Some studies in first-episode schizophrenia correlate shorter duration of untreated psychosis (DUP) with better prognosis, suggesting that timing of treatment may be important. A three-site prospective clinical trial in Norway and Denmark is underway to investigate the effect of the timing of treatment in first-episode psychosis. One health care sector (Rogaland, Norway) is experimental and has developed an early detection (ED) system to reduce DUP. Two other sectors (Ullevål, Norway, and Roskilde, Denmark) are comparison sectors and rely on existing detection and referral systems for first-episode cases. The study ultimately will compare early detected with usual detected patients. This paper describes the study's major independent intervention variable, i.e. a comprehensive education and detection system to change DUP in first onset psychosis. System variables and first results from the four-year inclusion period (1997-2000) are described. It includes targeted information towards the general public, health professionals and schools, and ED teams to recruit appropriate patients into treatment as soon as possible. This plus easy access to psychiatric services via ED teams systematically changed referral patterns of first-episode schizophrenia. DUP was reduced by 1.5 years (mean) from before the time the ED system was instituted (to 0.5 years). The ED strategies appear to be effective and to influence directly the community's help-seeking behaviour.