Improving access to psychological therapy on acute mental health wards

Aims To increase access to psychological therapy on mental health inpatient wards and therefore improve patient, staff, ward and economic outcomes. Background Last year, 103,840 people spent time on mental health wards. Each admission costs over £12,000, but inpatient care does not meet patients needs for evidenced-based psychological therapies. We propose an intervention to overcome this problem. Psychologists will deliver brief or longer therapy according to patients needs and wishes, as well as work with ward staff to create more psychologically-informed care environments. Research plan WP1. We will refine our proposed intervention to ensure that it is based on the best available evidence and acceptable to key stakeholders. This work will be informed by: a) a recently undertaken systematic review of therapy trials for inpatients; b) semi-structured interviews with 15-20 patients with recent inpatient experience, 10-15 carers and 15-20 staff exploring any adaptations we need to make to our intervention so it is acceptable to stakeholders and contextual factors that may impact on implementation; c) pilot study where we will implement the first iteration of the intervention to assess implementation and acceptability. Findings from a), b) and c) will be synthesised using consensus methods. The output will be a logic model for the intervention and a therapy manual to support delivery. WP2. We will evaluate the effect of the intervention on patient, staff, ward and health economic outcomes using a cluster randomised controlled trial with a 9-month internal pilot. Thirty four wards will be recruited in pairs (to stagger recruitment/intervention delivery). We will collect outcome measures at: baseline, 6-months and 9-months. Following baseline assessments, wards will be randomised to the intervention arm of the trial plus usual care, or usual care only. Co-primary outcome measures are serious incidents (risky behaviours) on the ward collected at the cluster level and patient well-being (Warwick-Edinburgh Mental Well-being Scale; WEMWBS) collected that the individual patient level for all patients within the clusters at baseline who consent complete self-report outcome measures. Assuming a mean number of 5.22 serious incidents in the control arm, 16 wards per arm has more than 80% power to detect a reduction in mean number of serious incidents to 3.13 (40% reduction). A sample of 384 (9 patients per cluster) at baseline will give us 85% power to detect an effect size of 0.04 for WEMWBS (3.5 point change on the scale) Secondary outcomes will be average length of stay, staff sickness, patient symptoms and functioning, perceptions of ward atmosphere and staff burnout. Analysis will follow intention-to-treat principles and we will follow the CONSORT statement for cluster RCTs. Treatment effects for the cluster level outcomes will be estimated using appropriate mixed models with a fixed effect for ward gender, and a random intercept for Trust. Individual (patient/staff) level outcomes measures will be analysed using an appropriate mixed models with random intercepts forward. Baseline measures assessed prior to randomisation will be used as covariates where possible. The EQ-5D-5L will be used to assess health status and to estimate Quality-Adjusted Life Years which will be the primary measure of health benefit for the economic analysis. A resource use inventory will be used at baseline and follow-ups to identify services used by trial participants and estimate the costs of health and social care. WP3. A mixed-methods process evaluation will be embedded within WP2 and build on qualitative work in WP1 to identify contextual factors that promote/inhibit implementation and routine incorporation into everyday practice. We will record the number of patients who were offered and received therapy and the number of and quality of therapy sessions delivered. Qualitative interviews will be carried out with 30-40 inpatient staff and 30-40