Abstract Objectives. There is a need for an observation scale for assessment and monitoring of acutely psychotic patients. Milestones based on such ratings should be defined, similar to the PANSS-based Remission criteria. FAST-O is such an instrument (11 items and a CGI rating). Methods. Reliability, validity and factor structure were analyzed in four separate studies-most importantly in one study of 33 chronic forensic in-patients rated concurrently by three independent raters, and another study of 91 Psychiatric Intensive Care Units patients and 20 newly admitted forensic patients, rated twice, each time by two independent raters. Results. The factor structure was simple, two factors and an orphan item (Depression). Reliability was adequate on item (>0.75) as well as scale (>0.85) level. There was no bias related to the rater's professional background. The instrument was sensitive to change. Percentile-based algorithms allow characterization of patients and groups. Tentative treatment milestones are defined; a clinical state "half-way" between the acute state and remission. Conclusions. FAST-O is a reliable, valid and easy to implement observation scale for patients with a psychotic illness, which can be used without bias by all staff.