximum medical therapy was considered a three-drug regimen (association of a fixed combination of timolol/dorzolamide [fctd] and a prostaglandin analogue [bimatoprost, latanoprost or travoprost]). cost-effectiveness ratio was defined as direct cost (us dollars) for each percentage of intraocular pressure (iop) reduction. horizon was 5 years and perspective is from the public health care service in brazil (sus). incremental cost-effectiveness ratio (icer) was calculated. results: direct cost for each percentage of iop reduction in 5 years (cost-effectiveness ratio) was us$ 10.19 for npds; us$ 37.45 for the association of a fctd and bimatoprost; us$ 39.33 for fctd and travoprost; and us$ 41.42 for fctd and latanoprost. npds demonstrated a better cost-effectiveness ratio, compared to maximum medical therapy. the icer was negative for all medical treatment options; therefore npds was dominant. conclusions: despite some limitations, npds was both less costly and more effective than maximum medical therapy. from the brazilian public health perspective, it was the most cost-effective treatment option when compared to maximum medical therapy (fctd and prostaglandin).