BackgroundThe REAL (Rehabilitation Effectiveness for Activities for Life) research programme, funded by the National Institute for Heath Research (NIHR) from 2009 to 2015, investigated NHS mental health rehabiliation services across England. The users of these services are people with longer-term, complex mental health problems, such as schizophrenia, who have additional problems that complicate recovery. Although only around 10% of people with severe mental illness require inpatient rehabilitation, because of the severity and complexity of their problems they cost 25–50% of the total mental health budget. Despite this, there has been little research to help clinicians and commissioners to plan and deliver effective treatments and services. This research aimed to address this gap.MethodsThe programme had four phases. (1) A national survey, using quantitative and qualitative methods, was used to provide a detailed understanding of the scope and quality of NHS mental health rehabilitation services in England and the characteristics of those who use them. (2) We developed a training intervention for staff of NHS inpatient mental health rehabilitation units to facilitate service users’ activities. (3) The clinical effectiveness and cost-effectiveness of the staff training programme was evaluated through a cluster randomised controlled trial involving 40 units that scored below average on our quality assessment tool in the national survey. A qualitative process evaluation and a realistic evaluation were carried out to inform our findings further. (4) A naturalistic cohort study was carried out involving 349 service users of 50 units that scored above average on our quality assessment tool in the national survey, who were followed up over 12 months. Factors associated with better clinical outcomes were investigated through exploratory analyses.ResultsMost NHS trusts provided inpatient mental health rehabilitation services. The quality of care provided was higher than that in similar facilities across Europe and was positively associated with service users’ autonomy. Our cluster trial did not find our staff training intervention to be clinically effective [coefficient 1.44, 95% confidence interval (CI) –1.35 to 4.24]; staff appeared to revert to previous practices once the training team left the unit. Our realistic review suggested that greater supervision and senior staff support could help to address this. Over half of the service users in our cohort study were successfully discharged from hospital over 12 months. Factors associated with this were service users’ activity levels [odds ratio (OR) 1.03, 95% CI 1.01 to 1.05] and social skills (OR 1.13, 95% CI 1.04 to 1.24), and the ‘recovery’ orientation of the unit (OR 1.04, 95% CI 1.00 to 1.08), which includes collaborative care planning with service users and holding hope for their progress. Quality of care was not associated with costs of care. A relatively small investment (£67 per service user per month) was required to achieve the improvement in everyday functioning that we found in our cohort study.ConclusionsPeople who require inpatient mental health rehabilitation are a ‘low-volume, high-needs’ group. Despite this, these services are able to successfully discharge most to the community within 18 months. Our results suggest that this may be facilitated by recovery-orientated practice that promotes service users’ activities and social skills. Further research is needed to identify effective interventions that enhance such practice to deliver these outcomes. Our research provides evidence that NHS inpatient mental health rehabilitation services deliver high-quality care that successfully supports service users with complex needs in their recovery.Main limitationOur programme included only NHS, non-secure, inpatient mental health rehabilitation services.Trial registrationCurrent Controlled Trials ISRCTN25898179.FundingThe NIHR Programme Grants for Applied Research programme.