CHERISH (collaboration for hospitalised elders reducing the impact of stays in hospital): protocol for a multi-site improvement program to reduce geriatric syndromes in older inpatients

The study will be conducted in four hospitals, including two metropolitan teaching hospitals and two large regional hospitals, in two Hospital and Health Services (HHS) serving populations in South-East Queensland, Australia. During study planning, senior clinical and administrative leaders from these hospitals committed resources to participate in this partnership project, and have assisted the study team to identify two medical or surgical wards in each hospital admitting at least 50% of admitted patients aged over 65, which are suitable for participation in the study. The 8 selected wards are three general medical wards, two specialty medicine wards, one orthopaedic ward and two general surgical wards.

We will use a cluster randomised control design, randomising one ward within each hospital (cluster) to implement the Eat Walk Engage program (intervention) and the other to continue usual care (control). Allocation will be undertaken by the off-site study statistician based on computer-generated random numbers using R [29].

Data will be collected on all 8 wards before and after an intervention period, with additional process data collected during implementation on intervention wards (Fig. 1). Outcome evaluation will compare measures in pre-intervention and post-intervention samples on the intervention wards, compared to measures in samples from the same time on control wards. This study design allows measurement of changes in processes and outcomes following intervention (to understand mechanisms of any observed outcome improvement) while controlling for systematic changes at hospital level which may affect the outcomes of interest. Patient-level data collection is planned for October 2015-March 2016 (pre-intervention) and October 2016-March 2017 (post-intervention).

Study oversight will be provided by a Steering Committee consisting of the Chief Investigators which is responsible for the protocol, study timelines, and supervision of analysis by the data manager (JS) and statistician (AB). Implementation will be overseen by a working group including experienced implementation scientists (GH, IB), expert and site facilitators, site investigators and a consumer representative.

Implementation will be guided by the i-PARIHS framework, designed to support implementation of complex interventions in healthcare [26]. The central element is enabling facilitation, recognising the importance of tailoring strategies to the local recipients (e.g. patient type, staff mix) and context (e.g. culture, resources, priorities). Each site will be provided with a facilitator (0.4 full time equivalent) who will be a health professional recruited within the study hospital and trained and supported by the expert facilitators (PM, AM). Support will include group-based training in evidence-based care strategies for older patients, quality improvement methods, and the i-PARIHS framework; a monthly face-to-face forum for sharing skills and experiences between facilitators; and in-person or telephone support by the expert facilitators as required.

The site facilitator will identify and prepare members of the ward-based multidisciplinary team (MDT, including senior nursing staff, physiotherapist, occupational therapist, dietitian, social worker, speech pathologist, assistant staff, medical staff and others). They will engage MDT members as local champions and informal leaders, both informally and through monthly team meetings. The facilitator will obtain patient feedback through structured interviews and will use a suite of process measures described below to assess current performance. These data will be fed back to the MDT to help them reach consensus on priorities for improvement and to identify potential barriers, enablers and solutions. Solutions might include individual or team actions, or changes to systems or resources. For example, if the baseline data suggests gaps in achieving early mobility, barriers identified by patients and staff might include inadequate pain control, inadequate encouragement, poor team communication of mobility goals, insufficient walking aids, and no walking destination. Solutions in that setting will require consideration of patient assessment, patient and team communication, and ward equipment and design, but may also require support by education and leadership, and multiple interacting strategies may be required to produce measureable improvements. The facilitator will support MDT members using strategies such as marketing, team building, role clarification, audit and feedback, reminders and education. The facilitators will also be encouraged to develop networks, leadership support and monitoring mechanisms at each site, important for sustaining changes [30, 31]. A senior physician designated as the site investigator will support the facilitator and MDT.

Previous successful programs have described an assistant workforce to help ward staff ensure consistent provision of basic care interventions, although the type of assistant varies between models (e.g. trained volunteers [32], nursing assistants [33] or healthcare students [34]). Intervention wards will be provided with a half-time trained multi-professional healthcare assistant, who will undertake self-directed tasks (e.g. providing reading materials, mealtime set-up, reorientation), as well as specific individual or group interventions delegated by health professionals (e.g. assisted mobilisation, exercise programs, encouraging supplement consumption, reminiscence strategies) to support the MDT strategies. The assistant will be recruited locally and undertake training and competency assessment at a central site, with mentoring by the facilitator and MDT, and peer support by monthly teleconference.

Multiple methods will be used to build a comprehensive understanding of how the complex intervention is operationalised in each site [25]. Ward process measures and patient interviews will be fed back by the facilitator to the staff on intervention wards to inform and evaluate improvement strategies. Implementation measures will assess the recipients, context and facilitation activities at each site. System-level data will estimate the effect of improvements at system level.

Ward processes will be assessed by activity mapping, mealtime audits and patient interviews performed on all wards in the pre-implementation and post-implementation periods, and mid-implementation on intervention wards. Mobility and engagement patterns will be assessed using activity mapping for an 8 h day-time period (8 am till 4.30 pm) [40]. In this method of systematic sampling, the auditor observes each patient in a room for a 2 min interval, noting the highest level of a hierarchical list of possible positions (lying, sitting, standing, walking) and activities (sleeping, receiving care, self-care, eating, TV/radio, reading/games/craft, talking, exercising) and in whose company (alone, with visitors, with staff) before moving to the next patient room. Observations are repeated continuously in a consistent order for the 8 h period and summarised as the average proportion of time spent at each level for each patient [40]. Meal-time processes including patient positioning, meal set-up, assistance, interruptions and estimated meal consumption will be measured by auditing one breakfast, one lunch and one dinner on the ward using a structured observation sheet, summarising observations using proportions or averages across all patients [41]. Mealtime audits will include all inpatients present on the ward provided with a meal at the audit time (excluding patients ordered nil by mouth or receiving end of life care).

Patient experience will be assessed by semi-structured 10–15 min interviews administered to inpatients on the study wards aged over 65 who are willing and able to participate in the interview, with a randomised list of bed numbers utilised to select consenting participants until 10 interviews are completed for each ward. Interviews will be analysed using a mixed methods approach, with summary of proportions or averages for quantitative responses and framework analysis of recorded interview transcripts for qualitative responses. Process audits and interviews will be conducted by the ward facilitators pre-intervention (before randomisation occurs) and mid-implementation, and quantitative data will be summarised and fed back to the intervention wards to inform improvement strategies. Post-intervention interviews and audits will be performed on all wards by a trained research assistant not involved in implementation, to reduce the risk of bias.

Implementation measures will include evaluation of context, recipients, facilitation process, and MDT engagement in each intervention ward. Context and recipients will be evaluated by the local facilitator using the i-PARIHS framework [26], assigning scores from −2 (barrier to program goals) to +2 (enabler) for individual domains (e.g. motivation, opinion leaders, networks, senior leadership, learning environment). This iterative evaluation will be supported by reflective discussion with the expert facilitators and other site facilitators, and inform tailored strategies. The facilitation experience will be assessed by interviews by an associate investigator (GH) not otherwise involved in outcome evaluation. Interviews will be undertaken at three time points (early, mid and late implementation) with the 4 site facilitators and the expert facilitators. Open-ended interviews will be recorded, transcribed and analysed thematically, providing transcripts to the participants for checking following completion of interviews. Engagement of the team will be evaluated by meeting attendance (MDT meeting minutes), reflective facilitator feedback (meeting minutes and journal), strategy uptake (meeting minutes and facilitator feedback) and multidisciplinary staff surveys using the NoMAD tool [42].

System-level outcome data will be routinely collected data on acute length of stay, discharge destination, falls and pressure injuries. It will be collected on all 8 wards from January 2015-March 2016 (pre-intervention) and April 2016-June 2017 (post-intervention), displayed graphically and analysed using interrupted time series methods.