A systematic review of integrated working between care homes and health care services

Background In the UK there are almost three times as many beds in care homes as in National Health Service (NHS) hospitals. Care homes rely on primary health care for access to medical care and specialist services. Repeated policy documents and government reviews register concern about how health care works with independent providers, and the need to increase the equity, continuity and quality of medical care for care homes. Despite multiple initiatives, it is not known if some approaches to service delivery are more effective in promoting integrated working between the NHS and care homes. This study aims to evaluate the different integrated approaches to health care services supporting older people in care homes, and identify barriers and facilitators to integrated working. Methods A systematic review was conducted using Medline (PubMed), CINAHL, BNI, EMBASE, PsycInfo, DH Data, Kings Fund, Web of Science (WoS incl. SCI, SSCI, HCI) and the Cochrane Library incl. DARE. Studies were included if they evaluated the effectiveness of integrated working between primary health care professionals and care homes, or identified barriers and facilitators to integrated working. Studies were quality assessed; data was extracted on health, service use, cost and process related outcomes. A modified narrative synthesis approach was used to compare and contrast integration using the principles of framework analysis. Results Seventeen studies were included; 10 quantitative studies, two process evaluations, one mixed methods study and four qualitative. The majority were carried out in nursing homes. They were characterised by heterogeneity of topic, interventions, methodology and outcomes. Most quantitative studies reported limited effects of the intervention; there was insufficient information to evaluate cost. Facilitators to integrated working included care home managers' support and protected time for staff training. Studies with the potential for integrated working were longer in duration. Conclusions Despite evidence about what inhibits and facilitates integrated working there was limited evidence about what the outcomes of different approaches to integrated care between health service and care homes might be. The majority of studies only achieved integrated working at the patient level of care and the focus on health service defined problems and outcome measures did not incorporate the priorities of residents or acknowledge the skills of care home staff. There is a need for more research to understand how integrated working is achieved and to test the effect of different approaches on cost, staff satisfaction and resident outcomes.


Background
In the UK care homes are the major provider of long term and intermediate care for older people [1][2][3]. There are 18, 255 care homes providing 459, 448 beds, almost three times as many as the 167, 000 hospital beds available [4]. Although people living in care homes have complex needs and represent the oldest and most frail of the older population in the UK, research consistently demonstrates that they have erratic access to NHS services, particularly those that offer specialist expertise in areas such as dementia and end of life care [5][6][7][8][9].
Inappropriate and unplanned hospital admissions, recognition of unmet health needs, concerns about supporting patient dignity, end of life care and access to health services have triggered multiple care home specific policy initiatives and interventions [10,11]. A consultation event that involved care home and health care representatives identified multiple examples of the NHS working with care homes to improve information exchange, palliative care, reduce falls, and unplanned admissions to hospital [12]. These interventions often involve the introduction of specialist health workers and teams or problem specific workers to achieve the desired outcomes [13,14].
Primary health care services in England spend significant amounts of time providing care for older people resident in these settings [15,16,7,8] (Goodman, C et al: Can clinical benchmarking improve bowel care in care homes for older people? Final report submitted to the DoH Nursing Quality Research Initiative PRP, Centre for Research in Primary and Community Care, University of Hertfordshire, 2007). However, relatively little is known about how health care services work with the (largely unqualified) workforce to provide care to a population that has complex physical and medication needs, experiences high level of cognitive impairment, depression and is in the last few years of life [17,18]. The involvement of health care services in care home settings is often defined by what care home staff are not allowed to do rather than a clear understanding of how the two sectors complement each other, or work together [19]. In addition, it cannot be assumed that health service definitions of problems and services reflect how older people and care home staff define health needs and the types of health care they would like (Evans, C: The analysis of experiences and representations of older people's health in care homes to develop primary care nursing practice, unpublished PhD King's College London, 2008).
Initiatives that support continuity and integration of care for older people with complex needs across health and social care with public and private providers are increasingly recognised as important for continuity and quality of care [20,21]. Integration of service provision can be defined as 'a single system of needs assessment, commissioning and/or service provision that aims to promote alignment and collaboration between the cure and care sectors [22]. There are different levels of integration between health care services [23]. In the context of integrated working with care homes, these can be summarised as:

Patient/Micro level
Close collaboration between different health care professionals and care home staff e.g for the benefit of individual patients.

Organisational/Meso level
Organisational or clinical structures and processes designed to enable teams and/or organisations to work collaboratively towards common goals (e.g. integrated health and social care teams).

Strategic/Macro level
Integration of structures and processes that link organisations and support shared strategic planning and development for example, when health care services jointly fund initiatives in care homes [24,25].
To understand the evidence for the benefits of different approaches to health care services supporting older people in care homes, we conducted a systematic review to identify studies using integrated working between primary health care professionals and care homes for older people; evaluated their impact on the health and well being of older people in care homes, and identified barriers and facilitators to integrated working.

Methods
The review was conducted according to inclusion criteria and methods pre-specified in a protocol developed by the authors before the review began.

Inclusion criteria
We included interventions designed to develop, promote or facilitate integrated working between care home or nursing home staff and health care practitioners. Interventions that involved staff going in to provide education or training to care home/nursing home staff were included as long as there was some description of joint working or collaboration. We excluded studies where staff were employed specifically for the purpose of the research without consideration of how the findings might be integrated into ongoing practice (i.e. project staff introduced for a limited time to deliver a specific intervention). For a study to be included there had to be evidence of at least one of the following: Clear evidence of joint working Joint goals or care planning Joint arrangements covering operational and strategic issues Shared or single management arrangements Joint commissioning at macro and micro levels Studies also had to report at least one of the following outcomes: Health and well being of older people (e.g. changes in health status, quality of life) Service use (e.g. number of GP visits, hospital admissions) Cost such as savings due to avoided hospitalisations Process related outcomes (such as changes in quality of care, increased staff knowledge, uptake of training and education and professional satisfaction) As the literature in this area is limited we included all studies that involved an element of evaluation. This included controlled and uncontrolled studies. However, because they are more susceptible to bias, studies without a control were used to describe and catalogue interventions rather than evaluate effectiveness. Process evaluations and qualitative studies including those using action research methodologies were included in order to identify facilitators and barriers to integrated working.

Identification of studies
The electronic search strategy was conducted in February 2009. We searched the following electronic databases: Medline (PubMed), CINAHL, BNI, EMBASE, PsycInfo, DH Data, Kings Fund, Web of Science (WoS incl. SCI, SSCI, HCI) and the Cochrane Library incl. DARE. In addition, we contacted care home related interest groups and used lateral search techniques, such as checking reference lists of relevant papers, and using the 'Cited by' option on WoS, Google Scholar and Scopus, and the 'Related articles' option on PubMed and WoS. We applied no restrictions by date or country but included English language papers only. Details of the search terms used can be seen in Table 1.

Data extraction and synthesis
Electronic search results were downloaded into EndNote bibliographic software. Two reviewers independently (SD, FB) screened all titles and abstracts of citations identified by the electronic search, applied the selection criteria to potentially relevant papers, and extracted data from included studies using a standardised form. Any disagreements concerning studies to be included were resolved by consensus or by discussion with a third reviewer (CG).
Due to substantial heterogeneity in study design, interventions, participants and outcomes we did not pool studies in a meta-analysis. Instead a narrative summary of findings is presented and where possible we have reported dichotomous outcomes as relative risks (RR) and continuous data as mean differences (MD) (with 95% confidence intervals). Data in the evidence tables is presented with an indication of whether the intervention had a positive effect (+), a negative effect (-), or no statistically significant effect (0). The qualitative studies were used to generate a list of potential barriers and facilitators to integrated working. Each paper was systematically read by two researchers (SD, CV) to highlight any factors that may have impacted on the process, both those that were explicitly referred to by the authors and those identified by the reviewers within the papers' narratives.
The quality of the included studies was assessed using design assessment checklists informed by the Cochrane Collaboration risk of bias tool [26] and Spencer et al's quality assessment checklist for qualitative studies [27]. The core quality-assessment domains are summarised in Table 2. As other non controlled studies were used to inform contextual understanding rather than evaluate effectiveness they were not formally quality assessed.
Data were extracted from each study on methodology, type of intervention, outcomes, participants, and location. In addition, an interpretive approach based on Kodner and Spreeuwenberg's (2002) work on integrated working, was used to compare and contrast the nature and level of integration across the studies using the principles of framework analysis [28]. Each study was categorised in terms of the degree of integration and the complexity classified as micro, meso and or macro. In addition, based on the assumption that care homes with a higher level of integration would show evidence of correspondingly greater levels of support and contact with health care professionals, each study was analysed to identify the amount of contact, support and training given by the health professionals involved in the study. Figure 1 shows the flow of studies through the selection process. Seventeen studies (reported in 18 papers) met our inclusion criteria.

Description of studies
Ten studies were quantitative, (four of which were RCTs), one used mixed methods, two were process evaluations, three were qualitative and one was action research (see Table 3).
Nine were conducted in the UK, five in Australia, two in the USA and one in Sweden. Eleven (65%) studies were conducted in nursing homes, five in residential homes and one in a combination of both. Study participants included residents, relatives, care home staff both residential and nursing, and health professionals including general practitioners, district nurses, nurse specialists, pharmacists, psychiatrists and psychologists.
Seven studies were focused on individual care, for example, specific health care needs such as end of life [29][30][31][32][33] or wound care [34] and dementia [35]. Six studies focused on residents' needs as a group, such as detection and treatment of depression [36], bowel related problems (Goodman, C.  [37] and improved prescribing [38][39][40]. A further four papers were service evaluations such as an in-reach

Sequence generation
Was the allocation sequence adequately generated?
Allocation concealment Was allocation adequately concealed?

Blinding
Was knowledge of the allocation intervention adequately concealed from outcome assessors?
Incomplete outcome data-Was this adequately addressed for each outcome?

Selective outcome reporting
Are reports of the study free of suggestion of selective outcome reporting?
Controlled studies (without randomisation) all scored as Yes/No/Unclear Baseline results reported Were baseline results reported for each group?

Groups balanced at baseline
Were there any significant differences in the groups at baseline?

Blinding
Was knowledge of the allocation intervention adequately concealed from outcome assessors?
Incomplete outcome data-Was this adequately addressed for each outcome?

Selective outcome reporting
Are reports of the study free of suggestion of selective outcome reporting?

Risk of bias
There were seven controlled studies of which four were RCTs. Although the RCTs could be expected to be less susceptible to bias than the non randomised studies the     Based on an assumed improvement in the healing rate from 15% to 30%, 108 wounds per arm were required to have an 80% chance of detecting a twofold increase in healing rates at a significance level of 5%.
To adjust for clustering this number was increased to 151 in each group.
Residents in the intervention arm received standardised treatment from a wound care team comprised of trained community pharmacists and nurses. A standard treatment protocol was developed based on the colour, depth and exudate method for assessing wounds and the group's clinical and academic experience. They met weekly to discuss any new wounds and treatment options within the protocol. Both nurses and pharmacists received training on wound healing and management.
Treatment recommendations, frequency and detail of dressing changes, measurement and photos of wounds, SF36, Assessment of Quality of Life index, Brief Pain Inventory -measures wound pain, total estimated cost of treatment per wound including, staff time, training, wound care products and waste disposal.
During the trial more wounds healed in the intervention than in the control group but this was not significant. The mean treatment cost of wound healing was significantly less in the intervention group. Standardised treatment by a multidisciplinary wound care team cut costs and improved chronic wound healing in nursing homes.

Crotty 2004
An outreach geriatric medication advisory service in residential aged care: a randomised controlled trial of case conferencing.

Cluster RCT
Evaluate the impact of multidisciplinary case conferences on the appropriateness of medications and on patient behaviours in residential care Population: residents with medication problems/ challenging behaviours Setting: 10 Highlevel aged care facilities Country: Australia 154 residents recruited with 54 in control, 50 in intervention, 50 in within facility control group 5 facilities randomised to the intervention and 5 to the control Staff nominated 20 residents for the intervention and 10 for the control, based on 2 criteria: Residents with a difficult behaviour they would like advice on, those prescribed 5 + medications An effect size based on patients aged 65 + with polypharmacy of 0.9 in the MAI between the intervention and control groups (power 0.9, type 1 error of 0.05) would be detected with 28 residents in each group 2 multidisciplinary case conferences chaired by the resident's GP, a geriatrician, pharmacist and residential care staff held at the nursing home for each resident. All facilities received a half day workshop on using the toolkit for challenging behaviour All residents had their medication chart reviewed pre and post intervention by an independent pharmacist using the MAI    Interviews explored participant's understanding of the project, their perceptions of issued involved in providing palliative care, benefits, limitations for staff and residents. Questionnaires were used to rate project performance, access, response time, liaison, benefits and limitations of the project. Services were also rated in order of their importance for care homes and residents.
The project helped to overcome the barriers to care between NHS services and the independent sector. Care home isolation was decreased through assistance with individual care and better access to specialist advice and training.  potential for bias in both groups of studies appeared to be high (see Tables 4 and 5).
A number of the studies appeared underpowered and for many follow up was short. The qualitative studies employed a range of methodologies including action research, interviews, focus groups and questionnaires. As with the quantitative studies, the quality was low, only two out of four [30,33] had a clearly defined purpose and design. With one exception [33] descriptions of the study sample, data collection and analysis were inadequate and evidence of their credibility and transferability was limited (see Table 6).

Effectiveness
The heterogeneity of outcomes and, in particular, the interventions meant that making comparisons between studies was problematic. Three studies looked at the effect on prescribing [38][39][40], three included mortality as an outcome [39,40,44] and two looked at disruptive behaviour [35,39]. The remaining outcomes, only included in single studies, were depression [36], hospital admissions [40], functional status [40], wound healing [34], and bowel related problems (Goodman, C et Table 7. Although there were some improvements in outcomes, the majority of studies showed that the intervention had either mixed effects (that is improvement in one outcome but no effect or negative effect in another outcome), or no effect when compared with the control group. Insufficient information was available to evaluate the cost of integrated working between care homes and primary health care professionals.

The nature of integrated working
There was a great deal of variation in how health care services and care homes worked together and the frequency of contact. For example, whilst some studies involved weekly multidisciplinary team meetings [43], monthly meetings were more common (Goodman, [30]. All the studies potentially increased care home staff access to health care professional's support and advice, with 15 out of 17 involving care home staff in multidisciplinary interventions or joint working. Care home staff were involved in multidisciplinary meetings and in some studies their opinions were sought [40], but they were led by health care professionals, with health care orientated and defined goals. Staff training was an integral part of all studies bar three; only a few studies consulted with care home staff on their perceived training needs [29,33]. The range of training input varied from as little as three hours [31] to seven seminars [37] or continuous training and support [43,44]. The level of integration for all studies and the degree of support and training provided by NHS staff for care home is reported in Table 8. The majority of studies showed micro integration at the clinical level involving close collaboration between care home staff and health care professionals to achieve specific outcomes (12 out of the 17) e.g. wound care techniques and wound healing. The remaining five studies were integrated at the clinical level but also showed greater complexity of integration in terms of funding and organisation or strategy, one at the meso level [42] and four at the macro level [31,41,43,44]. In service delivery, four studies used dedicated multidisciplinary teams to support staff and residents in care homes [42], three of which achieved their remit of avoiding unnecessary hospitalisation [41,43,44]. Two UK studies also had health service funded beds within care homes, one for use by a specialist health care nursing team [41] the other to provide end of life care [31]. A distinguishing feature of four out of the five studies classified at higher levels of integration was that care home staff received support and or training which was ongoing, as opposed to being offered at discrete time periods during the intervention. For example, nursing home staff were facilitated to recognise and manage acute conditions [43], to improve residents' overall care [44]. A number of cross cutting themes that influenced the achievement of integrated working were identified (See Tables 9 and 10). These included, care home access to services and the different working cultures of care home staff and health care professionals that acted as barriers and facilitators. Care home staff identified a lack of support from health care professionals and a failure to recognise their knowledge and skills [29,33,42]. There were negative perceptions on both sides with care home staff feeling that health care professionals were sometimes acting in a 'policing' rather than an advisory capacity [29,42] and health care professionals perceiving care home staff as lacking in knowledge and expertise, and unwilling to change their practice [30].
Whilst input and training from health care staff was valued, for care home staff to access it, dedicated time and finance from care home managers was necessary. Holding sessions within the care home and setting up a learning contract with the staff could facilitate training [32]. Examples of positive interactions included one care home support team described as acting as a link to 'the outside world' by the care home, and supporting clinical decision making across the multi disciplinary team [42]. Difficulty in maintaining levels of staff skills and knowledge were exacerbated by the high staff turnover experienced by care homes [29,32,33]. However, one study found a higher rate of staff turnover amongst the health care professionals involved in the intervention than the senior staff in the care homes (

Discussion
We found 17 studies, eight of which were controlled evaluations. Although some of the studies reported positive outcomes most interventions had mixed or no effects when compared with the control group. There was insufficient information available to evaluate the cost of integrated working between care homes and primary health care professionals. Some of the qualitative studies suggested that integrated working had the potential to improve the quality of life for older people in care homes through increased support for care home staff and increased access to health care services. A small number of studies which were integrated at the macro or meso level, involved care homes that were supported by dedicated health service teams and health service funded beds or managed care, showed more positive outcomes such as avoidance of hospitalisation. They also differed from the micro integrated studies in their capacity to give ongoing support and training for care home staff, which had the potential to address one of the main identified barriers to integrated working and ultimately improve resident's care. This indicates that Table 5 Non randomised controlled studies quality assessment results

Study
Baseline results reported?
Groups balanced at baseline?

Blinding of outcome assessment
Incomplete outcome data assessed?
Free from selective reporting?   Evercare rate significantly less than for control-in group but was slightly higher than control-out group (non significant)

Preventable hospitalizations
Rates of preventable admissions lower in Evercare than for either control but only significant when compared to control-out.
No differences in hospitalization rates overall. (0) for integrated working to be successful, formal structures may need to be in place for health service delivery and organisation of care for care homes.
Despite the lack of evidence on effectiveness, studies consistently demonstrated key issues that supported or militated against integrated working. These findings are significant for future research and the development of interventions that rely on integrated working between health care services and care home staff. Barriers to integrated working included a failure to acknowledge the expertise of care home staff, their lack of access to health care services, as well as high care home staff turnover and limited availability of training. Facilitators to integrated working were the care home manager's support for the intervention, protected time and the inclusion of all levels of care home staff for training and support by health care professionals.
A common feature of the interventions was the use of multidisciplinary teams to improve one or more aspect of older people's health care. However, all the studies were led and conducted by health care professionals. There was no evidence of care home staff being involved in the definition or focus of the studies and some evidence that care home staff felt that their knowledge and views were not valued. Seven studies employed external project staff in some capacity, which implies that integrated working may require some external facilitation.
Three studies used integrated care pathways as a means of improving the quality of end of life care for older people resident in care homes. Care pathways may increase integrated working for the individual older people who have them, but this will not necessarily extend to the care home residents as a whole. The use of a shared assessment and care framework and

Limitations of existing evidence
Given the limited number of studies in the review, their heterogeneity, poor quality, small size, and low level of detail, the scope for discussion of integrated working between care homes and primary health care professionals is limited and firm conclusions cannot be reached. Only five studies were conducted in residential care homes which reinforced previous findings that the majority of research is carried out in nursing homes, even though this is not where most older people in long term care live [14]. The absence of older people's views and resident centred outcomes from the studies was notable.
Moreover the majority of studies were only integrated at the micro level that is, close collaboration between care home staff and professionals, so little information was available on the impact of integration at meso and macro levels. There was wide variation amongst the studies in terms of their level of care home staff support and training, and the involvement of older people. Care home staff training and support ranged between those studies where it was ongoing and those where it was provided only on one occasion. Where there was support and training of care home staff it was not clear if the ultimate aim was to train staff to a level of expertise so that health services could withdraw.

Implications for research
There is a need for more research that addresses how integrated working can best be achieved and that evaluates the effect of integrated working on the health and wellbeing of older people, service use and cost. Research with care homes should reflect the context and constraints of working across public and independent services, and involve care homes in the planning and design of interventions. Moreover as this population is known to have multiple co-morbidities that are often compounded by cognitive impairment there is a need for more studies to look at improving the quality of care for the care home population as a whole. Future evaluations should be large enough to detect a difference and outcomes need to be meaningful to care home staff and residents.

Strengths and limitations of the review
We used systematic and rigorous methods to synthesise the current evidence on integrated working between care homes and health care services and highlight areas for further research. There are, however, a number of methodological issues that could have a bearing on the validity of the results. Owing to a lack of evidence in this area we included all studies types including uncontrolled studies. Only four of our included studies were randomised controlled trials. Whilst uncontrolled studies might be more likely to be biased these broad inclusion criteria enabled us to investigate integrated working more widely and identify barriers and facilitators. Although the studies reviewed were judged to have involved integrated working, it was not their main focus; only two studies referred to partnership working between care homes and health care services (Goodman, C et [44]. The information on integrated working was based on how the intervention was described, who was involved and at what level. It is possible that how this was reported in the studies reviewed did not capture the extent of the integration achieved.

Conclusions
Integrated working aims to ensure continuity of care, reduce duplication and fragmentation of services and places the patient as the focus for service delivery. This review identified a limited number of studies where the intervention supported integrated working between care homes and primary health care professionals. The narrow focus and single issue orientation of the majority of the studies did not engage with the needs of care home population or the context and organisation of their care. Outcome measures reflected the priorities of health care professionals rather than residents and care home staff. In view of the growing demand for residential and nursing home care together with funding constraints, more effective working between the NHS and care home providers is essential. There is an urgent need to develop and test interventions that promote integrated working and address the persistent divide between health services and independent providers.

Funding
This research was supported by the National Institute for Health Research Service Delivery and Organisation programme (project number 08/1809/231).