Background The evidence for manualized psycho-educative family intervention (FI) in first-episode psychosis (FEP) is well-established to reduce relapse and caregiver distress. Less is known, however, about type and duration of FI. Aim To compare two different types of manualized family interventions for FEP: Multi-Family Groups (MFG) and Single-Family Intervention (SFI). Methods This was a prospective, quasi-experimental cohort study of all participants of an early psychosis service (OPUS) with an ICD-10 diagnosis of F20 to F29 (excl. F21), aged 18 to 35 years, in Psychiatry Region Zealand, Denmark, during a 2-year period. All service users and their relatives are offered FI, either MFG or SFI. Assessment of level of participation, psychopathology measured by The Positive and Negative Syndrome Scale (PANSS), remission status and relapses was carried out at 3-year follow up. Results We found no differences between the service users participating in SFI (N = 25) or MFG (N = 18) on number of readmissions or relapses after baseline or psychopathology. A binary logistic regression analysis on remission status at follow up showed a trend in favour of MFG. A surprisingly high proportion of the families did not receive an FI. Conclusions SFI and MFG seem equally effective in an FEP programme. The low attendance of FI may be due to several issues-among others, the probability that the FI did not sufficiently match the needs of the service users. Further studies involving larger samples are needed, included randomized controlled trials and implementation studies.