RESEARCH ARTICLE Open Access

A systematic review of integrated working between care homes and health care services Sue L Davies1, Claire Goodman1*, Frances Bunn1, Christina Victor2, Angela Dickinson1, Steve Iliffe3, Heather Gage4, Wendy Martin2 and Katherine Froggatt5

Abstract

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 [1-3]. There are 18, 255 care homes providing 459, 448 beds, almost three times as many as the 167, 000 hospital beds avail- able [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 ser- vices, particularly those that offer specialist expertise in areas such as dementia and end of life care [5-9]. Inappropriate and unplanned hospital admissions,

recognition of unmet health needs, concerns about sup- porting patient dignity, end of life care and access to health services have triggered multiple care home speci- fic policy initiatives and interventions [10,11]. A

* Correspondence: [email protected] 1Centre for Research in Primary and Community Care, University of Hertfordshire, Hatfield, AL10 9AB, UK Full list of author information is available at the end of the article

Davies et al. BMC Health Services Research 2011, 11:320 http://www.biomedcentral.com/1472-6963/11/320

© 2011 Davies et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

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 signifi-

cant 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, Univer- sity 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 represen- tations 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 inte- gration 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 profes- sionals and care home staff e.g for the benefit of indivi- dual 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 organi- sations and support shared strategic planning and devel- opment for example, when health care services jointly fund initiatives in care homes [24,25]. To understand the evidence for the benefits of differ-

ent approaches to health care services supporting older people in care homes, we conducted a systematic review to identify studies using integrated working between pri- mary 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 bar- riers and facilitators to integrated working.

Methods The review was conducted according to inclusion cri- teria 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. Inter- ventions that involved staff going in to provide educa- tion 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

Davies et al. BMC Health Services Research 2011, 11:320 http://www.biomedcentral.com/1472-6963/11/320

Page 2 of 21

included controlled and uncontrolled studies. However, because they are more susceptible to bias, studies with- out a control were used to describe and catalogue inter- ventions 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 Febru- ary 2009. We searched the following electronic data- bases: 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 Sco- pus, 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, inter-

ventions, 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 sta- tistically significant effect (0). The qualitative studies were used to generate a list of potential barriers and facilitators to integrated working. Each paper was sys- tematically read by two researchers (SD, CV) to high- light 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

Table 1 Search terms on PubMed (search terms were suitably adapted for other databases)

Component 1

Search “Delivery of Health Care, Integrated"[Mesh] OR integrated[ti] OR team[ti] OR interdisciplinary[ti] OR integration[ti] OR integral[ti] OR integrat*[ti] OR seamless[ti] OR continuity[ti] OR interface[ti] OR multidisciplinary[ti] OR multiprofessional[ti] OR multiagency[ti] OR interprofessional [ti] OR multi sector[ti] OR model*[ti] OR coordinat*[ti] OR partnership*[ti] OR tufh OR continu*[ti] OR interagenc*[ti] OR stakeholder*[ti] OR network*[ti] OR systems [ti] OR team*[ti] OR shared[ti] OR joined-up[ti] OR pooling[ti] OR vertical*[ti] OR horizontal*[ti] OR collaborat*[ti] OR cross organi*[ti] OR multi- professional[ti] or intermediate care[ti] or multi agency[ti] or multiagency[ti] OR managed care[ti] OR joint care[ti] OR ((individual[ti] or separate[ti]) AND budget) OR partner*[ti] OR all-inclusive[ti] OR in-reach[ti] OR chain[ti] OR comprehensive[ti] or total care[ti] OR interface[ti] OR “service interaction” OR seamless[ti] OR interagency[ti] OR “Patient Care Team"[MAJR]

AND

Search Family Physicians OR general pract*[ti] OR general physician*[ti] OR family doctor*[ti] OR general medicine[ti] OR Primary Health Care OR Continuity of Patient Care OR “primary care” OR continuity of care OR physician*[ti] OR “Physicians"[Majr:NoExp] OR “Physicians, Family"[Majr] OR “Physician Assistants"[MeSH Terms] OR"Nurse Practitioners"[MeSH Terms] OR “Physician’s Practice Patterns"[MAJR] OR physician*[ti] or practitioner*[ti]

AND

Search Nursing Homes OR nursing home*[ti] OR “nursing home*” OR long-term care[ti] OR long term care [ti] OR nursing facilit*[ti] OR residential[ti] OR institutional care[ti] OR resident*[ti] OR continuing [ti] OR respite care OR nightingale home OR nightingale homes OR care home*[ti] OR long- term[ti] OR longterm[ti]

AND

Search geriatrics OR elderly OR older OR middle age OR middle-age OR senior OR frail OR care of elderly OR geriatric nursing OR geriatric assessment OR “Aged"[Mesh] OR “Health Services for the Aged"[Mesh] OR “Middle Aged"[Mesh] OR “Homes for the Aged"[Mesh] OR “Aged, 80 and over"[Mesh] OR senior*[ti] or pensioner*[ti] OR retire*[ti]

Component 2: Simplified, focused searches involving two aspects of the subject: NHS/Primary Care/Nursing homes

Search ("Physicians"[Majr:NoExp] OR “Physicians, Family"[Majr] OR “Physician Assistants"[MeSH Terms] OR"Nurse Practitioners"[MeSH Terms] OR “Physician’s Practice Patterns"[MAJR] OR physician*[ti] OR practitioner*[ti] OR specialist*[ti] OR primary care[ti]) (nursing home*[ti] OR residential care [ti] OR care home*[ti] OR residential home*[ti])

Nursing homes/Integrated Care

Search (nursing home*[ti OR residential care[ti] OR care home*[ti] OR residential home*[ti]) (integrat*[ti] or team*[ti] or cooperation[ti] OR multidisciplinary[ti])

Elderly/Integrated Care

Search (elderly[ti] or older[ti] or geriatric*[ti] OR senior[ti]) (integrat*[ti] OR team*[ti]) AND (community OR nursing homes)

Davies et al. BMC Health Services Research 2011, 11:320 http://www.biomedcentral.com/1472-6963/11/320

Page 3 of 21

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 loca- tion. 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.

Results 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 eva- luations, 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 partici- pants included residents, relatives, care home staff both residential and nursing, and health professionals includ- ing general practitioners, district nurses, nurse specia- lists, pharmacists, psychiatrists and psychologists. Seven studies were focused on individual care, for exam-

ple, specific health care needs such as end of life [29-33] or wound care [34] and dementia [35]. Six studies focused on residents’ needs as a group, such as detection and treat- ment of depression [36], bowel related problems (Good- man, C. et al: Can clinical benchmarking improve bowel care in care homes for older people? Final report sub- mitted to the DoH Nursing Quality Research Initiative PRP, Centre for Research in Primary and Community Care, University of Hertfordshire, 2007.) and or supporting the care home staff interactions with residents through training [37] and improved prescribing [38-40]. A further four papers were service evaluations such as an in-reach

Table 2 Quality assessment criteria by study type

Randomised controlled trials all scored as Yes/No/Unclear

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?

Qualitative studies – Scored as fully or mostly, partly or not at all

Scope and purpose e.g. clearly stated question, clear outline of theoretical framework

Design e.g. discussion of why particular approach/methods chosen

Sample e.g. adequate description of sample used and how sample identified and recruited

Data collection e.g. systematic documentation of tools/guides/researcher role, recording methods explicit

Analysis e.g. documentation of analytic tools/methods used, evidence of rigorous/systematic analysis

Reliability and validity e.g. presentation of original data, how categories/concepts/themes developed and were they checked by more than one author, interpretation, how theories developed

Generalisability e.g. sufficient evidence for generalisability or limits made clear by author

Credibility/plausibility e.g. provides evidence that resonates with other knowledge, results/conclusions supported by evidence

Davies et al. BMC Health Services Research 2011, 11:320 http://www.biomedcentral.com/1472-6963/11/320

Page 4 of 21

team for care homes [41], a care home support team [42], and nurse practitioners [43,44]. End of life care accounted for five papers [29-33], three of which focused on care pathways [30-32].

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

From: Moher D, Liberati A, Tetzlaff J, Altman DG, The PRISMA Group (2009). Preferred Reporting Items for Systematic Reviews and Meta- Analyses: The PRISMA Statement. PLoS Med 6(6): e1000097. doi:10.1371/journal.pmed1000097

• • • • •

• • • • •

Figure 1 PRISMA Flow Diagram. Systematic review process from electronic searching to study inclusion.

Davies et al. BMC Health Services Research 2011, 11:320 http://www.biomedcentral.com/1472-6963/11/320

Page 5 of 21

Table 3 Studies included in the systematic review of integrated working between care homes and health care services:

First Author, Year Title Study design

Research Question/aims and objectives

Study population, setting and country of study

Sample size/number of participants: Include power calculation if available

Description of intervention/ Study design

Main outcome variable(s)/ Areas of focus for qualitative studies

Main findings/ Conclusions

1. King, 2001 Multidisciplinary case conference reviews: improving outcomes for nursing home residents, carers and health professionals Controlled study

To determine whether multidisciplinary case conference reviews improved outcomes for nursing home residents and its impact on care staff.

Population: Older people in nursing homes Setting: 3 nursing homes Country: Australia

245 older people But only 75 residents were reviewed

Weekly case conference reviews, one review per resident, over 8 months attended by GPs, clinical pharmacist, senior nursing staff and other health professionals. Multidisciplinary discussion of all aspects of a resident’s care to make recommendations and devise a management plan for the resident. Reviews were led by GPs with data collection by the pharmacist. Baseline and endpoint comparisons were made between residents who were reviewed and those who were not.

Resident outcomes included: medication use, administered medications and weekly cost, health status and quality of life. Carer outcomes were based on resident interaction, workload or personal/ professional satisfaction.

• There were no significant reductions in medications orders, cost and mortality. 40% of the recommendations benefited residents, measured through their health status and quality of life. 26% of the recommendations benefited care staff, but no details were given. Multidisciplinary case conferences were seen as beneficial to patients and carers. Their future use was recommended.

2. Llewellyn-Jones, 1999 Multifaceted shared care intervention for late life depression in residential care: randomised controlled trial. RCT

To evaluate the effectiveness of a population based multifaceted shared care intervention for late life depression in residential care.

Population: Older people 65 years + with depression and no or low cognitive impairment Setting: Residential facility living in self care units and hostels not nursing homes (equivalent to residential care in UK) Residents were stratified and randomised to intervention or control Country: Australia

220 older people No power calculation

The shared care intervention included: 1. Multidisciplinary consultation and collaboration 2. Training of gps and carers in detection and management of depression 3. Depression related health education and activity programmes for residents. The control group received routine care.

Geriatric Depression Scale There was a significant reduction in adjusted depression scores for residents in the intervention group. Multidisciplinary collaboration, staff education, health education and activity programmes can improve depression in older people in residential care.

D avies

et a l. B M C H ea lth

Services R esea

rch 2 0 1 1 , 1 1 :3 2 0

h ttp

://w w w .b io m ed

cen tral.co

m /1 4 7 2-6

9 6 3 /1 1 /3 2 0

P ag

e 6 o f 2 1

Table 3 Studies included in the systematic review of integrated working between care homes and health care services: (Continued)

3. Opie, 2002 Challenging behaviours in nursing home residents with dementia: a randomised controlled trial of multidisciplinary interventions. RCT

To test whether individually tailored psychosocial, nursing and medical interventions to nursing home residents with dementia will reduce the frequency and severity of behavioural symptoms.

Population; Nursing home residents with severe dementia rated by staff as having frequent, severe behavioural disturbances. Setting: 42 Nursing homes Country: Australia

102 older people entered the study, (99 completed the 4 week trial, 2 RIPs 1 hospitalisation)

Residents selected on basis of CMAI scores and assigned to early or late intervention groups. Consultancy team with training in psychiatry, psychology and nursing met weekly for 30 minutes, to discuss referrals and formulate individualised care plans which were presented to nursing home staff to implement. Plans were reviewed at one week. 3 categories: medical, based on medication review, nursing, based on ADLs, and psychosocial including environment, sensory stimulation. The control was normal care, residents acted as their own controls by being in the early or late intervention groups.

Frequency and severity of disruptive behaviours and assessment of change by senior nursing staff. Tools included: Cohen-Mansfield Agitation Inventory (CMAI) which assesses frequency of 30 behaviours over previous 14 days Behaviour Assessment Graphical System (BAGS) which records a combined frequency and disruption score every hour for 24 hours.

There was a slight reduction in the daily observed counts of challenging behaviours. Individualised, multidisciplinary interventions appear to reduce the frequency and severity of challenging behaviours in nursing homes

4. Schmidt, 1998 The Impact of Regular Multidisciplinary Team Interventions on Psychotropic Prescribing in Swedish Nursing Homes RCT

To evaluate the impact of regular multidisciplinary team interventions on the quantity and quality of psychotropic drug prescribing in nursing homes Aim was to improve prescribing through better teamwork amongst physicians, pharmacists, nurses and nursing assistants

Population: Long term residents, 42% dementia, 5% psychotic disorder, 7% depression Setting: 33 Nursing homes Country: Sweden

1854 residents In 15 experimental homes and 18 control homes

Regular multidisciplinary team meetings over 12 months to discuss individual residents drug use. Training was provided for pharmacists but not for other staff. Control homes provided normal care.

Baseline and 12 month post resident medications

After 12 months the intervention group showed an improvement in the prescribing of hypnotics only. Prescribing practices can be improved through better teamwork between health care and nursing home staff using clinical guidelines.

D avies

et a l. B M C H ea lth

Services R esea

rch 2 0 1 1 , 1 1 :3 2 0

h ttp

://w w w .b io m ed

cen tral.co

m /1 4 7 2-6

9 6 3 /1 1 /3 2 0

P ag

e 7 o f 2 1

Table 3 Studies included in the systematic review of integrated working between care homes and health care services: (Continued)

5. Vu, 2007 Cost-effectiveness of multidisciplinary wound care in nursing homes: a pseudo-randomized pragmatic cluster trial Pseudo RCT

Trial to test the hypothesis that trained pharmacists and nurses working in collaboration with a wound treatment protocol would improve the wound healing and save costs.

Population: 176 residents with leg or pressure wounds Setting: 44 high care nursing homes Country: Australia

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 two- fold 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.

6. 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 High- level 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 reside