Objective: Among mental health outcome measures that have been developed for routine use, most of the information concerns the Health of the Nation Outcome Scales (HoNOS). This instrument is widely used in Europe, Australia, and New Zealand, but not in Canada. We tested its sensitivity and predictive validity under conditions that would resemble, as closely as possible, routine use. Method: Treating clinicians were asked to assess patients of all ages referred to outpatient mental health facilities of 2 district health authorities in Nova Scotia using either the HoNOS for adults or the HoNOS for Children and Adolescents (HoNOSCA). Data were entered using the existing routine administrative data system. Results: We obtained at least 1 rating on 4620 patients, giving a completion rate of 82%. On follow-up, ratings for the global score and most of the individual items were sensitive to change (n = 808). After adjusting for confounders, a baseline HoNOS score was significantly associated with subsequent in and outpatient service use including admissions, bed days, and psychiatric contacts (n = 1359). Conclusions: HoNOS has satisfactory sensitivity and predictive validity for routine use. We could introduce the adult version and HoNOSCA simultaneously and collect data using routine databases. Given the widespread routine use of HoNOS internationally, using the same outcome measure in Canada would enable comparisons of illness severity and outcomes between jurisdictions.