Demographic differences in the prevalence of blindness may be partly due to undertreatment of susceptible population subgroups. The authors examined the relation of age, race, and other demographic characteristics with initiation of treatment for glaucoma and compared treatment rates with expected rates based on known disease prevalence. Data were from Medicaid enrollees aged 65-99 years in New Jersey between March 1981 and February 1990. Based on review of all claims for prescription medications and laser and incisional surgery, there were 6,173 cases with at least 6 months of documented system eligibility before their initial treatment for glaucoma. The overall rate of new treatment was 11.5 cases per 1,000 person-years, and increased throughout the 1980s. The age-adjusted relative rate of new treatment was 1.58 times higher in blacks compared with whites; however, this was less than half the relative rate expected based on estimated relative incidence rates. Similarly, enrollees aged 70-99 years had only 7% to 27% higher treatment rates than those aged 65-69 years, substantially less than expected. Treatment for glaucoma was also less likely to be initiated in nursing home residents, compared with those living in the community. Blacks and the very old are much less likely to have treatment for glaucoma initiated than would be predicted based on the magnitude of disease burden in these populations.