Few studies have attempted to examine the meaning of health congruence, particularly in the oldest old. Participants were drawn from a longitudinal study of the oldest old (N = 151; M = 90 years). Dichotomized objective health was cross classified with dichotomized subjective health, producing four health congruence groups: good health realists, poor health realists, optimists, and pessimists. Both good health realists and pessimists had good objective health, yet pessimists had the highest depression, lowest functional status, and frequent reports of hospitalization. By contrast, the poor health realists and optimist groups had poor objective health, but the optimists had better outcomes on depression. This suggests that discrepancies between objective and subjective health may have significant implications for health outcomes.