From a sample of 239 patients diagnosed paranoid state and hospitalized between 1913 and 1940 at the Phipps Clinic, we particularly studied a group of 60 patients without previous hospitalizations, consisting of 57 patients with follow-ups of 5 or more years, and 3 patients who killed themselves (the ultimate follow-up) less than 1 year after discharge. These 60 patients had been retrospectively diagnosed with delusional disorders by DSM-IV criteria. On follow-up, 27% were rated recovered, whereas 52% were rated unimproved. Long-term follow-up was correlated with discharge status. Poor follow-up was significantly correlated with seclusive personality, poor premorbid history, onset 6 months or more before admission, gradual onset, lack of insight, single marital status, and lack of precipitating events. A prognostic scale constructed from the first four of these variables was predictive of long-term outcome. More recent, better treatment results have been contrasted with these findings from an earlier non-drug-treatment era.