Beyond the control of the care home: A meta‐ethnography of qualitative studies of Infection Prevention and Control in residential and nursing homes for older people

Abstract Objective This study aimed to develop interpretive insights concerning Infection Prevention and Control (IPC) in care homes for older people. Design This study had a meta‐ethnography design. Data Sources Six bibliographic databases were searched from inception to May 2020 to identify the relevant literature. Review Methods A meta‐ethnography was performed. Results Searches yielded 652 records; 15 were included. Findings were categorized into groups: The difficulties of enacting IPC measures in the care home environment; workload as an impediment to IPC practice; the tension between IPC and quality of life for care home residents; and problems dealing with medical services located outside the facility including diagnostics, general practice and pharmacy. Infection was revealed as something seen to lie ‘outside’ the control of the care home, whether according to origins or control measures. This could help explain the reported variability in IPC practice. Facilitators to IPC uptake involved repetitive training and professional development, although such opportunities can be constrained by the ways in which services are organized and delivered. Conclusions Significant challenges were revealed in implementing IPC in care homes including staffing skills, education, workloads and work routines. These challenges cannot be properly addressed without resolving the tension between the objectives of maintaining resident quality of life while enacting IPC practice. Repetitive staff training and professional development with parallel organisational improvements have prospects to enhance IPC uptake in residential and nursing homes. Patient or Public Contribution A carer of an older person joined study team meetings and was involved in writing a lay summary of the study findings.

pandemic, the virus may have been responsible for around half of all deaths in nursing or residential homes in European countries. 1 In England, it has been suggested that government policy privileging safeguarding the NHS and hospital discharge practices are possible reasons for the high number of deaths. 2 However, there was less discussion about mechanisms internal to care homes that contribute to the devastating impact of the Covid-19 pandemic, such as staff knowledge and resident behaviour, although a shortage of personal protective equipment, such as masks and gowns, was indicated. 3 It is therefore important to understand the factors that might promote or hinder the spread of an infectious disease like Covid-19 into and within care homes for older people.

| Getting started-Context for the research
Previous studies have examined staff adherence to Infection Prevention and Control (IPC) guidelines, mainly looking at self-reported behaviour through questionnaire surveys. Hand hygiene is one of the most basic strategies in IPC, and a cross-sectional study of compliance in nursing homes found that immediate access to disinfectant materials and role modelling by senior nursing staff were important factors for successful implementation. 4 Other approaches have also been proposed, including national initiatives, such as the use of inspection regimes or specialist infection control nurses. 5,6 Most of these interventions, as well as the bulk of relevant observational studies, had taken place in the United States.
A recent questionnaire study of nursing home staff in Italy found ambivalence and low uptake of influenza vaccination, with 34% of respondents expressing safety concerns. 7 A similar survey in France also found 'hesitancy' around influenza vaccination and recommended 'communication interventions' to improve staff uptake. 8 Pilot searches revealed that such issues had also been explored in more depth in at least one qualitative study, with issues around education and workload highlighted. 9 Staff education and training has been recommended to improve IPC in care homes, although this can be challenging, given high rates of staff turnover. 10 A recent systematic review of the effectiveness of IPC programmes in long-term care facilities by the World Health Organisation (WHO found that monitoring and feedback, in addition to staff education, had also been used, although efforts needed to focus on at least four elements of WHO's strategy (IPC Programmes, Guidelines, Training and Hospital-Acquired Infection surveillance) 11 to control infections. 12 A multimodal approach to improve hand hygiene and use of gloves noted the utility of training packages being contextualized in everyday practice. 13 A Swedish study that set out to examine care home staff knowledge and adherence to guidelines appeared to be hampered by the fact that carriage of bacteria, and thus experience of IPC, was very limited. 14

| Focus of the meta-ethnography
A metasynthesis of qualitative studies in IPC in nursing and residential care homes for older people was conducted. Meta-ethnography was chosen as a review and synthesis method, as it offers the opportunity to develop conceptual insights that go beyond the findings of qualitative studies. 17 The method is akin to a systematic review in quantitative effectiveness studies, although the way in which findings are brought together is more like primary qualitative research in the way that concepts, metaphors or findings 17 used by authors of original studies are systematically organized and compared.
Our aim was to develop interpretive insights into the factors that influence infection transmission in residents of care homes for older people. To achieve this, we set out to identify qualitative studies that would reflect the ways in which IPC is managed in care homes in practice and extract findings that yield insights into the enactment of IPC practices such as isolation, hand washing, environmental cleaning and antimicrobial management. Ethnographic and participant observation studies offer the potential to yield insights into actual (rather than self-reported) behaviour and advance current IPC understanding that is mostly based on selfreported data. Interview or focus group studies around knowledge, perceptions or adherence to IPC guidelines could help form hypotheses about how infection transmission might be either enabled or prevented.

| METHODS
This study report has been structured according to a framework for reporting standards for meta-ethnographies in health research. 18 As originally described by Noblit and Hare, 17 meta-ethnography is a seven-step process. 16 These steps can be understood as approxi-

| Search strategy
The bibliographic databases Medline, Embase, PsychINFO, CINAHL and ASSIA were searched from inception to May 2020 using a strategy with three modified blocks of terms (Mesh terms and keywords) derived from previously published reviews: Care homes for older people, 19 infections (IPC focus) 20 and some simple keywords found to have high utility in identifying reports of qualitative studies. 21 Searches are provided in Table S1A. A number of ad hoc searches were run in Google Scholar, which is considered a good source for identifying grey literature, such as unpublished theses and dissertations.

| Eligibility criteria
We included published reports of studies that fulfilled the following criteria: 1. Participants/setting: Involved residents, staff members or managers of nursing or residential homes for people aged over 60.
2. Studies design: Used qualitative methods of data collection (i.e., focus groups, interviews, observations) and analysis. Mixedmethods reports were included so long as there was presentation of a thematic analysis, or similar, at some point in the publication.
3. Outcome: Focused on IPC practices such as (but not limited to) isolation, hand washing, environmental cleaning and antimicrobial management.
4. Were written in English.

| Study selection
Titles and abstracts were independently double screened for 8% of the results (n = 50) by G. D. W. and S. G., with both agreeing which articles would be included. After establishing this high level of agreement, the first author completed the rest of the title/abstract screening. The full-text screening and data extraction were shared between each coauthor, although the first author completed around 30%, purposefully selecting studies concerning different topics.
Data extraction was completed using a modified version of a previous form used in a meta-synthesis of qualitative studies of patient safety in primary care 22 (see Table S1B). The quality of the included studies was assessed using five fundamental criteria for reporting quality in studies for a meta-synthesis. 23 In most instances, text was copied and pasted from the articles into the data extraction forms, making it harder for primary data and findings to get lost in translation. Each coauthor was assigned at least one study to complete full data extraction and quality assessment.

| Process for determining how the studies were related
The first author read the completed data extraction forms, having previously read the full texts (including the dissertation and thesis) in full, and looked for common issues or theme groups. First, the studies were divided into infection type (e.g., urinary tract infections and antimicrobials; methicillin-resistant Staphylococcus aureus; scabies and 'General Focus'), and tables of evolving theme groups were constructed. During this process, primary quotations from the studies were retained. At all stages of the translation process, groups of studies were analysed chronologically, beginning with the earliest published in each subset. Throughout the study, draft findings were circulated around the study team by the principal worker (first author) and discussed in weekly meetings to agree next steps.

| Process of translating studies
Comparison of these infection-specific frameworks of findings showed no differences, that is, they all spoke to common issues, e.g. around staff workload or relationships with health services. Accordingly, the articles were treated as a whole and a new framework was developed incorporating all studies. At each stage in the process, the translations were circulated to the wider team to garner alternative interpretations of meaning, significance or coherence of presentation.

| Synthesis process
As new iterations brought findings together in different groups, their comparison was used to develop second-order explanations by DAKER-WHITE ET AL. | 2097 either: (i) comparing refutational data within each row of each table or (ii) determining concepts or metaphors that described the contents of the cell or row. In some cases, second-order interpretations were found in the primary study reports, although they are not always found in descriptive studies. 24 The second-order interpretations (whether developed or reported) were themselves compared in each row of the Tables S2-S5 and used to form synthetic interpretations.
In the tables, synthetic interpretations are shown in blue text.

| Patient and public involvement
One experienced public contributor, who is an informal carer, attended our weekly research team meetings and contributed to discussions about refining research questions, searching and selecting studies and synthesizing the relevant data. Together with the first author, the public contributor coproduced a lay summary of the findings and advised authors on the interpretation and dissemination of results.

| RESULTS
Of 656 records screened, 28 full-text articles were initially included and assessed. A further 13 were excluded at full assessment, leaving 15 articles (including 13 unique studies because one study was reported in three different articles) eligible for inclusion (the PRISMA flowchart of the study selection process is presented in Figure 1).

| Characteristics of included studies
All included articles were published between 2007 and 2020. Most of the studies used semi-structured interview or focus groups, were F I G U R E 1 PRISMA chart descriptive in nature and used thematic or content analysis (Table 1).
Two studies, one of which was a doctoral dissertation, included observation of IPC behaviours in staff and residents. Articles reported studies mainly undertaken in North America (n = 8 studies from 10 articles), with others situated in the United Kingdom (n = 3), South Korea and Australia. The participants of the studies were usually nursing or care home staff, but some studies also recruited residents, administrators, leads of nursing or care home facilities and health professionals. It was noteworthy that most of the included articles did not show the demographic characteristics of study participants.
The studies were mainly concerned with types of infections such as methicillin-resistant Staphylococcus aureus, 27,28,33 C. difficile, urinary tract Infections 6,25 and scabies. 32 Some were focused on specific IPC practices (such as isolation, 30 vaccination, 26 antimicrobial management, 29,36 gown and glove use 33 or hand hygiene 34 ). Others had a more generic focus on IPC practice. 9,31,35,37 They were broadly acceptable at quality of reporting assessment, with one considered excellent 34 and one borderline unacceptable. 6

| Outcome of relating studies or study translation
Iterative reading and reorganisation of study findings eventually yielded three coherent theme groups focused on staff motivations and behaviour; the organisation of nursing or residential care homes; and interface with other health care services. Around 50% of the data and findings centred on an essential tension between staff knowledge, behaviour and attitudes set against the challenges of workload and shift patterns (Table S2). Other groups of findings were mainly focused on the barriers and facilitators to enacting IPC in care homes at the individual staff behaviour level (Table S3); the operationalization of IPC in a shared home environment, including resident perceptions (Table S4); and issues at the interface with medical services (Table S5).
To illustrate the process by which findings were analysed to generate interpretive insights, Table 2 represents an abridged version of It should be stressed that several attempts at grouping the findings were attempted before they appeared coherent.
Another column ('conflicting data') presents interview material that somehow countermanded the bulk of the data found. Comparison of data in this way can be useful in a so-called 'refutational synthesis', 17 where findings from different studies appear to be contradictory, although that was not the case in this metaethnography. However, comparison of the data in this way within each 'translation' (i.e., group of findings in Supporting Information Tables) helped derive the second-order 'interpretative findings' of the raw data. In a good qualitative study, such interpretations will be found in the original study reports, but where authors adopt a more descriptive approach, they come from comparison of findings within translations during the synthesis process. Accordingly, this column includes both. Finally, the right-hand column contains the higher conceptual interpretations, which were made by constant comparison of the contents of the rest of the table.
Turning to the content of Table 2, many first-order findings focused on the absence of clinical information relevant to IPC or difficulties obtaining it due to record-keeping or data management systems. Other findings centred on the fact that the necessary information was often located in another organisational entity, such as the hospital pharmacy. Those actors who were needed to formulate diagnoses and treatments, such as GPs, were not on hand and obtaining a diagnosis could present challenges for the priorities of staff on the ground. These issues led to delays in obtaining diagnostic information or treatments. Conflicting findings pointed to a distinction between the real-world intelligence of care home staff and clinicians who could at times be apparently sceptical about the clinical skills of care home staff or the need for treatment. A consequence of this was that treatment could be delivered in the absence of a relevant diagnosis, for example, by a possibly harassed locum doctor operating out of hours. This goes against the principles of IPC, especially in relation to the issue of antibiotic resistance.
Translating the findings into one another led to the interpretations that clinical knowledge in IPC is a contested area that can lead to questions about the credibility of information related to signs and symptoms. The information necessary to enact IPC is hard to come by ('a scarce commodity') and there is reliance on health workers located beyond the control of the care home. Ultimately, the tools necessary for the timely enactment of IPC are 'all off-site'.

| Outcome of translation
An interpretive reading of the completed theme group tables revealed certain domains of concern, including a perceived low-skills base in care assistant staff and a lack of effective monitoring or surveillance systems (Table S2); limits to IPC practice in the care home environment (Table S4); and diagnostic and management conflicts between offsite GPs, for example, and care staff who were perceived to lack training or competence (Table S5).

| Outcome of synthesis process
In terms of explaining IPC practice, the studies largely distinguished between nurses and nursing assistants; between care home staff and medical staff or services; between residents and their staff carers; or between care staff and other staff not involved in face-to-face personal care. A few studies appeared to perceive that poor IPC practice was due to subordinate and poorly paid staff. 9,28,33,35 Although the use of such staff appeared ubiquitous across the studies, it appears as an essential reality of current nursing and residential care home provision. Where concepts have been developed that go beyond the findings of the original studies, by the process described above, the text is shown in blue (Tables S2-S5). It was found that these concepts could be related as a theory of IPC in care homes. The main issues are encapsulated in Figure 2, where it can be seen that the control of IPC is understood to lie outside the nursing or residential care home.
There were two separate issues at play in viewing the control of IPC as something to lie outside the nursing or residential care home (left-hand side of the diagram in Figure 2 The right-hand box in Figure 2 presents a different group of issues concerning the availability or credibility of information critical to patient care including diagnosis, treatment and control. One major problem is that information in relation to IPC is both hard to come by and at times is actively challenged due to communication failures or hierarchical issues.
So far as the care home is concerned, all the clinical resources they need are off site and potentially without control or influence.
A smaller group of issues that did not fit in the synthesized concepts captured in Figure 2 formed a separate set of relationships that explained the variation in staff knowledge and behaviour related to IPC (Figure 3). lack managerial support, some may 'be happy to let colleagues do the work'. 40 This is of concern, given the importance of organisational culture in patient safety in general 41 and in realizing effective IPC practice in particular. 42 A comparative study of frontline care workers in Canada and Scandinavia went so far as to suggest that organisational factors in care home settings set the context for 'structural violence', usually experienced by staff from residents and sometimes on a daily basis. 43 In another publication from the same study, geographical differences in the experience of care staff were explained by different models of care: 'highly differentiated task-oriented work' (Canada, higher levels of violence against staff) versus an 'integrated relational care work model' (Sweden, lower levels of violence). 44 These issues present challenges for the enactment of IPC in different residential care settings and may go some way towards explaining some of the findings in this meta-ethnography.
In other studies, violence or abuse towards care home residents has also been linked to the organisation of care. 45 While the issue of abuse may appear tangential to the focus of this review, it would not be a stretch to argue that deficiencies in IPC could be seen to constitute a form of abuse and IPC is unarguably part and parcel of care quality. More importantly, the kinds of factors found to reduce the incidence of abuse, such as working on the professional development of staff and improving their morale and confidence, 45 are also likely to be effective in improving IPC practice in care homes. While it may seem intuitive to think in terms of education and training, a review found that education alone is insufficient and needs to be grounded in raising the status of care homes and adopting a relationshipcentred approach to IPC, 46 perhaps like that found in Sweden. 44

| Strengths, limitations and reflexivity
The strengths of this study included the fact that independent reliability checks were performed during the searches, published reporting standards 18 were used and that the synthesis resulted in second-and third-order concepts from the primary studies. Meta-ethnography is a form of primary qualitative data analysis applied to reports of qualitative studies. 17

| Implications for practitioners and policy makers
One major implication highlighted in this study is the importance of care homes implementing WHO recommendations on IPC. 11 Training of staff is necessary but not sufficient to improve IPC practice in nursing and residential care homes. Training needs to be embedded within a coherent programme also including guidelines, monitoring and testing. Overall, the meta-synthesis points to the utility of upgrading health care assistants to enhance their clinical responsibilities. This would require significant investment and might be unworkable within the current model of service provision. A conclusion is that IPC is not something that can be attended to in isolation; it requires wholesale attention to fundamental issues in the organisation and delivery of services.
Several studies appear to attribute responsibility for poor IPC to care assistants, nursing staff or GPs for deficiencies in IPC in care homes. This is of concern, given the ways in which the Covid-19 pandemic has shone a light on the largely marginalized status of the care home workforce. 47 A Swedish study found that healthcare assistants in long-term care facilities could detect early signs of infection, 48 and ways might be explored to better harness such professional skills for the furtherance of IPC in residential care settings for older people.
Another implication of the findings is the need to alter staff perceptions that infections may be seen as inevitable in residential care settings. The reasons underpinning these perceptions of infection inevitability and IPC pointlessness in care homes are unclear, but may simply reflect previous negative experiences with IPC in these settings. Behavioural science, organisational support and better safety climate could help towards challenging those perceptions that could act as barriers in implementing sustainable IPC improvements in nursing and residential care homes.

| CONCLUSION
The Covid-19 pandemic is likely to have had a significant impact on the enactment of IPC in care homes. The findings of this study represent IPC practice before the start of the pandemic, but they will be useful for those examining IPC behaviour in care homes during the Covid-19 pandemic and subsequently.