Knowing the long-term outcomes of schizophrenia and stability of a schizophrenia diagnosis are important from a clinical standpoint as well as essential to future research on diagnostic classifications and outcome. As in prior research on schizophrenia, prospectively designed long-term studies over the past 30 years find that the predominant course of illness includes chronically poor functioning, with little evidence of long-term improvement. Mortality due to suicide is significant at about 10% over 10-year periods of follow-up. Within studies, outcome domains are interrelated, and the relatively consistent predictors of poorer outcome include family history of schizophrenia, insidious onset, poor premorbid functioning, severity of negative symptoms, and severity and duration of untreated psychosis. Residing in a developed rather than a developing country is also associated with a poorer long-term course. The diagnostic stability of schizophrenia is less well studied. The positive predictive value exceeds 90%, and preliminary findings from the 10-year follow-up of the Suffolk County Mental Health Project cohort have found that the agreement across time increased from k = .52 (baseline to 10 years) to k = .76 (6 or 24 months to 10 years). After discussing several limitations of the existing body of research, we suggest that future studies incorporate more "modifiable" risk factors into the assessment battery that could potentially be used as building blocks in experimental intervention designs.