IMPORTANCE The OPUS 20-year follow-up is the longest follow-up of a randomized clinical trial testing early intervention services (EIS) among individuals with first-episode schizophrenia spectrum disorder. OBJECTIVE To report on long-term associations of EIS compared with treatment as usual (TAU) for first-episode schizophrenia spectrum disorder.DESIGN, SETTING, AND PARTICIPANTS A total of 547 individuals were included in this Danish multicenter randomized clinical trial between January 1998 and December 2000 and allocated to early intervention program group (OPUS) or TAU. Raters who were blinded to the original treatment performed the 20-year follow-up. A population-based sample aged 18 to 45 years with first-episode schizophrenia spectrum disorder were included. Individuals were excluded if they were treated with antipsychotics (> 12 weeks prior to randomization), had substance-induced psychosis, had mental disability, or had organic mental disorders. Analysis took place between December 2021 and August 2022.INTERVENTIONS EIS (OPUS) consisted of 2 years of assertive community treatment including social skill training, psychoeducation, and family involvement by a multidisciplinary team. TAU consisted of the available community mental health treatment.MAIN OUTCOMES AND MEASURES Psychopathological and functional outcomes, mortality, days of psychiatric hospitalizations, number of psychiatric outpatient contacts, use of supported housing/homeless shelters, symptom remission, and clinical recovery.RESULTS Of 547 participants, 164 (30%) were interviewed at 20-year follow-up (mean [SD] age, 45.9 [5.6] years; 85 [51.8%] female). No significant differences were found between the OPUS group compared with the TAU group on global functional levels (estimated mean difference, -3.72 [95% CI, -7.67 to 0.22]; P = .06), psychotic symptom dimensions (estimated mean difference, 0.14 [95% CI, -0.25 to 0.52]; P = .48), and negative symptom dimensions (estimated mean difference, 0.13 [95% CI, -0.18 to 0.44]; P = .41). The mortality rate was 13.1% (n = 36) in the OPUS group and 15.1% (n = 41) in the TAU group. Likewise, no differences were found 10 to 20 years after randomization between the OPUS and TAU groups on days of psychiatric hospitalizations (incidence rate ratio, 1.20 [95% CI, 0.73-1.20]; P = .46) or number of outpatient contacts (incidence rate ratio, 1.20 [95% CI, 0.89-1.61]; P = .24). Of the entire sample, 53 participants (40%) were in symptom remission and 23 (18%) were in clinical recovery.CONCLUSIONS AND RELEVANCE In this follow-up study of a randomized clinical trial, no differences between 2 years of EIS vs TAU among individuals with diagnosed schizophrenia spectrum disorders at 20 years were found. New initiatives are needed to maintain the positive outcomes achieved after 2 years of EIS and furthermore improve very long-term outcomes. While registry data was without attrition, interpretation of clinical assessments are limited by high attrition rate. However, this attrition bias most likely confirms the lack of an observed long-term association of OPUS with outcomes.