There are suggestions that some first-episode psychosis (FEP) patients can have favourable outcome without antipsychotic medication. However, there is very limited data regarding patients' characteristics on which the decision to propose medication free treatment could be based. FEPOS is a fi le-based study of an epidemiological sample of 704 FEP patients treated at EPPIC, Melbourne, between 1998 and 2000. Among the 661 patients where data was available, 108 consistently refused medication during the entire duration of their treatment at EPPIC. In this paper we compared, within this sub-group, patients who had a favourable outcome with those who did not. Patients were aged between 15 and 29 years (M = 21.9, SD = 3.40) and the majority were male (70.4%, n = 76). Symptomatic remission data was available on 105 patients; of these patients 41.0% (n = 41) had achieved remission. Functional remission data was available on 100 patients; of these patients 33.0% (n = 33) had achieved functional remission. Combined remission was evident in 23.0% (n = 23) of patients. Three factors were associated with symptomatic remission: better premorbid functioning (based on GAF, OR = 1.07, p = 0.006), higher number of years of education (OR = 1.43, p = 0.020), and being employed or studying at service entry (OR = 2.59, p = 0.034). Three factors were associated with functional remission: shorter duration of prodrome (OR = 0.50, p = 0.043), severity of psychopathology (CGI-S, OR = 0.51, p = 0.024), and vocational status at service entry (OR = 4.29, p = 0.003). While various aspects of pre-morbid functioning seem to correlate with the possibility of a favourable outcome in FEP patients who refuse medication, various limitations need to be taken into account in this study.