This study tested predictors of 2-year antipsychotic-free response from the Soteria study (older, better social functioning, fewer cardinal symptoms) using data from the Finnish Acute Psychosis Integrated treatment study. The quasi-experimental study compared need-adapted family-oriented psychosocial intervention within a 3-week antipsychotic-free trial to psychosocial intervention plus antipsychotic medications. Forty-six percent of experimental completers (37% of intent-to-treat subjects) were successfully treated without antipsychotic medications for the entire 2-year study. The DSM-III-R diagnoses of schizophrenia and schizophreniform disorder and Soteria-suggested predictors were not related to antipsychotic-free response. Different variables within the same domains of good prognosis and fewer schizophrenia symptoms predicted antipsychotic-free response or nonresponse with 74% accuracy. The 6-month duration of symptom criterion distinguishing schizophrenia from schizophreniform disorder does not separate medication-free treatment responders from those requiring medications. Prognosis appears related to antipsychotic-free response and may be helpful in distinguishing schizophrenia from schizophreniform disorder in early episodes.