On the basis of their definition of outcome, Vincent & Curtis challenge the conclusions of our national study of mental health in-patient rehabilitation services where we found that the quality of services was positively associated with better clinical outcomes. In doing so, they seem to adopt a medical model and ignore important insights arising from patient-centred concepts of recovery. They regard ‘symptoms, functioning and relapse’ as key, whereas a recovery-oriented approach would also value the quality of therapeutic relationships, the promotion of autonomy and better quality of life.

Our aim was to try to bridge this ideological divide by placing patient-reported outcomes on an equal footing with so-called clinical measures. At the risk of excessive pedantry, it would be fair to say that measures of patient experiences of care, such as those we used in our study, could be considered assessments of process rather than outcome, although they are nevertheless an important aspect in the measurement of service quality. In any case, such measures are usually referred to as patient-reported outcome measures. We found a positive association between our service managers' assessments of the quality of their own services (using our standardised measure, the Quality Indicator for Rehabilitative Care, QuIRC) and patients' experiences of care. This corroborates the service managers' ratings and strongly suggests that improving service quality will result in a better service user experience – surely an ‘outcome’ everyone can relate to as worthwhile.

We also found that greater quality of mental health rehabilitation services was associated with greater service user autonomy. We gave a number of possible reasons why we did not find a positive association between service quality and service user quality of life, the main one being that the measure we used focuses on experiences outside of an in-patient setting (relationships with family/partner, work, income, etc.). Our findings reflect the reality for people in in-patient mental health rehabilitation units who tend to have lengthy admissions (in our study, their current admission was 18 months on average, with 8 of these in the rehabilitation unit) due to the severity and complexity of their symptoms and severe impairment of social functioning, all of which have a negative impact on their social inclusion and quality of life. Nevertheless, it is absolutely correct that rehabilitation services should (and do) aim to help service users achieve a successful community life which, ultimately, should be reflected in their quality of life.

Our findings reflect the focus of our study – in-patient mental health rehabilitation services deal with individuals at the beginning of their rehabilitation, when they are most severely unwell and least able to engage in the community. Later phases of our research will provide further data on the longitudinal outcomes, including social functioning and successful community living.