In the present study, the superior factorial structure of the Health of the Nation Outcome Scales (HoNOS) was investigated. The aims of the study were to test the fit of our new four-factor/subscale structure within and across a wide range of clinical populations compared to competing structures, and to identify any clinical populations where the HoNOS might require modification in order to improve its psychometric properties. Ordinal confirmatory factor analysis was performed for four competing HoNOS subscale structures using a national sample of 80,161 ratings taken at the point of referral to mental health services in England. A new four-factor structure with good/acceptable internal consistency for each of the factors/subscales demonstrated a significantly better fit to the data for 21 of the 24 data groups compared to three competing subscale models. A new four-factor/subscale structure was shown to be the most generalizable subscale structure currently available for HoNOS. However, although superior to other factorial structures, the statistical fit was less than optimal for some clinical populations. Therefore, we would recommend that additional items relevant to all mental health patients are identified and added to the HoNOS in order to allow the development of a single, more generalizable subscale structure, and a much improved version of the HoNOS.