Secure mental health service outcomes are commonly derived from post-discharge recidivism, readmission and mortality rates. Information about less rare behavioural, functional, and symptom-related outcomes across the sometimes lengthy span of admission is scant. We aimed to determine whether patients underwent reliable and clinically significant change in risk-related need, behaviour, functioning, and symptoms during admission from analysis of routinely collected HoNOS-Secure data. Outcomes (N=418) were examined to determine between-group differences in baseline scores and rates of change, the proportion of patients for whom change was reliable, and, of those, the proportion whose scores fell during treatment by a clinically significant margin. Reliable change was demonstrated for just 4.8% and 5.7% of patients on the HoNOS-Secure scale and security scale respectively, and was rarely clinically significant. In a context in which services rarely report routinely collected data for a range of outcomes, this study found that HoNOS-Secure captured little of any clinical change that may have occurred. Further research is needed to determine whether the HoNOS-Secure is the most suitable tool for routine outcomes reporting, but, in any event, patients, clinicians, and the public should reasonably expect services to routinely demonstrate their effectiveness.