Reducing unplanned hospital admissions from care homes: a systematic review

Article history

The research reported in this issue of the journal was funded by the HSDR programme or one of its preceding programmes as project number NIHR133884. The contractual start date was in November 2021. The final report began editorial review in November 2022 and was accepted for publication in March 2023. The authors have been wholly responsible for all data collection, analysis and interpretation, and for writing up their work. The HSDR editors and production house have tried to ensure the accuracy of the authors’ report and would like to thank the reviewers for their constructive comments on the final report document. However, they do not accept liability for damages or losses arising from material published in this report.

Copyright statement

Copyright © 2023 Chambers et al. This work was produced by Chambers et al. under the terms of a commissioning contract issued by the Secretary of State for Health and Social Care. This is an Open Access publication distributed under the terms of the Creative Commons Attribution CC BY 4.0 licence, which permits unrestricted use, distribution, reproduction and adaptation in any medium and for any purpose provided that it is properly attributed. See: https://creativecommons.org/licenses/by/4.0/. For attribution the title, original author(s), the publication source – NIHR Journals Library, and the DOI of the publication must be cited.

2023 Chambers et al.

Patient and public involvement

The project was planned to incorporate public involvement led by a public co-applicant (Lynne Wright, who replaced Cynthia Atkin) supported by regular meetings of a public advisory group. The strategic public advisory group of the Sheffield Health and Social Care Delivery Research Programme (HS and DR) Evidence Synthesis Centre also discussed the review at an early stage and advised on recruitment of the public advisory group.

The topic-specific public advisory group and review team had four one-hour meetings, covering:

1. introductions and discussion of the project context, the methods used (including the public involvement plan and offer of optional training on evidence synthesis methods) and the objectives

2. discussion of ‘emerging’ findings

3. further discussion of review findings

4. focus on dissemination of the review findings and production of the final report (including plain language summary and reporting public involvement).

Members of the group contributed to the review by

• highlighting issues from their own experience, primarily related to how services work in practice, for example inadequate needs assessment of people with complex problems leading to placement in a care home that is unable to meet their needs and hence increasing risk of unplanned hospital admission. Members also discussed examples of paramedic assessment and decision-making about taking people to hospital, and raised awareness of pressures on some care home managers to adopt a ‘risk averse’ approach to managing residents’ health

• providing details of potentially relevant research based on involvement in other research projects, specifically a tool for identifying people at risk of deterioration

• suggesting channels for disseminating review findings outside academia, for example charities such as Age UK, local carers’ groups (with links to both the NHS and local authorities), the Care Quality Commission and possibly Parliament (Health and Care Select Committee or All-Party Parliamentary Groups)

• suggesting that many carers were unlikely to access material on academic websites even when this was labelled as being in plain language. Short printed summaries were seen as being more suitable for decision-makers than for carers, care home residents and the general public.

We also made contact with an existing group based in Leicester with an interest in older people’s health and care and we hope to work with them on the dissemination of the research findings.

We encountered a number of challenges:

• The public co-applicant was unable to participate as planned for personal reasons. A member of the patient and public involvement (PPI) group, Lynne Wright, took on the role and attended meetings with the research team and clinical experts as well as contributing to the report.

• Despite our best efforts, we were only able to recruit three active members for the topic-specific PPI group. This was obviously not ideal, although the members were knowledgeable and enthusiastic, represented different geographical areas and had diverse experiences of the care home sector.

• We were unable to attend a meeting until late in the project with the existing PPI group based in Leicester but the group was interested in the project and we hope to work with them on dissemination of the review findings.

Equity, diversity and inclusion

As a systematic literature review, this project did not involve care home residents, family carers or members of the public as research participants. The research team included a public co-applicant (Lynne Wright, who replaced Cynthia Atkin) with extensive experience of the care home sector as carer for her late husband (see Patient and public involvement). The strategic and topic-specific public advisory groups included people from a diverse range of backgrounds (age, ethnicity and place of residence), although those involved in the topic-specific group tended to be older. We also discussed the project with a Leicester-based PPI group with a specific focus on health of older people (see Patient and public involvement).

Direct involvement of care home residents with appropriate capacity and interest would have made the project more inclusive but would have required considerable time and resources. The project was a systematic review, which meant that most of the research team had backgrounds in this field and lacked informal links with the care home sector. This probably acted as a barrier to including funding for this type of work in the research proposal.

The academic members of the research team were all highly experienced in their respective fields, reflecting the knowledge and experience required to gain funding and deliver the research. The overall team was balanced in terms of gender. The project was a development opportunity for the principal investigator (a late entrant to academic research) who undertook this role on a separately commissioned project for the first time.

Reflections on equity, diversity and inclusion arising from the review findings are presented below (see Chapter 5).

Review questions

The overall research questions were:

1. What interventions are used in the UK health and social care system to minimise unplanned hospital admissions of care home residents?

2. What candidate interventions used in other applicable settings could potentially be used in the UK?

3. What can we learn from research studies and ‘real-world’ evaluations about the effects of such interventions on admissions?

4. What is known about the feasibility of implementing such interventions in routine practice and their acceptability to care home residents, their families and staff?

5. What is known about the costs and value for money associated with these interventions?

Identification of evidence

A broad search was conducted to identify published and peer-reviewed literature on interventions to reduce unplanned admissions from care homes in the UK and other high-income countries. Additionally, a search was undertaken to retrieve relevant grey literature.

The search strategy was developed on MEDLINE and then agreed with the research team (see Appendix 1). The search includes thesaurus and free-text terms and relevant synonyms for the population (residents in care homes for older people) and intervention (interventions to reduce unplanned admissions and named interventions) and makes use of proximity operators where appropriate and the different terms for each concept were combined using the Boolean operator OR. Population and intervention search terms were then combined using the Boolean operator AND. Outcome terms were not included in the search as information on outcomes is not always included in title or abstracts, so including these terms could mean that relevant studies would potentially not be retrieved. The search was limited to research published in English from 2014 to December 2021 to reflect developments since the previous review. Methodological search filters were not applied to keep the search broad and to ensure that all relevant study types were retrieved. However, an attempt was made to remove non-empirical research using the Boolean operator NOT for letters, editorials, news, historical articles, comments and case reports. Additionally, to ensure that studies retrieved were on humans not animals, the Boolean operator NOT was used to remove terms likely to be in studies on animals not humans. The National Institute for Health and Care Excellence filter for Organisation for Economic Co-operation and Development countries was used to aid retrieval of studies from UK and other high-income countries. 9

Once the MEDLINE search had been agreed it was translated to the other major medical and health-related bibliographic databases in December 2021.

The following databases were searched:

• Cumulative Index to Nursing and Allied Health Literature

• EMBASE

• PsycINFO^® (American Psychological Association, Washington, DC, USA)

• Cochrane Central Register of Controlled Trials and Cochrane Database of Systematic Reviews

• Science and Social Sciences Citation Indexes

• Health Management Information Consortium

• Social Care Online

• Social Service Abstracts.

Following the main search, an extra focused search was conducted to identify studies investigating interventions to reduce falls in care homes in January 2022. The search used the MeSH term Accidental Falls/pc (Prevention and Control) and free-text terms, and was then combined with the main search population terms; the MEDLINE search is provided in Appendix 1.

Targeted grey literature searches were carried out to identify reports, guidelines and policy in January 2022. The websites of the following organisations were searched:

• Department of Health and Social Care (https://www.gov.uk/government/organisations/department-of-health-and-social-care)

• Health Foundation (www.health.org.uk)

• National Institute for Health and Care Excellence (www.nice.org.uk)

• Nuffield Trust (www.nuffieldtrust.org.uk)

• The database OpenGrey (https://easy.dans.knaw.nl/ui/advancedsearch) was searched, although it is now an archive and no new items are being added.

Citation searching of the 49 initially included studies, from the screen of the main and extra falls searches, was undertaken on Web of Science on 9 March 2022.

Reference checking of included studies and relevant existing reviews was completed. Search results were downloaded to a bibliographic management database (EndNote X9, Clarivate, Philadelphia, PA, USA) and deduplicated. Records were exported to EPPI-Reviewer (EPPI Centre, University College London, London, UK) systematic review software for coding and analysis.

Inclusion and exclusion criteria

Study selection

Selection of studies for the review was carried out in stages. In view of the large number of records retrieved, keyword searching of EPPI-Reviewer for relevant terms in titles and abstracts was used as a preliminary filter. Search terms included ‘care home(s)’, ‘nursing home(s) (NHs)’, ‘assisted living’, ‘extra-care’, ‘ambulance’, ‘paramedic’, ‘skilled nursing facility’ and ‘residential aged care facility (RACF)’. Records that contained relevant terms but were obviously not relevant based on their title were excluded by a single reviewer. Titles and abstracts of remaining records were screened by two reviewers independently using the inclusion criteria above. Discrepancies were resolved by discussion and, if necessary, by reference to a third reviewer. Full-text items that appeared potentially to meet the inclusion criteria were obtained and evaluated by two reviewers independently, with disagreements resolved as above. Records of the process were maintained in EPPI-Reviewer.

Data extraction and quality (risk of bias) assessment

Data were extracted from included studies in EPPI-Reviewer Web (EPPI Centre, University College London, UK) using a customised set of codes covering the study characteristics, key findings/conclusions and strengths/limitations. Effect measures were extracted as reported by study authors. Data on intervention components and delivery were extracted using the Template for Intervention Description and Replication (TIDiER-Lite) checklist. We used the Promoting Action on Research Implementation in Health Services (PARIHS) framework to support extraction of relevant data on implementation of interventions from included UK studies and the FITAR tool to assess applicability of international evidence to the UK context. PARIHS incorporates domains covering evidence, context and facilitation. 12 FITAR covers elements of the intervention/initiative, features of workforce, features of services, systems leadership, financial and commissioning processes and patients/populations. 11

We assessed risk of bias for studies using recognised research designs using the following tools:

• Joanna Briggs Institute checklists for randomised controlled trials (RCTs) and quasi-experimental studies. 13

• National Heart, Lung and Blood Institute checklist for cohort and cross-sectional studies. 14

• Mixed Methods Appraisal Tool for mixed methods and qualitative studies. 15

Assessments were performed by two reviewers independently, with discrepancies resolved by consensus or referral to a third reviewer.

In addition to risk of bias assessment, we extracted data on the strengths and limitations of each included study. Strengths and limitations were those reported by study authors and/or identified by members of the review team. Evidence sources with major limitations, namely absence of a control group and/or a small sample size (judged qualitatively) were described as low quality to reflect the need to treat such studies with caution as evidence of intervention effectiveness. This does not mean that the research was poorly conducted or not of value in its own setting.

Synthesis of evidence

The synthesis of evidence adopted a narrative synthesis approach as specified in the review protocol. Narrative synthesis has been described as ‘an approach to the systematic review and synthesis of findings from multiple studies that relies primarily on the use of words and text to summarise and explain the findings of the synthesis’. 16 Narrative synthesis typically involves four stages: developing a theoretical model of how the intervention(s) of interest might work; developing a preliminary synthesis; exploring relationships in the data; and assessing robustness of the synthesis (strength of evidence). 16

Interventions to reduce unplanned hospital admissions of care home residents are diverse and involve different health and care professionals intervening at different stages of residents’ care pathways. Our taxonomy of interventions (see Table 1) identified these and formed the theoretical model for the narrative synthesis.

We grouped studies by type of intervention and setting (UK or international) and the preliminary synthesis was performed within these groups. We summarised study characteristics, findings and study quality (risk of bias assessment where applicable plus strengths and limitations) for each group, together with any general issues about implementation or applicability to the UK setting using the PARIHS and FITAR tools, respectively. For studies that reported sufficient detail, we extracted information on intervention characteristics and delivery using the TIDiER-Lite checklist. Studies were assigned to one intervention group, but the synthesis took account of links between intervention types; for example, ACP can be a stand-alone intervention, part of a quality improvement (QI) programme or linked to an approaching need for palliative/end-of-life care. We used the totality of extracted data for each type of intervention to seek to identify factors that might make the interventions more or less effective and/or influence their implementation in routine practice (described as ‘exploring relationships in the data’ by Popay et al. ). 16

We classified the overall strength of evidence for intervention effectiveness as ‘stronger’, ‘weaker’, ‘very limited’ or ‘inconsistent’ based on the following criteria:17

• ‘stronger evidence’ represents generally consistent findings in multiple studies with a comparator group design

• ‘weaker evidence’ represents generally consistent findings in one study with a comparator group design and several noncomparator studies or multiple noncomparator studies

• ‘very limited evidence’ represents an outcome reported by a single study

• ‘inconsistent evidence’ represents an outcome for which less than 75% of the studies agree on the direction of effect.

Evidence on effectiveness was considered alongside that on feasibility, acceptability and ‘cost-effectiveness’ to assist decision-makers in forming an overall assessment of the value of the intervention. We specifically aimed to identify which interventions are best supported by UK evidence and which interventions in use elsewhere may be suitable for adaptation and evaluation in the UK context. All studies included in the review were included in the analysis of the overall strength of evidence, with no exclusions based on study design or risk of bias.

Evidence summary tables are presented in the main text. Detailed tables on intervention characteristics, implementation and applicability are presented in (see Appendices 2 and 3, Tables 33–41). Risk of bias tables for different study designs are presented in Appendix 4, Tables 42–45.

Variations from protocol

As noted above, keyword searches were undertaken to prioritise records for screening in EPPI-Reviewer because of the large number of records retrieved by the search. A random 10% sample of remaining records was checked and no further references were selected for abstract or full-text screening. One relevant reference that would have been overlooked was subsequently discovered by chance as part of a separate EPPI-Reviewer search.

Results of literature search

The database searches retrieved 24,656 references which were imported into Endnote X9. After the removal of 7691 duplicates there were 16,965 unique references. The unique references were then imported in EPPI-Reviewer systematic reviews software and a check for duplicates found a further 120 duplicates leaving 16,845 unique references. The large number of references would have taken up too much time and resources to screen; thus keyword searches were undertaken on EPPI-Reviewer to prioritise references for screening (see Variations from protocol). Single-reviewer title screening was then undertaken on 6141 references and 576 were then screened on abstract by two reviewers with 234 included for full-text screening. A total of 96 references were included in the full review from the original database search.

The extra focused database search for falls prevention retrieved 198 references after duplicates within the falls and from the original search were removed. All 198 references were screened on title and 22 were included for abstract screening; 17 references were included for full-text screening and 7 were included in the review.

Citation searches retrieved 620 references. After deduplication within the citation search results and the Endnote library, 406 references were imported into Endnote for screening. Title screening included 84 of these references for abstract screening and 32 were included for full-text screening with 15 included in the final review.

Reference checking of included studies found four further studies for inclusion in the review. A further two included studies were found from alerts. Search of websites of relevant organisations retrieved six potential additional publications, none of which were included.

The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram18 (see Figure 1) illustrates the study selection process. In summary, 124 publications were included in the review. Given that some interventions/programmes were represented by multiple publications, and that we included qualitative and implementation research as well as trials and service evaluations, the term ‘study’ is used to refer to any type of publication (primarily peer-reviewed journal articles) or report (primarily grey literature).

FIGURE 1.

PRISMA 2020 flow diagram. 18

Summary of included studies

Of the 124 studies from which we extracted data, 30 were from the UK, 44 from the USA, 24 from Australia, 4 from New Zealand, 20 from other countries and 2 from multiple countries (see Table 2). The most common types of intervention were integrated working (particularly in the UK and Australia) and QI programmes (particularly in the USA).

In terms of study design, the largest single group was cluster RCTs (17 studies), followed by uncontrolled before–after (15), controlled before–after (13), non-RCTs (11), qualitative studies (11) and mixed-methods studies (10). Four studies used a step-wedge design, involving randomisation to introduce the intervention at different times during the study. Thirty-two studies used other designs, including cohorts and secondary data analyses. The studies reported a wide range of quantitative and qualitative outcomes but no patient-reported outcome measures were recorded.

Quality improvement programmes

Integrated working

Training/workforce development

Interventions classified as training/workforce development were more specific than QI programmes. These interventions focused on training within a specific speciality, for example dementia, or to manage a specific problem such as delirium or recognising acute physiological deterioration.

Management of specific problems

Paramedic assessment/non-conveyance

Only one included study dealt with paramedic assessment and decision-making when there is uncertainty about taking a care home resident to hospital or treating them in situ. 101 This qualitative focus group study involved 18 German paramedics. Participants perceived that hospital transfers were potentially avoidable in some cases of urinary catheter complications, infections and falls. Uncertainties about legal liability were seen as a key factor favouring transfer of residents to hospital. Although not a UK study, the findings may be applicable to UK paramedics and could inform implementation of specialist paramedic services and training for paramedics attending care homes. See also the qualitative study of decision-making at the end of life by Murphy-Jones et al. 102 described below (see Palliative/end-of-life care).

Emergency department interventions

ED interventions were defined as interventions which were initiated in the ED or conducted by ED staff. Two studies met this definition (see Table 22) and were both conducted in Australia. 70,103 Both studies are before and after studies conducted at single regional EDs which receive admissions from multiple care homes within their catchment area. One study included 5991 and the other study included 1271 ED attendances in analysis. Marsden et al. 103 included no control group and was rated as higher risk of bias than Hullick et al. ,70 which included control care homes which did not receive the intervention; however, allocation was not randomised.

Marsden et al. 103 present a nurse-led multidisciplinary intervention delivered in the ED who receive referrals from care homes and expedite geriatric assessment and other clinical management in the ED. When controlling for all other variables, the hazard ratio for discharge for those patients presenting in the post intervention period was 1.15 (95% CI 1.05 to 1.26), indicating a significant reduction in hospital admission in the post-intervention period. When controlling for all other variables the hazard ratio for reduced length of hospital stay for those patients presenting in the intervention period was 1.49 (95% CI 1.24 to 1.78) and 1.44 (95% CI 1.30 to 1.60) for the interim intervention period. Hullick et al. 70 evaluated a video telehealth consultation with hospital ED staff in addition to an existing QI programme aimed at reducing ED attendances from care homes. Fourteen months’ pre- and post-intervention data were analysed. There was no significant difference in hospital admission or ED visits after the introduction of telehealth.

The evidence is low quality from single centres outside the UK setting. Therefore, we are unable to recommend use of either intervention. That said, models of geriatric assessment in the ED like that described by Marsden et al. are used in some centres in the UK as part of a frailty service104,105 and both the clinical and cost effectiveness require evaluation.

Advance care planning

Promoting ACP was part of most QI interventions included in the review. This section includes studies of ‘stand-alone’ interventions to encourage ACP.

Palliative/end-of-life care

Other interventions

A single study from the USA evaluated a decision aid in a mixed-methods RCT. 139 Residents and families rated the decision aid as very helpful but it did not decrease hospital transfers compared with the control group. Installation of ‘compliant flooring’ to reduce fall-related injuries and hospitalisations also had no effect in a randomised trial conducted in Canada. 140

The Veterans’ Administration Medical Foster Home Programme is discussed below.

UK economic evidence

Table 29 summarises UK studies reporting economic outcomes. The majority of studies reported cost savings but weak study designs and limited reporting mean that the findings should be treated with great caution. One report extrapolated findings from one area across the whole of England to estimate possible cost savings. 56 Some reports were produced by manufacturers56 or professional bodies55 with a potential conflict of interest. One report indicated that the intervention increased costs in the first two years,52 probably reflecting the introduction of new ways of working. The programme would require reductions in costs (preferably from reduced hospital admissions) to be sustainable without additional funding.

The two most rigorous studies83,89 collected promising economic data but they were both feasibility trials and in one case the intervention proved impossible to implement, making a full trial unviable. 89

Interventions in use in the UK

That the quantity of the evidence was lower for the UK compared with other countries combined is to be expected, given that the UK is only one country, but it covered almost all the intervention categories in our taxonomy.

A particular focus in the UK literature has been on links with primary care and community services through the care home vanguards programme, which has subsequently provided the basis for EHCH. The vanguards, and subsequently EHCH, combined elements of integrated working with QI and locally important elements. The vanguards programme involved substantial support being given to the localities selected as pilot sites, something that has not been continued at the same level for national roll-out of EHCH, raising questions about how easy it might be to embed and sustain such programmes longer term. Early research has outlined the considerations during implementation that enable care home interventions to embed and sustain, but this has been based upon a small number of studies, largely focussed around QI. 143 There is much potential for researchers and practitioners to both test these processes and learn from successes and failures during national roll-out of EHCH.

Other interventions used in the UK included local QI and training programmes, initiatives to tackle specific problems such as delirium and inappropriate prescribing, or to introduce components of ACP and palliative/end-of-life care. Quantitative studies of paramedic-led interventions were lacking, although qualitative research highlighted the complexities and tensions involved in decision-making. These findings reinforce long-standing findings about the difficulty of negotiating care during acute crises in care homes due to uncertain organisational boundaries, roles and responsibilities when residents’ health deteriorates. 144,145 A decade after these findings first appeared, work is needed to better standardise operating procedures between care homes and ambulance providers during medical crises.

UK care home research is in its infancy,144,145 but research considering those receiving social care in other settings, for example assisted living (with more limited services available on site), is even less well developed. It was to a certain extent unsurprising that we found a paucity of publications around such settings.

International interventions

The range of interventions used in other high-income countries was broadly similar to that seen in the UK. The US Medicare system has evaluated a range of QI programmes with memorable acronyms such as INTERACT and OPTIMISTIC. New Zealand is also a source of important studies of QI interventions.

An important group of studies from Australia evaluated responsive services provided to groups of care homes by hospital-based MDTs, often with an emergency medicine background. Details of the service models varied (e.g. in-reach, outreach, HiNH) but studies suggested substantial reductions in unplanned admissions and other outcomes, albeit sometimes with weak study designs. Such models are not standard in the UK. Hospital at Home in the UK mirrors some of the approaches undertaken in Australia. Care homes were excluded from the largest RCT of Hospital at Home in the UK,146 but are supported by some local Hospital at Home services. 147 Urgent Community Response, which describes short-duration intensive healthcare support in the community for older people who have experienced acute deterioration and which is part of the NHS England Ageing Well programme, also excludes care home residents from its core specification. 148 The national roll-out in England of Hospital at Home as part of the Frailty Virtual Wards initiative, alongside Urgent Community Response, provides an opportunity to evaluate the applicability of approaches evidenced to work in Australia in the UK setting. This would, though, require care home residents to be included and for focused evaluation around these questions.

Telehealth interventions have been evaluated in the USA and other settings (including to a small extent in the UK) with mixed results. There is a fairly substantial international evidence base evaluating ACP and palliative care interventions, although results have been inconsistent (see Table 32). The considerations when implementing telehealth in care home settings have been well described, although there remain specific challenges around implementing video consultations for residents with cognitive or sensory impairment and further research is required to better understand the role of telehealth in such groups. 149

We found only one study based in an assisted living setting (in the USA). The Veterans’ Administration Medical Foster Home model is another evaluated intervention with no apparent equivalent in the UK.

Effectiveness of interventions

We graded evidence for effectiveness according to a relatively simple scheme. Much of the evidence was graded as ‘weaker’ because of a shortage of controlled studies or ‘inconsistent’. Interventions with ‘stronger’ evidence included improved links with primary/community care and QI programmes with staff attached to care homes. Diversity of outcome measures and heterogeneity of interventions precluded meta-analysis but claimed reductions (against control or baseline) were often of the order of 30–50%, which would imply significant reductions in resource use. Groupings presented in our tables are not the only ones and readers can form their own assessment (but note that single studies always constitute ‘very limited’ evidence).

Implementation in UK

The barriers to implementation of healthcare interventions in care homes are well described. and include staff shortages, high staff turnover and care homes changing ownership or going out of business. Problems affecting service delivery in care homes affect the wider urgent/emergency care system and vice versa. A recurrent theme, throughout the literature, both prior to and during the pandemic, is insufficient recognition or engagement with expertise held in and by the care home sector. Serial studies have shown a willingness and enthusiasm among care home staff to make healthcare better for their residents, but a tendency among NHS providers to engage with them late in service design, if at all, with services designed to give care home staff little latitude in how they are implemented at an individual care home level. 89,150–153

Facilitators to implementation were identified, including building on existing services/relationships and the role of committed leaders in both the care home and health sectors. Key within this will be empowerment of care home providers and staff to shape the design and implementation of care home healthcare interventions. Integrated care boards bring together local authorities and NHS commissioners and providers. Care home providers, though, are predominantly corporate or third sector and remain largely ‘outside the room’. There is promise that this may change with the appointment of a chief nurse for social care, who might better facilitate consultation with national care home provider organisations, but specific work is required to build relationships with care home providers and staff at a local and regional level. 154 The situation will not improve without investment in such relationships.

Applicability of international evidence

Much is made in the literature of the differences in service delivery between care home sectors internationally. This is important in the way that it influences: the core competencies of staff employed in care homes (for example through mandatory minimum training in the USA and Australia): the availability of in situ resident medical staff (e.g. elderly care physicians in the Netherlands); the role of medical staff in determining care home policies (for example through the medical director initiative in the USA) and the sense of care homes as being ‘inside’ the system (more common where care homes are publicly funded, as in Denmark or the Netherlands).

An important overarching consideration, however, is that the cohorts of residents living in care homes are very similar around the globe. 155 This was evidenced in the markedly similar way that COVID-19 affected care home residents internationally. 156 Long-term care systems are no more similar or different than healthcare systems. UK and US long-term care arguably have more in common than do acute hospital systems in the respective countries. It is therefore possible to transfer approaches between countries, so long as attention is paid to the differences and similarities between systems, and so long as new models from overseas are evaluated in parallel with implementation when introduced for the first time.

The FITAR tool proved useful for identifying a wide range of issues that could affect applicability of international evidence. Examples included mixtures of long-stay and short-stay residents in some nursing homes (USA), cultural attitudes to ACP/palliative care (USA/Europe), workforce regulations and roles of different grades (especially nurses; USA/Australia) and public ownership of care homes (Denmark). Extraction of data at the level of the country rather than the individual study would have saved time in the overall review process.

Economic evidence

We identified a relatively large amount of economic evidence for the UK but this was mostly of low quality. International evidence was included for completeness but there is a a need for particular caution in extrapolating from other countries or time periods. Costing social care interventions has long provided difficulties in the UK, as NHS and social care costs are collated differently and separately. 157 Additional challenges come from understanding how costs are transferred between health and social care commissioners and providers. 158 An intervention that saves the NHS money may ultimately generate cost, opportunity cost, or both for care home providers. Finally, quality of life measures have limited applicability in care home residents because of the prevalence of dementia in this setting, and this limits some types of health economic evaluation. 159 Substantial work is required to develop frameworks to better evaluate costs, and cost-effectiveness of interventions that straddle the interface between health and social care, and in care home research more generally.

Strengths and limitations

Strengths of the review include its broad coverage of UK and international evidence, a thorough search including grey literature and citation searching into 2022 and reference list checking and use of two independent reviewers for most standard review processes.

We extracted data on details of interventions, which is important for understanding complex interventions; the relevant tables may be a useful resource for designing new care models.

Coverage of the international literature was sufficient to answer our research question about possible candidate interventions that could be relevant to UK practice. We evaluated risk of bias for most included studies using design-specific tools. Statements about overall strength of evidence were informed by risk of bias assessment even where the key considerations are not explicitly stated.

Our findings were broadly similar to a review of specialist health-care services in care homes by Buck et al. published in 2021. 8 This review used a slightly different framework of interventions to ours but with many common elements. The study reported that interventions addressing residents’ general health needs, assessment and management services and non-training initiatives involving medical staff can reduce hospital admissions, while there was also promising evidence for services targeting residents at imminent risk of hospital admission or post-hospital discharge and training-only initiatives. There was evidence that end-of-life care services may enable residents to remain in the home at the end of life, but the high number of poor-quality studies (authors’ terminology) undermined confidence in this result. Buck et al. 8 used a ‘vote counting’ approach to evidence synthesis with binomial testing to assess statistical significance of the findings. This contrasts with our more qualitative approach to assessing strength of evidence by intervention and setting. Our assessment of implementation and applicability issues and our specific focus on unplanned admissions differentiates our approach from that of Buck et al. 8 Both reviews have relevance for supporting evidence-informed decision-making by service commissioners and providers.

The large number of records identified by our searches led us to prioritise items for screening by searching for relevant keywords. This is not uncommon and is allowed for in the latest version of the PRISMA flow diagram. We checked a random sample of unscreened records which confirmed that the risk of missing relevant items through this approach was likely to be low. We did not attempt any meta-analyses because of the heterogeneity of included studies for most types of intervention together with lack of availability of the required data and/or diversity of effect measures in some cases.

A further possible limitation was the framing of our analysis around a predefined taxonomy of interventions. This worked well in most cases but inevitably, there was an element of subjectivity in placement of a few studies. Our approach was to assign each study to one group but it would be possible to conduct a sensitivity analysis by adding or removing ‘marginal’ studies to identify any effects on the overall assessment of strength of evidence.

Our overall assessment was less comprehensive than that provided by Grading of Recommendations, Assessment, Development and Evaluation (GRADE) and similar systems. This reflects our decision to produce a broadly focused review and the recognition that much of the evidence either fell outside the domains of GRADE or would be heavily downgraded because of the use of nonrandomised study designs. It is also worth noting that qualitative evidence, while relevant to some aspects of the review, is not adequately assessed by either our system or GRADE.

Equity, diversity and inclusion

This section complements the discussion of review processes above (see Chapter 2). Diversity in care home populations is likely to be increasing in many countries in line with that of society as a whole, although research on this topic was outside the scope of our review. We cannot be certain how far the care homes, residents and staff in our included studies were representative of the wider care home population. However, there is evidence that care home residents worldwide are relatively similar at the population level (see Chapter 5). Equity and diversity issues affecting care homes may thus reflect those present in society as a whole and specifically in the health and care system.

Decision-making about admission of care home residents to hospital can raise issues of equity, although given our focus on implementing and evaluating interventions these were only touched on in our review. Family advocacy may influence the decision-making process (see Chapter 3). Residents with dementia represent a particularly vulnerable group whose needs require attention to support equitable treatment.

The health and social care workforce is disproportionately female and highly diverse in terms of characteristics such as ethnicity and socioeconomic status. Research involving care home staff should seek to ensure that all groups are appropriately represented, for example in qualitative studies and surveys. Our review included too small a sample of such studies to draw meaningful conclusions, but this could be an interesting topic for further research.

Implications for service delivery

We have identified the following implications for service delivery:

• Opportunities to reduce unplanned admissions exist at all stages of residents’ care journeys from routine care to palliative care at end-of-life.

• Types of intervention such as QI programmes or integrated working between care homes and primary care/community services differ in aspects such as workforce requirements, technology, intensity of the intervention etc. Services can consider adapting described interventions to their own context, including possibilities for simplification where evidence supports this.

• Evidence suggests that care home managers and staff support proposed interventions that will help them to deliver better care for their residents. Early and genuine consultation to assess feasibility and acceptability of interventions could be a major factor in successfully implementing new service models.

• Specific work is required to build relationships between NHS and care home providers and staff at a local and regional level.

• There is some evidence to guide where changes to services are more likely to improve outcomes; for example, care homes without nursing may benefit more than those with nursing from some forms of support because of their lower baseline level of staffing and because of differences in case mix.

• In attempting to transfer approaches between countries, attention should be paid to the differences and similarities between systems, and new models from overseas should be evaluated in parallel with implementation in the UK setting. Changes to how care is delivered to older people in the community, including Hospital at Home and Urgent Community Response, mirror some initiatives targeted at care homes overseas and could provide an opportunity to test and evaluate imported models of working if care homes are adopted within remit.

• Work is needed to better understand and standardise operating procedures between care homes and ambulance providers seeking to negotiate care for residents during medical crises, particularly with regard to lines of responsibility and shared liability for decision making.

Implications for research

We have identified the following priorities for research:

• Researchers should carefully consider what is realistic in terms of study design and data collection given the current UK context of extreme pressure on care homes. As with changes to service delivery, genuine involvement of care home residents, family members and staff is required to design and deliver high-quality research. Development and reporting of appropriate patient-reported outcome measures is recommended.

• Research is needed to better understand the factors that enable effective interventions to become embedded and sustained in practice over the long term.

• There is a need for rigorous economic evaluations, ideally using measures that can be used to compare different interventions and taking into account costs associated with implementation, particularly how costs are transferred between health and social care commissioners and providers.

• The national roll-out in England of Hospital at Home as part of the Frailty Virtual Wards initiative, alongside Urgent Community Response, provides an opportunity to evaluate the applicability of approaches evidenced to work in Australia in the UK setting.

• Further research is required to evaluate approaches based on paramedic assessment and potential non-conveyance, including assessment of safety and qualitative studies of resident, family carer and care home staff/management perspectives.

• Further research is required to better understand the role of telehealth in reducing unplanned admissions of residents with cognitive or sensory impairments.

• We found only one study based in an assisted living setting (in the USA). Research to evaluate interventions to reduce unplanned admissions from such settings in the UK is required, bearing in mind the lower levels of both resident need and on-site services.

Conclusions

This updated and extended systematic review has identified a substantial body of relevant research (124 publications) published between 2014 and early 2022. Much of the evidence is of low quality because of flaws such as uncontrolled study designs and small sample size. Despite this, it is clear that interventions at various stages of residents’ care pathways can reduce unplanned admissions if implemented effectively. Most interventions are complex and require co-operation between health and social care staff. This suggests that intervention effectiveness in routine practice is likely to be influenced by local contextual factors.

In the UK and internationally, integrated working and QI programmes have the strongest evidence of effectiveness. Integrated working takes a variety of forms, with primary and community care links being emphasised in UK practice and the role of hospital-based MDTs being more prominent in Australia. Simpler training and staff development initiatives have shown mixed results. ACP is key to ensuring that residents’ and families’ wishes are respected at the end of life but ‘do not hospitalise’ orders cannot be recommended. Qualitative research has identified key themes around decision-making involving paramedics, care home staff and residents/family carers but further work is required to optimise ways of working together and minimise unplanned admissions. Access to high-quality palliative care for care home residents is an ethical requirement but the best way to deliver this is unclear in the face of inconsistent and generally low-quality evidence.

The relevance of international evidence to UK practice is often debated. We were able to identify key factors that decision-makers may want to take into account in evaluating the suitability of international care models for testing in UK settings. Fundamental similarity of the characteristics of care home residents included in research studies tends to support transferability of evidence between different health and care systems.

The UK care home sector, like the health and care system generally, is under great pressure as a result of well-known problems including staff shortages, financial instability and the COVID-19 pandemic. The PARIHS framework identified these and other barriers to implementing interventions to reduce unplanned admissions. Facilitators were also identified, and the Care Home Vanguard studies included in this review led to the roll-out of the EHCH care model in England. Even when interventions have been successfully implemented, the pressures on the system mean that longer-term sustainability cannot be taken for granted and should be considered from an early stage.

Finally, we identified priorities for research, including better economic evaluations; evaluation of ‘HiNH’ and similar approaches; and appropriate interventions to reduce unplanned admissions from assisted living settings. Research studies should have realistic design and objectives and involve care home staff, managers and residents/families from the outset.

Contributions of authors

Duncan Chambers (https://orcid.org/0000-0002-0154-0469) (Research Fellow in Public Health) managed the project and contributed to all review processes and report writing.

Anna Cantrell (https://orcid.org/0000-0003-0040-9853) (Research Fellow in Health Economics and Decision Science) performed the literature searches and contributed to all review processes and report writing.

Louise Preston (https://orcid.org/0000-0001-7477-4517) (Senior Research Fellow in Health Economics and Decision Science) contributed to all review processes and report writing.

Carl Marincowitz (https://orcid.org/0000-0003-3043-7564) (NIHR Academic Clinical Lecturer in Emergency Medicine) acted as a topic expert advisor and contributed to all review processes and report writing.

Lynne Wright (https://orcid.org/0009-0008-6637-2911) (public co-applicant) acted as a topic expert advisor and contributed to writing the plain English summary and PPI sections of the report.

Simon Conroy (https://orcid.org/0000-0002-4306-6064) (Professor of Geriatric Medicine) acted as a topic expert advisor and contributed to writing the report.

Adam Lee Gordon (https://orcid.org/0000-0003-1676-9853) (Professor of the Care of Older People) acted as a topic expert advisor and contributed to writing the report.

Ethical review

Not required.

Information governance statement

As a literature review, this study did not handle any personal data.

Data-sharing statement

Any additional data not included in this report and its appendices are available on request. All queries should be submitted to the corresponding author.

Disclaimers

This report presents independent research funded by the National Institute for Health and Care Research (NIHR). The views and opinions expressed by authors in this publication are those of the authors and do not necessarily reflect those of the NHS, the NIHR, the HSDR programme or the Department of Health and Social Care. If there are verbatim quotations included in this publication the views and opinions expressed by the interviewees are those of the interviewees and do not necessarily reflect those of the authors, those of the NHS, the NIHR, the HSDR programme or the Department of Health and Social Care.

Main search strategy

Database: Ovid MEDLINE(R) and Epub Ahead of Print, In-Process, In-Data-Review and Other Non-Indexed Citations and Daily <1946 to 14 December 2021>

Search strategy:

1. residential facilities/or homes for the aged/or nursing homes/(47,896)

2. ‘residential care’.ab,ti. (3754)

3. ‘care home*’.ab,ti. (4721)

4. ‘nursing home*’.ab,ti. (32,554)

5. Assisted Living Facilities/(1528)

6. ‘assisted living’.ab,ti. (2410)

7. ‘extra care housing’.ab,ti. (16)

8. or/1–7 (65,730)

9. exp *aged/or exp *geriatrics/or exp *geriatric nursing/or (centarian* or centenarian* or elder* or eldest or frail* or geriatri* or nonagenarian* or octagenarian* or octogenarian* or old age* or older adult* or older age* or older female* or older male* or older man or older men or older patient* or older people or older person* or older population or older subject* or older woman or older women or oldest old* or senior* or senium or septuagenarian* or supercentenarian* or very old*).ti,kf. (306,918)

10. 8 or 9 (355,492)

11. Patient Admission/(25,826)

12. (unplanned adj3 (admission* or hospital*)).ab,ti. (2672)

13. (avoidable adj3 (admission* or hospital*)).ab,ti. (947)

14. ‘community geriatric* service*’.ab,ti. (8)

15. Case Management/(10,401)

16. ‘case management’.ab,ti. (11,442)

17. (discharg* adj3 plan*).ab,ti. (5382)

18. ‘Delivery of Health Care, Integrated’/(13,761)

19. (integrated adj3 (working or care)).ab,ti. (12,152)

20. Delirium/pc (Prevention and Control) (1248)

21. (prevent* adj3 deliri*).ab,ti. (1173)

22. [(medicine* or medication*) adj3 (manag* or monitor* or review*)].ab,ti. (25,092)

23. Terminal Care/(30,296)

24. ‘terminal care’.ab,ti. (1553)

25. ‘end of life care*’.ab,ti. (10,925)

26. care, end-of-life.ab,ti. (128)

27. eol.ab,ti. (2217)

28. Advance Care Planning/(3619)

29. ‘advance care planning’.ab,ti. (3566)

30. Fluid Therapy/(21,238)

31. (hydration adj3 nutrition).ab,ti. (1154)

32. (‘non convey*’ or nonconvey*).ab,kw,ti. (52)

33. ‘specialist paramedic*’.ab,ti. (8)

34. (paramedic* adj3 assess*).ab,ti. (196)

35. ‘training course* ’.ab,ti. (5569)

36. [(vocational or educational) adj qualification*].ab,ti. (804)

37. ‘hydration and nutrition’.ab,ti. (232)

38. ‘geriatric* specialist’.ab,ti. (40)

39. Vaccination/(90,966)

40. [(influenza* or flu*) adj3 vaccin*].ab,ti. (26,583)

41. covid-19 vaccines/or influenza vaccines/ (32,036)

42. [(covid* or corona* or ‘SARS CoV 2’) adj3 vaccin*].ab,ti. (11,950)

43. Pneumococcal Vaccines/(8162)

44. ‘pneumo* vaccine*’.ab,ti. (4107)

45. Oxygen Inhalation Therapy/(15,411)

46. (oxygen adj1 therap*).ab,ti. (12,517)

47. or/11–46 (320,076)

48. 10 and 47 (16,212)

49. letter/(1,162,749)

50. editorial/(589,748)

51. news/(210,552)

52. exp historical article/(406,561)

53. anecdotes as topic/(4746)

54. comment/(942,690)

55. case report/(2,232,728)

56. (letter or comment*).ti. (172,085)

57. or/49–56 (4,669,568)

58. randomized controlled trial/or random*.ti,ab. (1,404,034)

59. 57 not 58 (4,640,014)

60. animals/not humans/(4,896,931)

61. exp animals, laboratory/(924,650)

62. exp animal experimentation/(9971)

63. exp models, animal/(613,504)

64. exp rodentia/(3,391,961)

65. (rat or rats or mouse or mice).ti. (1,384,553)

66. or/59–65 (10,383,418)

67. 48 not 66 (14,900)

68. afghanistan/or africa/or africa, northern/or africa, central/or africa, eastern/or ‘africa south of the sahara’/or africa, southern/or africa, western/or albania/or algeria/or andorra/or angola/or ‘antigua and barbuda’/or argentina/or armenia/or azerbaijan/or bahamas/or bahrain/or bangladesh/or barbados/or belize/or benin/or bhutan/or bolivia/or borneo/or ‘bosnia and herzegovina’/or botswana/or brazil/or brunei/or bulgaria/or burkina faso/or burundi/or cabo verde/or cambodia/or cameroon/or central african republic/or chad/or exp china/or comoros/or congo/or cote d’ivoire/or croatia/or cuba/or ‘democratic republic of the congo’/or cyprus/or djibouti/or dominica/or dominican republic/or ecuador/or egypt/or el salvador/or equatorial guinea/or eritrea/or eswatini/or ethiopia/or fiji/or gabon/ or gambia/ or ‘georgia (republic)’/ or ghana/ or grenada/ or guatemala/ or guinea/ or guinea-bissau/ or guyana/ or haiti/ or honduras/ or independent state of samoa/ or exp india/ or indian ocean islands/ or indochina/ or indonesia/ or iran/ or iraq/ or jamaica/ or jordan/ or kazakhstan/ or kenya/ or kosovo/ or kuwait/ or kyrgyzstan/ or laos/ or lebanon/ or liechtenstein/ or lesotho/ or liberia/ or libya/ or madagascar/ or malaysia/ or malawi/ or mali/ or malta/ or mauritania/ or mauritius/ or mekong valley/ or melanesia/ or micronesia/ or monaco/ or mongolia/ or montenegro/ or morocco/ or mozambique/ or myanmar/ or namibia/ or nepal/ or nicaragua/ or niger/ or nigeria/ or oman/ or pakistan or palau/or exp panama/or papua new guinea/or paraguay/or peru/or philippines/or qatar/or ‘republic of belarus’/or ‘republic of north macedonia’/or romania/or exp russia/or rwanda/or ‘saint kitts and nevis’/or saint lucia/or ‘saint vincent and the grenadines’/or ‘sao tome and principe’/or saudi arabia/or serbia/or sierra leone/or senegal/or seychelles/or singapore/or somalia/or south africa/or south sudan/or sri lanka/or sudan/or suriname/or syria/or taiwan/or tajikistan/or tanzania/or thailand/or timor-leste/or togo/or tonga/or ‘trinidad and tobago’/or tunisia/or turkmenistan/or uganda/or ukraine/or united arab emirates/or uruguay/or uzbekistan/or vanuatu/or venezuela/or vietnam/or west indies/or yemen/or zambia/or zimbabwe/(1,186,145)

69. ‘Organisation for Economic Co-Operation and Development’/(404)

70. australasia/or exp australia/or austria/or baltic states/or belgium/or exp canada/or chile/or colombia/or costa rica/or czech republic/or exp denmark/or estonia/or europe/or finland/or exp france/or exp germany/or greece/or hungary/or iceland/or ireland/or israel/or exp italy/or exp japan/or korea/or latvia/or lithuania/or luxembourg/or mexico/or netherlands/or new zealand/or north america/or exp norway/or poland/or portugal/or exp ‘republic of korea’/or ‘scandinavian and nordic countries’/or slovakia/or slovenia/or spain/or sweden/or switzerland/or turkey/or exp united kingdom/or exp united states/(3,363,943)

71. European Union/(17,009)

72. Developed Countries/(21,011)

73. 69 or 70 or 71 or 72 (3,379,076)

74. 68 not 73 (1,099,831)

75. 67 not 74 (14,273)

76. limit 75 to english language (12,989)

77. limit 76 to year=‘2014–Current’ (6385)

MEDLINE extra falls search

Database: Ovid MEDLINE(R) and Epub Ahead of Print, In-Process, In-Data-Review and Other Non-Indexed Citations and Daily <1946 to 14 January 2022>

Search strategy:

1. residential facilities/or homes for the aged/or nursing homes/(48,006)

2. ‘residential care’.ab,ti. (3775)

3. ‘care home*’.ab,ti. (4774)

4. ‘nursing home*’.ab,ti. (32,705)

5. Assisted Living Facilities/(1531)

6. ‘assisted living’.ab,ti. (2433)

7. ‘extra care housing’.ab,ti. (16)

8. or/1–7 (65,979)

9. exp *aged/or exp *geriatrics/or exp *geriatric nursing/or (centarian* or centenarian* or elder* or eldest or frail* or geriatri* or nonagenarian* or octagenarian* or octogenarian* or old age* or older adult* or older age* or older female* or older male* or older man or older men or older patient* or older people or older person* or older population or older subject* or older woman or older women or oldest old* or senior* or senium or septuagenarian* or supercentenarian* or very old*).ti,kf. (308,664)

10. 8 or 9 (357,423)

11. Patient Admission/(25,879)

12. [unplanned adj3 (admission* or hospital*)].ab,ti. (2699)

13. [avoidable adj3 (admission* or hospital*)].ab,ti. (954)

14. ‘community geriatric* service*’.ab,ti. (8)

15. Case Management/(10,408)

16. ‘case management’.ab,ti. (11,481)

17. (discharg* adj3 plan*).ab,ti. (5413)

18. ‘Delivery of Health Care, Integrated’/(13,793)

19. [integrated adj3 (working or care)].ab,ti. (12,255)

20. Delirium/pc (Prevention and Control) (1261)

21. (prevent* adj3 deliri*).ab,ti. (1192)

22. [(medicine* or medication*) adj3 (manag* or monitor* or review*)].ab,ti. (25,285)

23. Terminal Care/(30,376)

24. ‘terminal care’.ab,ti. (1556)

25. ‘end of life care*’.ab,ti. (10,999)

26. care, end-of-life.ab,ti. (129)

27. eol.ab,ti. (2238)

28. Advance Care Planning/(3646)

29. ‘advance care planning’.ab,ti. (3612)

30. Fluid Therapy/(21,285)

31. (hydration adj3 nutrition).ab,ti. (1161)

32. (‘non convey*’ or nonconvey*).ab,kw,ti. (53)

33. ‘specialist paramedic*’.ab,ti. (8)

34. (paramedic* adj3 assess*).ab,ti. (197)

35. ‘training course* ’.ab,ti. (5611)

36. [(vocational or educational) adj qualification*].ab,ti. (810)

37. ‘hydration and nutrition’.ab,ti. (235)

38. ‘geriatric* specialist’.ab,ti. (39)

39. Vaccination/(91,649)

40. [(influenza* or flu*) adj3 vaccin*].ab,ti. (26,718)

41. covid-19 vaccines/or influenza vaccines/(32,861)

42. [(covid* or corona* or ‘SARS CoV 2’) adj3 vaccin*].ab,ti. (13,156)

43. Pneumococcal Vaccines/(8196)

44. ‘pneumo* vaccine*’.ab,ti. (4126)

45. Oxygen Inhalation Therapy/(15,473)

46. (oxygen adj1 therap*).ab,ti. (12,610)

47. or/11–46 (322,683)

48. 10 and 47 (16,328)

49. letter/(1,166,322)

50. editorial/(592,842)

51. news/(210,646)

52. exp historical article/(407,003)

53. anecdotes as topic/(4746)

54. comment/(947,871)

55. ase report/(2,239,554)

56. (letter or comment*).ti. (173,104)

57. or/49–56 (4,685,317)

58. randomized controlled trial/or random*.ti,ab. (1,411,760)

59. 57 not 58 (4,655,618)

60. animals/not humans/(4,909,832)

61. exp animals, laboratory/(928,170)

62. exp animal experimentation/(10,021)

63. exp models, animal/(616,756)

64. exp rodentia/(3,403,886)

65. (rat or rats or mouse or mice).ti. (1,387,516)

66. or/59–65 (10,416,012)

67. 48 not 66 (15,012)

68. afghanistan/or africa/or africa, northern/or africa, central/or africa, eastern/or ‘africa south of the sahara’/or africa, southern/or africa, western/or albania/or algeria/or andorra/or angola/or ‘antigua and barbuda’/or argentina/or armenia/or azerbaijan/or bahamas/or bahrain/or bangladesh/or barbados/or belize/or benin/or bhutan/or bolivia/or borneo/or ‘bosnia and herzegovina’/or botswana/or brazil/or brunei/or bulgaria/or burkina faso/or burundi/or cabo verde/or cambodia/or cameroon/or central african republic/or chad/or exp china/or comoros/or congo/or cote d’ivoire/or croatia/or cuba/or ‘democratic republic of the congo’/or cyprus/or djibouti/or dominica/or dominican republic/or ecuador/or egypt/or el salvador/or equatorial guinea/or eritrea/or eswatini/or ethiopia/or fiji/or gabon/or gambia/or ‘georgia (republic)’ /or ghana/or grenada/or guatemala/or guinea/or guinea-bissau/or guyana/or haiti/or honduras/or independent state of samoa/or exp india/or indian ocean islands/or indochina/or indonesia/or iran/or iraq/or jamaica/or jordan/or kazakhstan/or kenya/or kosovo/or kuwait/or kyrgyzstan/or laos/or lebanon/or liechtenstein/or lesotho/or liberia/or libya/or madagascar/or malaysia/or malawi/or mali/or malta/or mauritania/or mauritius/or mekong valley/or melanesia/or micronesia/or monaco/or mongolia/or montenegro/or morocco/or mozambique/or myanmar/or namibia/or nepal/or nicaragua/or niger/or nigeria/or oman/or pakistan/or palau/or exp panama/or papua new guinea/or paraguay/or peru/or philippines/or qatar/or ‘republic of belarus’/or ‘republic of north macedonia’/or romania/or exp russia/or rwanda/or ‘saint kitts and nevis’/or saint lucia/or ‘saint vincent and the grenadines’/or ‘sao tome and principe’/or saudi arabia/or serbia/or sierra leone/or senegal/or seychelles/or singapore/or somalia/or south africa/or south sudan/or sri lanka/or sudan/or suriname/or syria/or taiwan/or tajikistan/or tanzania/or thailand/or timor-leste/or togo/or tonga/or ‘trinidad and tobago’/or tunisia/or turkmenistan/or uganda/or ukraine/or united arab emirates/or uruguay/or uzbekistan/or vanuatu/or venezuela/or vietnam/or west indies/or yemen/or zambia/or zimbabwe/(1,192,175)

69. ‘Organisation for Economic Co-Operation and Development’/(411)

70. australasia/or exp australia/or austria/or baltic states/or belgium/or exp canada/or chile/or colombia/or costa rica/or czech republic/or exp denmark/or estonia/or europe/or finland/or exp france/or exp germany/or greece/or hungary/or iceland/or ireland/or israel/or exp italy/or exp japan/or korea/or latvia/or lithuania/or luxembourg/or mexico/or netherlands/or new zealand/or north america/or exp norway/or poland/or portugal/or exp ‘republic of korea’/or ‘scandinavian and nordic countries’/or slovakia/or slovenia/or spain/or sweden/or switzerland/or turkey/or exp united kingdom/or exp united states/(3,372,732)

71. European Union/(17,054)

72. Developed Countries/(21,036)

73. 69 or 70 or 71 or 72 (3,387,922)

74. 68 not 73 (1,105,653)

75. 67 not 74 (14,378)

76. limit 75 to english language (13,092)

77. limit 76 to year=‘2014–Current’ (6488)

78. Accidental Falls/pc (Prevention and Control) (10,120)

79. [fall* adj3 (prevent* or avoid* or manag* or intervention* or project* or program*)].ab,ti. (10,401)

80. 78 or 79 (16,211)

81. 8 and 80 (1071)

82. 81 not 66 (1016)

83. 82 not 74 (980)

84. limit 83 to english language (869)

85. limit 84 to year=‘2014–Current’ (333)

86. hospitalization/or patient admission/(147,809)

87. hospitali?ation.ab,ti. (160,421)

88. 12 or 13 or 86 or 87 (267,445)

89. 85 and 88 (33)

90. 89 not 77 (21)