A complex intervention to reduce avoidable hospital admissions in nursing homes: a research programme including the BHiRCH-NH pilot cluster RCT

Aim

In the original proposal, the aim was to reduce rates of hospital admissions from nursing homes for ambulatory care-sensitive conditions (ACSCs), conditions whose early detection and active management in nursing homes could have prevented the exacerbation that led to the unplanned admission. We aimed to optimise and pilot test a pragmatic, acceptable, evidence-based intervention for ensuring active management of ACSCs in nursing homes, thus reducing avoidable hospital admissions.

The overall aim was revised to the following: to develop and pilot test an intervention to improve rates of early detection and treatment for four conditions. This allowed us to have an explicit focus on the active management needed for early detection and treatment in order to reduce hospitalisations. This change in emphasis meant that the project was more closely aligned with policy on improving health care for residents as a route to decreasing hospitalisations. Therefore, we called it the Better Health in Residents in Care Homes with Nursing (BHiRCH-NH) study.

Use of interviews and informal conversations rather than focus groups to gather feedback during the feasibility study

We had intended to conduct two focus groups with staff in the two feasibility homes to further refine the intervention and implementation support. These proved difficult to arrange as the staff in both homes were under significant time pressure and unable to take time out to join a group. Indeed, even seeking one-to-one time with them proved difficult, because of pressure of work. Instead, we conducted one semistructured interview (during which there were frequent interruptions), which we transcribed, and otherwise made notes during impromptu conversations with staff. We took a pragmatic approach to analysis, noting key points to share with the team.

Location of Yorkshire pilot homes

In Yorkshire, we had to recruit nursing homes outside Bradford for the pilot study. Following a presentation we made to the Bradford and District Clinical Commissioning Group (CCG) in autumn 2017, tools similar to the project intervention tools were sent to all nursing and care homes in the district, meaning that these homes were essentially pre-exposed to the intervention. We therefore had to move the intervention to areas that had not had this exposure.

Duration of post-intervention pilot study period

In the original proposal, we expected to have a 12-month post-intervention study period. Owing to delays of approximately 10 months in obtaining ethics approval, the post-intervention study period was reduced to 6 months. This meant that we had a shorter period in which to embed the intervention and were unable to look at the seasonal effects on the primary and secondary outcome measures.

Number of, and approach to, clinical record reviews

We reviewed 11, as opposed to 30, nursing home residents’ records because of the shortened post-intervention period and the lower than anticipated numbers of hospital admissions. We had planned to review a random sample of residents’ records, using a structured audit template comparing the care pathways developed for the intervention with a pathway analysis of the events leading to hospitalisation or treatment for an ACSC, and noting if changes were made to the care pathway documentation.

This approach depended on the intervention being adopted as planned. However, in the event, only 16 Stop and Watch Early Warning Tool (S&W) forms were completed in the study period, with 11 of them coming from a single home. Furthermore, we were able to locate only three of the eight completed care pathway forms. During interviews, some nursing home staff stated ‘we do this anyway’, and some were able to point to local practices that formalised and supported those claims.

We therefore changed the approach to focus on any reference to the observations or actions in the care pathway. We hoped that this would allow us to confirm (or refute) the claim that the principles of the intervention were already in use as normal practice. We also hoped that it would give us insights into which aspects of the intervention were seen as most useful. We thought that this would be the most helpful approach in deciding whether or not to pursue a definitive trial.

Use of laptops for data collection and entry

We had intended that we would enter data while in the field using laptops. We instead relied on completing paper-based case report forms, which were then manually entered into the database prepared by University College London’s Clinical Trials Unit, Priment.

Use of additional human resources

Methods of data collection and analysis

We worked with stakeholders, including nursing home staff and our Carer Reference Panel (CRP), to develop and adapt the INTERACT tools for use in the UK.

The following were conducted:

1. Co-applicant specialists in geriatric medicine (JY) and primary care (LR) reviewed care pathways identified in a search for care pathways (see Appendices 1 and 2). This led to version 1 of the care pathway.

2. We obtained expert panel input via e-mail on this version of the care pathway (see Appendix 3). Experts included three international advisory group members, two members of the British Geriatrics Society and three members of the Royal College of General Practitioners. This input was reviewed by John Young and Louise Robinson, and led to version 2.

3. We held a consensus workshop with 17 stakeholders with diverse areas of expertise, including eight nurses (four community/district nurses, three nursing home nurses and one nurse consultant), two nursing home managers, three care assistants, one geriatrician and three family members (see Appendix 4). This led to the final draft of the care pathway.

Limitations

The main limitation was the limited amount of previous research to guide our approach. Only 22 papers were identified in the literature review. Overall, the quality of intervention studies was rated as being highly variable. There was a lack of robustly conducted RCTs (only two trials were rated as being of ‘strong’ quality). Insufficient attention was paid to key methodological issues, particularly the clustering effect within nursing homes. Furthermore, there were few studies conducted in the UK, thereby having questionable relevance to the UK context.

Key finding

We identified care pathways for ACSCs in nursing homes for our four conditions (respiratory infection, exacerbation of CHF, UTI and dehydration) and adapted them in consensus co-design workshops with staff and family carers. We developed a care pathway to facilitate assessment and diagnosis of the four index conditions from existing care pathways used to facilitate diagnosis and treatment of health conditions in care home residents. The final version of the care pathway is presented in Appendix 5.

Inter-relationship with other parts of the programme

The care pathway was one of three intervention tools that were included in the feasibility study.

Methods for data collection and analysis

We used three methods of data collection: semistructured interviews, rapid research review and consensus-building.

Any limitations

There was a limited literature to inform us about the knowledge and skills required for nursing home care assistants and nurses to achieve early detection, assessment and reporting of acute changes in residents’ health.

Key findings

We derived a set of key knowledge and skills for nurses (see Appendix 8) and for care home staff and family members that are required for early detection of the four ACSCs. These included:

• knowledge of how ACSCs may manifest

• ability to detect these changes during daily care

• knowledge of residents’ medical conditions, care plans and their baseline behaviour and symptoms

• continuing assessment skills

• leadership skills

• communication skills.

Inter-relation with other parts of the programme

We examined the use of the knowledge and skills matrix in our feasibility study. We expected Practice Development Champions (PDCs) would use it to conduct a gap analysis and to direct care staff and family members to learning resources that could help to address these gaps.

Methods for data collection

We made detailed notes of the consensus workshop discussions and participants generated written notes.

Method for data analysis

We analysed the notes, along with written materials generated by the participants, looking for key themes.

Key findings

The criteria for PDCs were thought to be appropriate. Participants identified situations in which there could be gaps in coverage of staff implementing the intervention. In particular, staff from evening and weekend shifts would need to be recruited to the PDSGs to ensure coverage across the week. Further issues with the language and structure of the staff handbooks were noted, in particular with respect to readability for care assistants and family members.

Participants reported that the purpose of the PDCs and the training methods needed to be clearer. The staff handbooks, which had been designed to provide information about the intervention, and about introducing and embedding change, were felt to be in need of improvement, including the following: be more visually attractive, use simpler language and provide less information.

Based on these findings, we developed the final version of the purpose, role and person specification of PDCs (see Appendix 9). We agreed the key topics to cover in the preparation workshop, using both presentations and interactive exercises. Box 1 presents the topics for the PDC preparation workshop.

Limitations

We intended to provide a 2-day preparation workshop for PDCs. Owing to a diary mix-up with one of the PDCs, we reduced the workshop to 1 day.

Inter-relationship with other parts of the programme

We finalised our approach to recruiting and supporting PDCs and identified the personal characteristics and professional skills required of them. The support to be provided included a preparation workshop, monthly telephone support (with AB, the programme manager) and a website with additional resources for nursing home staff. We tested this implementation support in our feasibility study in WS2, WP1.

Aim

The aim was to explore family members’ perspectives on their involvement in the timely detection of their family members’ changes in health in UK nursing homes. Specifically, we sought to address:

• How are family carers involved in timely detection of changes in their relatives’ health in nursing homes?

• How can family carers be supported to engage in timely detection of their family members’ changes in health?

Methods for data collection

We conducted 14 semistructured interviews with family carers of residents living in 13 different nursing homes (see Appendix 10). This allowed us to obtain in-depth information on their perceived and preferred role. All interviewees were adult children and lived near the nursing home. We recruited family members who were regularly involved with their relative. We used telephone interviews to gather data from family members who were unable to attend a face-to-face interview.

In addition, we finalised drafts of the following:

• a handbook on the intervention and its implementation for the PDCs

• a handbook about the intervention and its implementation for care home owners and managers

• a workbook on implementation for the PDSGs.

Method for data analysis

Interview data were transcribed verbatim and analysed using thematic analysis. 67

Any limitations

We were unable to identify family members who lived further away from the care home, despite advertising nationally. We had hoped to include these, as they provide an additional perspective on family involvement.

Key findings

Families were involved in the timely detection of changes in health in three key ways:

1. noticing signs of changes in health

2. informing care staff about what they noticed

3. educating care staff about their family members’ changes in health and how they manifest.

Families suggested that they could be supported to detect early changes in health by developing effective working practices with care staff. 66

Inter-relation with other parts of the programme

Family involvement was examined in the feasibility study in WS2, WP1. The intention was for care staff to establish how best to work with family members who would like to be involved in the intervention, recording preferences in each resident’s care record, as well as explaining the purpose and application of the S&W.

Five components of the complex intervention

Implementation support

Implementation support included asking nursing home managers to engage in a change management programme, which included:

• identifying two PDCs

• establishing a PDSG to support the work of the PDCs in introducing and embedding the intervention tools.

The study team provided training for the PDCs’ role. We provided two implementation support handbooks:

1. for care home staff (for managers and PDCs)

2. for staff and PDSG members.

The handbooks were to be used in the nursing home to help staff to understand the intervention and support its implementation. The handbooks offer background information on some of the approaches to change used in this project, promote an understanding of how the intervention can be implemented in the differing contexts of each home, facilitate and improve the learning of others in the nursing home, help teams learn and act alongside the people for whom they care, maximise opportunities for all team members to enhance their leadership capabilities, and offer information for residents and their family carers to help them become active participants in developing practice in their care homes.

Can the intervention be delivered as intended?

At the end of the 3 months, we carried out a semistructured interview with one intervention champion to learn about their experience of taking part in the feasibility study, the barriers to and facilitators of using the intervention, the training workshop and the implementation support. In addition, we held informal conversations with six further members of staff from the two nursing homes to explore what had worked well, what had worked less well, and what facilitated and hindered adoption of the intervention. Notes were taken during the conversations. The interview was analysed alongside the notes from impromptu conversations with care staff.

What further refinements are required to the implementation guidance?

Is the approach to collecting data feasible?

What are the resource requirements to collect and analyse the data?

Although we could collect individual-, process- and care home-level data, it was more time-consuming than expected.

Stop and Watch Early Warning Tool

We agreed to emphasise the purpose and use of the S&W during the training workshop for PDCs, including our expectation that family carers be offered the opportunity to have a role in alerting staff to any changes they noticed in their relatives’ health.

Care pathway

We clarified the links between sections and changed the form so that data could be entered directly. We expected that PDCs would provide further education to their nursing colleagues about its purpose and aims.

The situation, background, assessment, recommendation technique

We expected that nursing home nurses would use the SBAR technique, when communicating with primary care, as an integral part of their engagement with the intervention.

Involving families

We integrated family involvement into use of the S&W by emphasising the role of family members in early detection of changes in health. Family members were expected to notify staff if they observed any changes in their relative, which, in turn, would prompt the staff member to use the S&W.

We expected nursing home staff to inform family carers about the research at the start of the project and invite them to participate. Nursing home staff were expected to describe to family carers, who had expressed an interest in participating, the potential roles they could play in the project.

Knowledge and skills development

As the PDCs had not engaged in this aspect of the intervention, we decided to use the knowledge and skills matrix as a way of describing nurses’ knowledge and skills at the outset of the study. This would provide us with useful contextual information.

Components

Data collection

Resource requirements

We recruited a research facilitator from the nursing home staff to support the research team with recruiting residents and family carers and with accessing residents’ records (see Appendix 16). We also had support from the local CRN with recruitment and consent processes. We reduced the number of questionnaires for use in the pilot trial. For example, we removed the Context Assessment Index as it was too burdensome. Furthermore, we found that some of the questionnaires used in the feasibility study did not provide a great deal of additional information; for example, there was overlap between the Dementia Quality of Life (DEMQOL) scale and the EuroQol-5 Dimensions, five-level version (EQ-5D-5L), as measures of quality of life. We retained the EQ-5D-5L and the EQ-5D-5L proxy (reported by a carer).

Trial design

We conducted a pilot cluster RCT of the BHiRCH-NH study intervention in nursing homes and an embedded process evaluation.

Study population

Randomisation and masking

Nursing homes were randomised prior to the intervention starting (considered to be the day after the training workshop); four in West Yorkshire and three in Greater London (seven in total) were randomised to the intervention arm and four in Yorkshire and three in Greater London (seven in total) were randomised to the treatment-as-usual (TAU) arm, stratified by location, by an independent statistician from the Priment Clinical Trials Unit (CTU). The statistician and health economists analysed the data while remaining blind to randomised allocation. Data were not collected by people who were blind to allocation because of resource issues and the visibility of the intervention in nursing homes.

Ethics and consent

Ethics approval was obtained from the Queen Square London Research Ethics Committee (reference number 17/LO/1542). We gained written permission from the manager, regional manager or owner for the intervention to be implemented at nursing home level. We obtained individual consent for the collection of individual-level outcome data from residents. If they lacked capacity, we consulted a personal or professional consultee, as per the Mental Capacity Act 2005,74 for their agreement for the person to participate. A member of the nursing home staff identified potentially eligible residents. Adhering to the Mental Capacity Act 2005,74 the research team conducted a capacity assessment with respect to participation. Consent was sought from family carers associated with residents recruited to the study and from nursing home staff to answer questionnaires and/or take part in qualitative interviews.

Care record reviews

To explore fidelity to the intervention, two nurse researchers (SG, HP) reviewed a convenience sample of five records for residents who had been admitted to hospital, or received treatment in the nursing home, for an ACSC. The nurse researchers used free text to record reference to the observations and actions in the care pathway.

To ascertain whether or not hospital admissions were avoidable, one of the co-applicants (JY) and two nurse researchers (SG, HP) extracted data from the care home records of 11 residents who had been hospitalised, and completed the first section of the structured implicit record review (SIRR) tool. 75

Trial procedures

The trial ran for 10 months between November 2017 and August 2018. Each nursing home appointed a research facilitator to support the research team with recruitment activities and to ensure that nursing home-level data collected without consent from resident files were pseudoanonymised prior to being given to the research team. The intervention was delivered as described in Workstream 2, work package 1: test the intervention – feasibility and acceptability study. Nursing homes in the control arm received TAU according to existing local policy and practice. Data were collected for 3 months prior to the intervention phase starting and for 6 months after. All medications and treatments were permitted.

Trial measures

We collected data across three domains (Table 3):

1. Individual-level data on nursing home residents, their family carers and staff – these were collected from staff, family carers or residents who had given informed consent or from residents from whom we obtained agreement from a consultee.

2. System-level data – collected from each nursing home as a whole using existing record systems, pseudoanonymised and provided by the nursing home manager or research facilitator.

3. Process-level data – we examined how the intervention was implemented in practice, including fidelity. To understand the effectiveness of the trial procedures, we collected data on recruitment rates and consent and the numbers of family carers who wished to be involved in the residents’ care, and assessed the completeness of outcome measures, data collection and the return rate of questionnaires.

We conducted semistructured interviews with five nursing home managers, six PDCs, three nurses, seven care assistants and five family carers from across the five homes that were able to deliver the intervention. We explored participants’ views on their knowledge of, engagement with, and views on effectiveness of, the intervention. Family carers were purposively sampled to ensure a range of genders, ages and types of family carer. The interviews lasted ≈ 40 minutes, and were conducted either face to face or by telephone.

Serious adverse events

Participants were monitored each month for potential adverse events. Serious adverse events (SAEs) were defined, using the standard operating procedures of the Priment CTU, as any untoward occurrence that resulted in death, was considered life-threatening at the time of the event, required hospitalisation or prolongation of existing hospitalisation, resulted in persistent or significant disability or incapacity or was any other important medical condition. Each untoward event was then considered and designated as ‘related’, that is it resulted from administration of any of the research procedures, or ‘unexpected’, that is the type of event is not listed in the protocol as an expected occurrence.

Statistical analysis

We followed Consolidated Standards of Reporting Trials (CONSORT) guidelines for reporting randomised trials. Given that this was a pilot trial, analyses were mainly descriptive, focusing on recruitment, participant characteristics, other baseline and outcome variables, loss to follow-up and tabulation of SAEs. We summarised completeness of data collection on all outcome measures. Analyses were carried out in Stata^® version 14 (StataCorp LP, College Station, TX, USA). We aimed to consider the rates of hospital admission for ACSCs to inform the sample size calculation for a full trial.

We calculated costs associated with the intervention. Resource use associated with NHS and social care provision was collected and costs were calculated from the NHS and Personal Social Services (PSS) perspective. We assessed the feasibility of calculating quality-adjusted life-years (QALYs) for residents and family carers. The health economic analysis was conducted using utility values (to calculate QALYs) from resident self-reported EQ-5D-5L questionnaires.

We provided an initial estimate of the incremental mean cost per QALY gained in the intervention arm, compared with the TAU arm, and reported the probability that the intervention is cost-effective for a range of willingness-to-pay (WTP) values for a QALY gained. We then estimated the price that a health-care decision-maker would be willing to pay to have perfect information regarding all factors that influence which treatment choice is preferred as the results of a cost-effectiveness analysis in the form of expected value of perfect information (EVPI) and expected value of partial perfect information (EVPPI). 76,77

Qualitative analysis

Verbatim transcripts of the interviews were made. Key themes were coded using framework analysis. 66 This involved the following stages:

1. Familiarisation – reading through each transcript and making notes.

2. Identifying a thematic framework – identifying key themes, issues or discussion points. We used the topic guide as a starting point.

3. Indexing – annotating transcripts to identify patterns.

4. Charting – rearranging the data and framework to create order.

5. Mapping and interpretation – understanding how themes relate to one another.

The process was assisted by the use of NVivo qualitative data analysis software, version 10 (QSR International, Warrington, UK). For the family carer transcripts, stages 1 and 2 were conducted by the CRPs in collaboration with the research team. A workshop on framework data analysis was held with the CRP and members were also offered the opportunity to send input by post (see Box 2).

Stages 3–5 were conducted by members of the research team. A sample of the interviews was independently coded by an additional member of the team to check levels of agreement. For interview transcripts from nursing home staff, the research team conducted stages 1–5. The thematic framework developed at stage 2 for the family carers was adapted for these staff transcripts. We explored the extent of engagement with the intervention in terms of knowledge about, and use of, its three elements: the S&W, the care pathway and the SBAR technique. We explored the effectiveness of the key elements of the implementation strategy: PDCs, the training workshop, monthly coaching telephone calls, the handbook and the web-based resources. We presented the qualitative data under these headings as they directly address the questions posed in the pilot study about the knowledge of, and engagement with, the intervention and its implementation strategy. We have added commentary to the discussion about how these findings link to the PARiHS framework, with an emphasis on context and facilitation.

Trial settings

Fourteen (which was the target number) nursing homes were recruited and randomised as planned: seven to the intervention (three in London and four in West Yorkshire) and seven to the control (three in London and four in West Yorkshire). One Yorkshire intervention nursing home dropped out during pre intervention as it was closed by its owners. A further intervention group nursing home in London dropped out following PDC training as it was unable to implement any aspects of the intervention.

Nursing homes were predominantly privately managed; one was a local authority nursing home. They had a median of 50 residents [interquartile range (IQR) 34–68 residents]. The majority (73%) were ‘Dementia Registered’ and were rated overall as ‘good’ by the CQC. Table 4 allows comparison with the national average overall CQC rating for care homes: 3% are rated ‘inadequate’, 31% are rated as ‘requires improvement’, 65% are rated as ‘good’ and 1% as ‘outstanding’. Therefore, overall CQC ratings for nursing homes in this trial were above average, when compared with UK care homes.

Table 5 shows nursing home characteristics at baseline. There were 13 nursing homes, with a median of 60 resident places (IQR 36–71 resident places). A median of 20 residents had dementia (IQR 15–33 residents) in each nursing home. There were relatively few hospital admissions, ambulances called and A&E attendances at the nursing home level in the month before baseline. There was a median of three hospital admissions per participating home (IQR 2–5), three ambulances called (IQR 2–6) and three resident A&E attendances (IQR 2–5).

As can be seen in Table 6, all nursing homes had a dedicated GP and access to chiropody. Support from other external agencies varied. For example, only one-quarter of the nursing homes received regular input from a geriatrician and 38% had input from a district nurse.

As can be seen in Table 7, few nursing homes reported having used the interventions in this trial before. Prior to the trial, one nursing home used the S&W, two nursing homes used care pathways for ACSCs and four used the SBAR technique to aid communication with primary care.

Table 8 presents the nursing home characteristics at baseline, by randomised group. There was a higher median number of resident places in the TAU group than in the intervention group, and a higher median number of residents with dementia in the TAU group, possibly because the TAU group had dementia-specialist homes, whereas there were no such homes in the intervention group.

As can be seen in Table 9, there is little difference at baseline between the homes in terms of receipt of medical and allied health services.

There were minor differences between intervention and TAU nursing homes at baseline in terms of reported use of intervention tools prior to the trial (Table 10). None of the intervention homes reported using the S&W or the care pathway; three nursing homes in the TAU group reported using the SBAR technique.

Residents

As can be seen in Table 11, we recruited 237 residents: two-thirds were female, they were predominantly white (90%) and the median age was 86 years (IQR 80–91 years). The median Barthel Index score was low, at 27 points (IQR 9–64 points), indicating a high level of dependence. Only 3% of the residents had an admission for an ACSC in the month before baseline, 11 (5%) residents had had at least one ambulance called and 12 (5%) had had an unscheduled GP visit or telephone contact.

There were similar resident characteristics between the intervention and TAU groups at baseline (Table 12).

Family carers

As can be seen in Table 13, we recruited 91 family carers, two-thirds of whom were female. Their median age was 63 years (IQR 57–71 years). The majority wanted a role in noticing changes in their relative’s health and informing staff of this. Other roles that family carers wished to take on included informing staff of their relative’s background, personal history and preferences, managing finances and taking the person out for visits. Table 13 presents details of family carer characteristics at baseline, and Table 14 presents family carer characteristics at baseline by randomised group.

Staff

We recruited 132 staff, with a median age of 42 years (IQR 30–53 years); 12% were male, 50% of nurses spoke English as a first language and 59% of staff described themselves as white. Most staff had worked at the nursing home for ≥ 1 years (71%) and 30% were qualified nurses (Table 15).

Scores on the Person-centred Care Assessment Tool (P-CAT) (possible range 13–65 points) were generally positive, with a median score of 49 points (IQR 46–53 points) (see Appendix 22). The majority of nurses were positive about the quality of communications they had with the residents’ GP (see Appendix 23). Nurses were positive about their knowledge and skills (see Appendix 24).

As can be seen in Table 16, the intervention group had a greater percentage of white staff. Two-thirds of participating TAU staff and three-quarters of intervention staff were not nurses.

Practice Development Champions

The implementation strategy involved establishing two PDCs working with a PDSG in each participating care home to introduce and embed the three intervention tools: the S&W, the care pathway and the SBAR technique. PDCs were supported in their role with a training workshop, a handbook, web-based resources and monthly telephone coaching from the co-applicant practice development experts (KF and BM).

One of the intervention homes withdrew from the study after their two PDCs had participated in the training workshop. They were unable to comply with the requirement that the PDCs lead the implementation of the intervention. This was a result of a combination of a lack of knowledge, skills and expertise on the part of the PDCs, and a lack of management support.

Across the five remaining intervention homes, none introduced the intervention as expected, although most PDCs engaged in some kind of information sharing and training about the project. One PDC described her role as:

To come back and train staff on what we’d learned and implement the tool and monitor its effectiveness within our setting.

Nursing home 6, PDC3

Some reported that they had used a cascade model of training the nurses to train the care staff in using the S&W and had noticed that care assistants were now reporting changes in residents’ health quicker. In another home, the PDC took time out to brief nurses one to one and then used time at the start of each shift to remind small groups of staff to be aware of possible changes in the residents’ health.

The PDCs’ effectiveness at implementing the intervention (introducing and integrating it into everyday practice) was variable. In one home, PDCs had integrated elements of the complex intervention into their regular morning meetings, where they reviewed each resident identifying any concerns and integrated it into the electronic care records.

Practice development support groups

None of the PDCs set up PDSGs. Some tried but found timetabling very difficult. One PDC chose not to set one up as they had only five nurses in total. Therefore, they thought that there was no need to set up a separate group as they already had a mechanism for effective communication with staff. One PDC got support for the intervention from her/his nurses, as they were the ones using the tools and reporting back at staff meetings. One PDC felt that, had the groups been set up, they would have been able to more effectively engage family carers in the BHiRCH-NH study.

Training workshop

One PDC reported that the training workshop at the outset of the study was very helpful. They noted, in particular, the input from the consultant geriatrician (JY) regarding signs of early deterioration and input from the family member. They also valued the discussion about how to share knowledge with colleagues:

It actually was very nice to have a relative there giving her personal experience, which was that little bit of an extra insight. That is one of the main things.

Nursing home 9, PDC1

Several PDCs suggested that the study team communicating more directly with staff about the project and providing staff training would have been helpful.

The handbook

Engagement with the handbook was variable. One PDC said they had not had time to look at it, whereas another considered it to be ‘brilliant’ and others pointed to it being helpful to have information about the research underlying the project for when their colleagues asked them questions. The handbook provided them with a reference point. They felt that having the handbook made it easier to communicate the goals of the BHiRCH-NH study to their colleagues, as they could present the underlying research evidence:

I had to deliver it [details of the intervention] to the carers and other nurses. I knew exactly what the goals were and then, when actually delivering it, I knew a little bit more knowledge instead of just saying ‘this is this sheet, this is’. I could answer some of their questions and then I had it [the handbook] by my side, so if they asked me something I didn’t know, I could quickly flick through it [the handbook] and say ‘this is the reason why’.

Nursing home 9, PDC1

The monthly coaching telephone calls

Several of the PDCs pointed to the helpfulness of the monthly telephone calls with the practice development expert. They found these to be encouraging, providing useful reminders, and found it reassuring to know there was someone at the end of the telephone if they had queries:

[The telephone call] was helpful because I told them [the practice development coaches] what I’ve done. They said to me ‘well done’. When I do something good, and someone says to you ‘well done’, then our work is not for nothing.

Nursing home 12, PDC2

It keeps it fresh in your mind as well, because you come to the training day having not spoken to anyone again I probably would have just forgotten about it. It would go to the bottom of the pile. The fact that I knew she was ringing me every month kept my mind fresh and I knew that I would still keep using it because I was going to have this conference call with her.

Nursing home 6, PDC3

One of the practice development coaches thought that the offer of monthly calls was not welcome in the home that dropped out. Despite numerous attempts at making contact and establishing a schedule of telephone calls, no telephone calls were established. He/she concluded that continuing the process potentially breached the ethics procedures of voluntary participation; thus, further attempts were withdrawn.

Support from the manager

The PDCs pointed to the critical role of the manager in lending them authority, putting their weight behind the project and being someone with who they could discuss progress. One PDC felt that the manager should have had helped more to disseminate the BHiRCH-NH intervention across the care home because she had more authority. Those who were given more time by their managers felt more able to implement the BHiRCH-NH intervention. The PDC also felt that staff lacked motivation and would only undertake training and complete forms that were mandatory and part organisational policy:

I found it very difficult to chase everyone and to explain, and that’s why I asked my manager a few times . . . ‘when you have a meeting or can you please just pass it over?’ because it is different when it comes from the managers than when it comes from me . . . it’s the boss . . . not all the staff considered [it] really important because . . . everyone saw [it as] something optional.

Nursing home 4, PDC4

Relationships with primary care and out-of-hours services

Relationships with primary care and out-of-hours services also had an impact on implementation. Most PDCs reported excellent relationships with their primary care providers. Although one PDC felt that they were able to communicate changes well, at the weekend they had greater difficulty getting the needed services:

The only problem now in our area is that, over the weekend, the Rapid Response, there is a wait time. They don’t have a nurse prescriber. Sometimes they will say ‘I am a nurse like you, it’s your responsibility in that nursing home’. They don’t have enough doctors over the weekend. So they say ‘I’m so sorry, we don’t have a doctor. If you consider it an emergency, just send them to the hospital’. So we don’t have any antibiotics or anything over the weekend. We have to wait until Monday.

Nursing home 12, PDC2

Nursing home staff engagement

Rates of hospitalisation for ambulatory care-sensitive conditions

The aim of the study was to develop and pilot test an intervention to improve early detection, assessment and treatment of four common health-care conditions (respiratory infection, exacerbation of CHF, UTI and dehydration) to reduce potentially avoidable hospital admissions for these conditions. At baseline, the rates of hospitalisation for any of these conditions were low: 0.4% of the cohort had an admission for respiratory infections, 1% for UTI, none for dehydration and 0.4% had an admission for chronic heart failure in the month before the trial started. There were six admissions in total from the 235 residents in the study. Considering the whole 6-month study period, 25 study participants (15%) had an unplanned hospital admission for one of the ACSCs studied. Rates of events (per 100 person-months) are given in Table 17. The incidence of the primary outcome of interest suggests that this may not be the best potential primary outcome measure in a future study (see Table 17).

Considering secondary outcome measures for a definitive trial, we found that a slightly higher proportion of participants (n = 38, 16%) had an A&E attendance during the follow-up period, 42 (18%) had at least one ambulance called, 29 (12%) had an unscheduled (out-of-hours) GP visit and 21 (11%) had died. The incidence of these events was still relatively low and not sufficiently frequent to be used as a primary outcome in a definitive study. Comparing TAU with intervention nursing homes, the difference for those who had an admission who were in the study at baseline was 3.0% [95% confidence interval (CI) –6.4% to 12.4%]. This comparison is limited because both of the nursing homes that dropped out were in the intervention arm; thus, their residents were lost to follow-up, resulting in fewer admissions from that arm (see Appendix 25).

Test of the assumption that a hospitalisation for an ambulatory care-sensitive condition is a proxy for avoidable hospital admission

One nursing home had no eligible residents, because those admitted to hospital had died and/or their care records were no longer available. We intended to review only residents admitted for ACSCs, but we expanded our sample to include any hospital admission, partly because it was not always possible to identify if an admission was for an ACSC and because of low numbers overall. For the same reason, we also included residents who died in hospital, as long as their records were still available.

Three of the 10 admissions that could be assessed were deemed to be potentially avoidable. This is consistent with national findings79 that suggest that about one-third of admissions for older people are potentially avoidable. This provides some support for the validity of our approach.

None of the care records provided a complete picture of a resident’s health in the period leading up to admission. Nonetheless, the panel agreed on 10 of the 11 records. Only one was classified as ‘unable to decide’ because of lack of information. These included definitely yes (n = 1), probably yes (n = 2), probably not (n = 5), definitely not (n = 2) and insufficient information in records (n = 1). Only 3 of the 10 hospital admissions rated were judged to be avoidable. From this review of care records, we concluded that an ACSC admission is not a proxy for avoidable admission.

Serious adverse events

There were no differences in SAEs between the TAU and intervention groups (Table 18). There were 104 SAEs in 74 residents during the study, and 34 residents died (19 in the TAU group and 15 in the intervention group). Of the 103 SAEs, hospitalisation (any cause) was the most common (n = 50, 48% of SAEs).

Care staff-related data

We were able to collect data on most of the recruited care staff at baseline (n = 132). For example, 129 (98%) care staff gave demographic details and 127 (96%) completed the P-CAT. All nurses (n = 40) completed the nurse–GP communication tool and 34 out of 40 (85%) completed the knowledge and skills tool. Subsequent attrition in response to these scales was secondary to nursing homes dropping out of the study, rather than staff being unwilling or unable to complete the forms.

Resident-related data

Demographic and functional ability data were available on most residents at baseline (n = 235), including gender (99%), ethnicity (95%), marital status (98%) and functional ability (Barthel Index) (98%). Data on years of education were more difficult to collect, as nursing home staff often did not know this about the residents. We cannot be sure that no admissions or visits to acute hospital were missed, as nursing homes did not have centralised systems for collecting these data.

At subsequent time points, one or two nursing homes from both intervention and TAU groups did not report some data. The number of data reported decreased over time, notwithstanding nursing home dropout.

Resident involvement

We also sought to integrate the perspective of care home residents to inform how we introduced and implemented the intervention to the care home setting, with specific reference to a review of the study information materials for residents in the pilot trial.

We conducted an internal evaluation to review the process and outcomes of the PPI work by the CRPs. 80 The evaluation of the PPI involvement was conducted at two time points: baseline data were collected in January 2017 and data towards the end of study were collected in December 2018.

Further PPI representation was present on the National Institute for Health Research (NIHR) Programme Steering Committee (n = 1) and the International Advisory Group (n = 2).

Pre-funding preparation

Shirley Nurock and Barbara Woodward-Carlton initiated the focus on family involvement in health care in nursing homes. They were involved in writing the grant application, ensuring a prominent focus on family involvement. At this stage, we drew on our study experience of working with a CRP and expanded it to use two panels, one at each study site.

Supporting the study

There are five areas in which PPI was influential in the study’s delivery.

Evaluation of the patient and public involvement

Evaluation of the PPI work in the study was undertaken at two points, January 2017 and December 2018, using a survey questionnaire, based on the Public Involvement Impact Assessment Framework (PiiAF) approach. 80 In January 2017, we sent invitations to the 11 active CRP members. Eight members responded (seven females, one male; seven were a family member of a person living/lived in a care home and one was a person living with dementia). Their prior involvement in PPI ranged from 2 to > 20 years (not specified). This was an experienced PPI group, whose members had all been involved in other research studies or as monitors for the Alzheimer’s Society, with many advocating for and supporting care for people with dementia in their communities.

In December 2018, 11 evaluation forms were circulated to CRP members. The evaluation addressed the nature of involvement, what had gone well from a PPI perspective and what could have been improved, and asked for panel members to identify ways in which the PPI element influenced the outcomes of the study. Nine responses were returned.

Development of the intervention

It is surprising that, as we developed a complex intervention in collaboration with key stakeholders, the intervention had little traction. NIHR have stressed the importance of effective partnership-working as essential for practice development in nursing homes. 83 We explicitly sought end-user input into the design of the intervention and its implementation support.

It is possible that the collaborative approach we took did not fulfil many of the key tenets of co-production. 84 The intervention may have been more acceptable and effective had we developed it collaboratively with each nursing home. In this way, we could have addressed the specific issue any particular nursing home was facing, if any, with respect to avoidable hospital admissions. Such an approach would be closer to the ambitions of quality improvement, rather than research per se, albeit cognisant of its challenges. 85 An example of this is ‘My Home Life’ [http://myhomelife.org.uk (accessed 20 February 2020)], which focuses on working in partnership with care home providers to identify and collaborate on improving the quality of care in care homes. Furthermore, such an approach would have entailed closer collaboration with all levels of nursing home staff, including care assistants, domestic and estates staff. 86–88 Certainly, closer working with care assistants would have been of benefit, whether as advisors or participants. In retrospect, we had limited engagement with them, even though they play a key role in early detection.

In essence, we came to the consensus workshops with a proposed solution based on findings from our study, which we sought to refine, which suggests a power imbalance both in terms of the definition of the problem and its proposed solution. We succeeded in including all perspectives and skills, but perhaps would have benefited from having more detailed engagement from a reference/advisory group solely consisting of nursing home nurses and care assistants. We respected and valued each person’s knowledge. Indeed, we purposely sought stakeholders with expertise to complement our own. We did not assess stakeholders’ perspectives on what they got from engaging in consensus workshops, but attendance figures suggest that they derived benefit. Over the period of three consensus workshops, we had 17, 10 and 12 participants, respectively. The collaboration with the CRP has led to establishing close relationships within the panel and between the panel and the research team.

Although we applied a theory-informed approach to implementation, it is possible that, with additional external facilitation and better-resourced internal facilitation, we could have engaged in closer co-design of the implementation support. Furthermore, it may have been beneficial to offer different levels of support for different PDCs working in different contexts. PDCs did differ with respect to levels of support needed and acquired, whether from internal colleagues or external facilitation.

Despite working closely with a diverse range of stakeholders, we did not achieve our ambition to develop an intervention, the implementation of which would be led by the nursing home and adapted to each home’s context. Arguably, we imposed a solution from outside the nursing home, rather than identifying the problem and its potential solution from within.

Testing of the intervention

In testing the intervention, we examined engagement with the tools and views on the implementation support.

What are effective approaches to ensure that NHS and care home staff work collaboratively?

Although our study was conducted within nursing home, more attention needs to be paid to effective collaboration between the NHS and care homes. In this study, we wanted to improve nursing home staff detection, assessment and reporting of changes in residents’ health.

What is the extent of the problem of avoidable admissions from care and nursing homes?

We need to establish hospitalisation rates for all conditions from nursing homes and care homes in the UK. This will enable future studies to target high-admitting homes and the conditions that lead to these high rates of admission. In this study, we focused on ACSCs, yet found relatively low hospitalisation rates for these conditions. This may be because of the lack of routine data collection in nursing homes and under-reporting of acute hospital admissions. New automated methods have recently been developed and may give a more reliable and scalable technique to assess nursing home resident acute hospital resource use. 140 Since we designed this intervention study, the research field has moved forward and routinely collected health and social care data have been used to give insight into the effects of interventions that aim to improve care. For example, the Health Foundation briefing on the Rushcliffe vanguard project46 linked care home data with Hospital Episode Statistics to conduct a matched comparison study between residents who received the enhanced health-care model and similar individuals residing in similar care homes in comparable areas of England. Such data are less subject to the possible recruitment bias found in RCTs and give a ‘real-world’ picture of change. They do, however, have disadvantages. For example, the data are based on aggregate service-level data and do not give information on contextual variables or on what the ‘active ingredient’ of the new intervention may be.

In addition, we need to establish the extent to which care homes problematise detection and recurrent hospital admissions, as the NHS does.

What are the current processes of care for residents who experience deterioration in their health?

We need a better understanding of the processes of care for residents who experience ACSCs that result in hospitalisation and for those who experience ACSCs that do not result in hospitalisation. Bowers et al. 119 used a qualitative grounded theory to develop a conceptual model to explain common and divergent (from the Green House model) care practices. Such an approach could be used to propose a conceptual model to explain differences in care processes between homes with low admission rates and homes with high admission rates in the UK.

Future research would benefit from a breadth of data, including care record reviews, participant or non-participant observation and interviews with key personnel. Further work might also usefully consider the value of nationally standardised record formats for early detection and assessment in nursing homes, which would reduce variation around what is recorded in everyday practice.

What are the important contextual variables to consider in achieving early detection?

Bowers et al. 119 remind us of a host of contextual variables to consider when seeking to establish and embed care processes associated with low rates of hospital admission. These include having opportunities to communicate, and having a culture of taking advantage of such opportunities. Small-scale housing models of nursing home care provide these more easily than large-scale provision. Opportunities to communicate include communication with the full range of stakeholders, including residents, family carers, all levels of nursing home staff and visiting health professionals, including primary care.

The legitimate role and involvement of care assistants is a particularly important contextual variable. Bowers et al. 119 argue that the notion of ‘empowering’ care assistants needs to be tempered with them being accountable to a care team. Future research should, at a minimum, gather data on the primary care context and the role of care assistants in health care in each home. Describing the primary care context would include describing the NHS service delivery model, whether weekly clinics or individual GPs visiting individual residents. Accessing their perspective on the adequacy and completeness of nurse reports of resident ill health would be an important perspective not sufficiently included in the current study.

What resources are needed for internal and external facilitation of change?

We need to identify the resources required to ensure effective support for both external and internal facilitation of change. 61,113 With respect to externally supported facilitation, although previous studies have attempted to determine the optimal dose and mechanism (e.g. Seers et al. 95), further studies need to adequately resource internal–external facilitator relationships. With respect to internal facilitation, research could identify the human resources (personal qualities and professional knowledge and skills regarding facilitation of change), as well as the time required, for leadership in introducing and embedding change in care practice.

Might this intervention be useful to residential care homes?

We found a lower than expected rate of admission to acute hospitals in our nursing home sample. This may be because these are not documented or may reflect very recent evidence that the overall rate of admissions from nursing homes is lower than that found in residential care homes. 141 Our intervention may be more suitable in residential care homes, where there is no on-site nursing care and more potential to reduce emergency care use, something suggested during interviews with nursing home staff. 142

Victor et al. 10 query whether or not residential care home staff have the expertise that residents assume they have: expertise in both noticing changes in health and determining whether or not they warrant medical attention. Residents relied on care home staff (less commonly family) as intermediaries in arranging access to GPs. Almost half expected that care home staff would monitor their health; knew them sufficiently well to notice health changes; and would initiate help from the GP, without prompting by the resident. They conclude that care home staff may need support in fulfilling this ascribed role of day-to-day monitoring of residents’ health. They argue that more attention needs to be paid to this aspect of care home staff training and that we need evidence about the best ways to equip care home staff in this area. We sought to support nursing home staff in detecting and reporting early signs of health deterioration in residents. Some of our interviewees and several of the participants at our dissemination event mentioned this.

During the interviews conducted as part of the process evaluation, the S&W was frequently mentioned. It is possible that using this tool in combination with family involvement and knowledge and skills enhancement may prove useful, particularly in residential care homes. 142 This focus could be enhanced by adopting approaches used in the growing body of research on implementing advance care planning in care homes (e.g. see Kezirian et al. 121). Our intervention, however, was not succesfully implemented in any of the participating nursing homes; therefore, we cannot assume that implementation would be successfull in residential care homes.

What is the cost to care homes and nursing homes of reducing hospital admissions?

Future research incorporating economic models of hospital reduction programmes should measure the cost of care to the care home.

How can we effectively engage in co-production of interventions and implementation strategies with diverse stakeholders?

To optimise impact, we need to facilitate engagement in co-production of applied research with the full range of stakeholders, for example care and nursing homes, GPs and paramedics. 84 This includes assessing care home readiness and capacity to engage, including releasing all levels of staff to engage in intervention design. 143

Without care homes search strategy

1. practice guideline/ (20,201)

2. clinical protocols/ (21,088)

3. algorithms/ (188,266)

4. critical pathways/ (4923)

5. exp guideline/ (26,446)

6. Health Planning Guidelines/ (3891)

7. guideline adherence/ (22,798)

8. (compliance adj (protocol? or policy or guideline?)).tw. (65)

9. (guideline? adj2 (introduc* or issu* or impact or effect? or disseminat* or distribut* or implement*)).tw. (5784)

10. nursing protocol?.tw. (110)

11. professional standard*.tw. (975)

12. (practice guidelin* or practice protocol* or practice parameter or clinical practice guidelin*).tw. (14,792)

13. (consensus development conference or consensus development conference nih).pt. (9785)

14. ((clinical or critical or care) adj3 path).tw. (765)

15. ((care adj map*) or plan*).tw. (677,527)

16. or/1-15 [guidelines] (944,931)

17. hospitalization/ and (avoid* or prevent* or unnecessary).tw. (7908)

18. Health Services Misuse/(3871)

19. 17 or 18 [avoidable hospitalisation] (11,692)

20. ambulatory care/ (36,848)

21. chronic disease/ (224,731)

22. exp pulmonary disease, chronic obstructive/ (38,800)

23. exp heart failure/ (91,251)

24. exp respiratory tract infections/ (299,937)

25. exp urinary tract infections/ (39,673)

26. exp dehydration/ (10,763)

27. (COPD or chronic obstructive pulmonary disease or respiratory infect* or respiratory tract infect* or congest* heart failure or dehydrat* or urinary tract infect* or UTI).mp. (191,968)

28. or/20-27 [ambulatory care sensitive terms] (771,218)

29. 19 or 28 [avoidable hospitalisation or ambulatory care sensitive terms] (780,792)

30. exp aged/ (244,7169)

31. geriatrics/ (26,985)

32. geriatric nursing/ (12,538)

33. (gerontol* or ageing or aging or aged or elder* or geriatric* or seniors or senior citizen* or old age or late* life).mp. (4,385,665)

34. (older adj (person* or people or adult* or patient*)).tw. (77,780)

35. veterans/ (10,477)

36. veteran*.mp. (27,581)

37. or/30-36 [older adults] (4,405,212)

38. 16 and 29 and 37 [guidelines and avoid hosp or ambulatory care sensitive terms and elderly] (9704)

39. limit 38 to (english language and yr=”2000 -Current”) (6709)

* etc. etc.

With care homes search strategy

1. practice guideline/ (20,201)

2. clinical protocols/ (21,088)

3. algorithms/ (188,266)

4. critical pathways/ (4923)

5. exp guideline/ (26,446)

6. Health Planning Guidelines/ (3891)

7. guideline adherence/ (22,798)

8. (compliance adj (protocol? or policy or guideline?)).tw. (65)

9. (guideline? adj2 (introduc* or issu* or impact or effect? or disseminat* or distribut* or implement*)).tw. (5784)

10. nursing protocol?.tw. (110)

11. professional standard*.tw. (975)

12. (practice guidelin* or practice protocol* or practice parameter or clinical practice guidelin*).tw. (14,792)

13. (consensus development conference or consensus development conference nih).pt. (9785)

14. ((clinical or critical or care) adj3 path).tw. (765)

15. ((care adj map*) or plan*).tw. (677,527)

16. or/1-15 [guidelines] (944,931)

17. hospitalization/and (avoid* or prevent* or unnecessary).tw. (7908)

18. Health Services Misuse/ (3871)

19. 17 or 18 [avoidable hospitalisation] (11,692)

20. ambulatory care/ (36,848)

21. chronic disease/ (224,731)

22. exp pulmonary disease, chronic obstructive/ (38,800)

23. exp heart failure/ (91,251)

24. exp respiratory tract infections/ (299,937)

25. exp urinary tract infections/ (39,673)

26. exp dehydration/ (10,763)

27. (COPD or chronic obstructive pulmonary disease or respiratory infect* or respiratory tract infect* or congest* heart failure or dehydrat* or urinary tract infect* or UTI).mp. (191,968)

28. or/20-27 [ambulatory care sensitive terms] (771,218)

29. 19 or 28 [avoidable hospitalisation or ambulatory care sensitive terms] (780,792)

30. Homes for the Aged/ (11,446)

31. exp Nursing Homes/ (32,977)

32. nursing home?.tw. (21,734)

33. or/30-32 [nursing home terms] (43,502)

34. Nursing Care/ (26,994)

35. Rehabilitation Nursing/ (1175)

36. Community Health Nursing/ (18,583)

37. Hospitals, Convalescent/ (270)

38. Rehabilitation Centers/ (6980)

39. Institutionalization/ (4824)

40. Group Homes/ (857)

41. Assisted Living Facilities/ (974)

42. Residential Facilities/ (4769)

43. Long-Term Care/ (22,346)

44. care home?.tw. (1849)

45. ((group or residential) adj home?).tw. (1360)

46. ((residential or long-term or longterm or long-stay) adj5 (care or facility or facilities or ward? or institution)).tw. (22,874)

47. ((sheltered or retirement or residential) adj5 (hous* or home? or accommodation)).tw. (2623)

48. (life care cent*or continuing care cent*or extended care facility or extended care facilities).tw. (209)

49. ((skilled or intermediate) adj2 (nursing facility or nursing facilities)).tw. (1522)

50. assisted living.tw. (1326)

51. or/34-50 [institutional care/nursing care terms] (103,301)

52. exp aged/ (2,447,169)

53. geriatrics/ (26,985)

54. geriatric nursing/ (12,538)

55. (gerontol* or ageing or aging or aged or elder* or geriatric* or seniors or senior citizen* or old age or late* life).tw. (641,501)

56. (older adj (person* or people or adult* or patient* or inpatient*)).tw. (77,861)

57. veterans/ (10,477)

58. or/52-57 [aged terms] (2,809,474)

59. 51 and 58 [institutional care or nursing care and aged terms] (31,907)

60. 33 or 59 [nursing homes or all other institutional care or nursing care and aged terms] (66,135)

61. 16 and 29 and 60 [guidelines and ambulatory care or avoidable admissions or nursing homes or other aged facilities/care] (477)

62. limit 61 to yr=“2000 -Current” (366)

Introduction

Thank you very much for agreeing to act as a key informant for ‘Better Health in Residents in Care Homes’ (BHiRCH). This interview will take approximately 30 minutes to complete.

• Are you happy with the interview being recorded? If not, I can take notes.

• You do not have to answer all the questions if you do not wish to. If you would like to end or take a break from the interview at any time, please let me know.

Context

As you know, our primary aim is to improve the early detection and timely referral of readily treatable health conditions (respiratory infection, exacerbation of CHF, UTI and dehydration) affecting residents in nursing homes.

The project is over a 3-year period and we are currently in year 1. In year 1 we are developing and testing the best approach to reduce avoidable admissions from care homes. In years 2 and 3 we will test the intervention, first of all in two nursing homes, and then in 14 nursing homes. We will be interested to see whether the complex intervention affects rates of avoidable hospital admissions and outcomes of importance to residents, staff and families.

At the moment we are finalising a care pathway that can be used in nursing homes that will involve senior care assistants, nurses in care homes, care home managers, family members and primary care professionals.

We would like your assistance with the next part of our project, which involves developing an optimal approach to enhancing knowledge and skills of nursing home staff, as well as implementation support.

Therefore, I will ask you questions about:

• key knowledge and skills for health care in care homes

• literature you are aware of in this area

• best practice you are aware of nationally and internationally.

Do you have any further questions before we start?

For all

Are you aware of any literature that addresses:

• perspectives on core competencies required for ensuring proactive health care

• effective approaches to enhancing the knowledge and skills of nursing home staff

• existing knowledge and skills enhancement programmes for active management of ACSCs in nursing homes

• optimal approach to implementation support?

Can you give examples of best practice you are aware of nationally and internationally?

• What do you think are the core competencies that are required for nursing home staff to (ensure proactive health care)?

  Prompts: what skills do you think are required for staff to recognise ACSCs? (Heart failure, UTI, respiratory infection, dehydration.)

• What are the current provisions of specific training required for medical professionals (physicians, nurses, geriatricians) who are seeing nursing home residents?

  Prompts: prevention, health gain, health maintenance and palliative care.

• How do you think specific knowledge about older people’s medical conditions could be improved to recognise ACSCs?

• What are the current provisions of specific training in care for older people required for nursing home staff?

• How can competencies be enhanced within the environment of a nursing home?

  Prompts: (e.g. around existing time pressures).

  Use of telecare? Could this be developed to assist communication between staff and primary care? Are there training requirements here?

  How can workload be balanced across GPs, nurses and care assistants?

• What else is needed in addition to medical care to manage the four ACSCs we are concerned with (CHF, respiratory infection, UTI, dehydration)?

  Prompts:

  Access to physiotherapy, occupational therapy, speech and language therapy, pharmacy

  Psychologist, dentist, social work

  Do these systems exist already?

International advisory group members

• What do you think are they key knowledge and skills needed to provide good health care in care homes?

• What do you think are the best ways to get these knowledge and skills?

• What do you think is the best way to ensure early detection of changes in a resident’s condition?

Care home staff

• What do you think you need to know in order to help a resident who is about to become ill?

• How do you get this information? [e.g. telemedicine; types of training, e.g. online; digital versatile disc (DVD); induction; supervision; peers/colleagues; handover; magazines; web].

  • How might you get this information?

• What signs do you notice in a resident about to become ill?

• What skills do you think you need in order to help a resident who is about to become ill?

  • How do you get these skills?

  • How might you get these skills?

Nurses

• What do you think nurses need to know in order to help a resident who is about to become ill?

  • How do they get this information?

  • How do you think they can best get this information?

• What skills do you think nurses need in order to help a resident who is about to become ill?

  • How do they get these skills?

  • How do you think they can best get these skills?

Care assistants

• What do you think care assistants need to know in order to help a resident who is about to become ill?

  • How do they get this information?

  • How do you think they can best get this information?

• What skills do you think care assistants need in order to help a resident who is about to become ill?

  • How do they get these skills?

  • How do you think they can best get these skills?

Domestic staff

• What do you think domestic staff need to know in order to help a resident who is about to become ill?

  • How do they get this information?

  • How do you think they can best get this information?

• What skills do you think domestic staff need in order to help a resident who is about to become ill?

  • How do they get these skills?

  • How do you think they can best get these skills?

Learning outcomes

By the end of the workshop you will be able to:

• Understand and implement the BHiRCH intervention.

• Identify the key staff needed to facilitate the change process, who will go on to form the PDSG.

• Have the skills necessary to negotiate with managers/leaders and the membership of the PDSG.

• Collaborate in identifying changes that need to be made within the care home to embed the BHiRCH intervention.

• Engage with all staff and family members.

These outcomes are worked towards through a focus on specific aims pertaining to the three key elements of facilitation, evidence and context.

Facilitation

• Demonstrate the skills of a facilitator who adopts a person-centred approach.

• Enable members of the PDSG to participate in all aspects of the implementation activity.

Evidence

• Enable the PDSG to understand the BHiRCH Intervention.

• Help other staff to develop their own knowledge and skill in better supporting care home residents in maintaining their health care.

Context

• Enable members of the PDSG to understand the importance of developing:

  • a conducive context, for example collaborative decision-making processes, enabling power and authority processes, sharing resources, effective communication and feedback processes

  • workplace cultures of effectiveness

  • effective leadership and team-working

  • effective care review processes at individual, team and systems levels.

Aim

The aim of the economic evaluation was to evaluate the cost-effectiveness of delivering the BHiRCH-NH intervention, compared with TAU, over 6 months. This is reported as the incremental cost per QALY gained and the probability of cost-effectiveness for a range of values of WTP for a QALY gained.

Outcomes

• The EQ-5D-5L scores at baseline and 6 months [residents’ self-completed questionnaires and carer-perception (about the resident) questionnaires].

• Resource use at baseline (covering the previous month) and monthly for the following 6 months.

The economic analysis was conducted from the NHS and PSS perspective. We calculated the incremental cost per QALY gained over 6 months and the probability of cost-effectiveness for a range of values of WTP for a QALY gained. Analyses followed the Consolidated Health Economic Evaluation Reporting Standards (CHEERS) Good Reporting Practices. 144

Analyses