Objective

This meta-synthesis of qualitative studies examines stakeholders' understanding, beliefs and attitudes towards screening, intervention and care in the context of frailty and pre-frailty, in order to understand perceived or observed value for these healthcare activities, or experiences of individuals in relation to them. It aims to examine views across the range of involved stakeholders: frail and healthy older persons, informal carers, and health and social care practitioners.

This meta-synthesis is conducted in the framework of an EU funded project aimed at reducing the burden of frailty in the European Union, Frailty management Optimisation though EIP AHA Commitments and Utilisation of Stakeholders input (FOCUS).

Background

Frailty is defined as a state of high vulnerability to the risk of adverse events when exposed to a stressor [1], for example, a new diagnosis or exacerbation of a previously stable chronic condition, a fall or a traumatic life event. The frailty process appears to be a transitional state in the dynamic progression from robustness to functional decline [2], during which physiological reserves decrease and become less likely to be sufficient for maintenance or repair of the consequences of conditions. That is, frailty can be seen as the absence of resilience.

There are two main operational descriptions of frailty in the literature, which are both used to define frailty in clinical contexts. The first is a phenotype [3] which defines frailty using physical markers focused on muscle strength/weakness and sarcopenia. The second is that of an accumulation of deficits [4] which is flexible to context but, in addition to physical weakness, includes accumulated morbidities, polypharmacy, cognition, mental health and activities of daily living, although debate continues as to the range of approaches that are appropriate to detect frailty. Within these two frameworks, many tools have been developed to define a person as robust, pre-frail or frail [5,6,7]. The choice of tool can be determined by the purpose of the specific assessment, for example, directed towards an intervention that is to be provided based on screening. Prevalence of frailty in the population aged over 65 varies from 4 to 17% depending on definition and population, with 26% of over 84s being frail [8].

Development of the accumulation of deficits models has underlined the necessity to integrate medical with social domains to really understand frailty, recognising that frailty also defines a social group. Nicholson integrates the concepts of marginal groups and liminality between a possibly active and empowered group (the Third Age) and an invisible, totally dependent and near to death group (the Fourth Age) [9]. According to this theory, the defining feature of this Third Age/Fourth Age conceptualisation is that it constructs the possibility that individuals can be "in transition", without membership of either group. This liminality also represents the idea of fluid space where some form of power can be still experienced, at the same time as vulnerability and dependence [10]. While adding complexity to the view of frailty, the social conceptualisation offers a clear framework to the possibility of resilience and some form of power within frailty.

However we conceptualise it, frailty likely has a huge social impact. There are clear indications that frailty increases the risk of isolation [11], and therefore necessitates the commitment of services to address a difficult-to-reach group. Nevertheless, it is increasingly clear that frailty is malleable [12]. This understanding is timely. It relates to our context whereby European and worldwide healthcare systems are undergoing transformation efforts to move from single diagnosis focus, and systems where medical and social care function in separate silos, to more holistic care which considers the accumulation of effects on a patient and aims for an integrated approach [13].

Frail older people are high users of informal and formal care, and healthcare services [14]. However, studies illustrate examples of older people and their carers refusing services offered [15], or expressing reluctance to seek care even whilst acknowledging that such support is needed [16]. There are also indications that dependency on services, but limited participation in decision making, can produce feelings of devaluation and disempowerment in older persons [17]. It is essential to take into account older adults’ views, attitudes and understandings related to their own coping and care planning strategies, and to care and interventions in the context of frailty.

Informal caregivers receive support in many jurisdictions, but there is a perception that what support they get is based on fixed formulae regarding eligibility for fixed support, rather than their own perceptions of needs or what support would make a difference [18]. We wanted to integrate their points of view as well.

Previous studies on screening and intervention programmes in other health areas have found not only that potential patients are often reluctant to be screened or to follow health advice, but GPs and Practice Nurses (PNs) may be reluctant or unable to implement the intervention as was structured [19,20], suggesting lack of fit with understanding, perceived value, and skillset. An awareness of issues faced by practitioners in implementing care programmes or supporting care of frail elders, including their perceptions of the roles, needs integrating into the knowledge base in this area.

Qualitative research works to produce knowledge from subjective experiences. This can be both necessary and powerful when those points of view are from at risk and marginalised groups whose voices would otherwise not be heard. Frail older people are considered such a group. Qualitative research enables frail older people to express their needs and describe experiences in terms of what matters to them in their everyday lives. These data are then analysed systematically to represent the needs of older adults when evaluating and planning services. Widening the evidence-base in health and social care service development to include qualitative research is especially powerful when assessing acceptability and feasibility of interventions within particular population groups, services and communities [21].

The review teams consists of research centres across Europe with expertise in frailty, care for older people and disability related to mental disorders spanning the disciplines of psychology, health and social sciences, and medicine. We saw a need to review the qualitative evidence on this issue because there were questions about the experience of disability, well-being, frailty and health that were being obscured in more traditional reviews owing to their inherent relationship with services, institutions and power. Prioritising the voices of older adults and those with responsibilities towards them is commensurate with the review team's commitment to a recovery-oriented approach to care for older people that encompasses the concepts of resilience and empowerment in an authentic way.

We have conducted this meta-synthesis in order to build on the qualitative literature of frailty and reach a more encompassing vision of the issue. In doing this, three main ideas have orientated our work. These were: how much frailty is seen as malleable and preventable by stakeholders; the scope for integrated and empowering care; the importance of the knowledge and practices that stakeholders, frail older people in particular, use to cope with the conditions of frailty.

Design

We conducted a meta-synthesis of qualitative evidence because the objective was to make sense of the evidence base rather than simply collate it. An interpretative synthesis method was therefore appropriate [22–24] in order to understand the conceptualisation of frailty through the views, and attitudes of stakeholders. Such evidence is needed to inform the development of frailty interventions that are both feasible and acceptable within the context of health and social care in Europe.

Search strategy

Three electronic databases were searched in July 2015: PubMed, CINAHL and Web of Science. The search strategy included terms relating to frail older people (population), frailty (phenomenon of interest), qualitative or nursing methodology research and mixed methods studies (methodology) (S1 Text; search strategy is shown for PubMed, with the same applied to CINHAL and Web of Science with appropriate modifications).

Inclusion and exclusion criteria

Included studies were: qualitative research; with frail older adults or stakeholders involved in their care, e.g. nurses, allied health professionals, family caregivers; about the conceptualisation of frailty; beliefs about interventions to prevent or manage frailty. We selected qualitative research studies as identified through the MeSH terms and several appropriate keywords in the searched databases (S1 Text, Search strategy). The MeSH terms used in PubMed were "qualitative research", "nursing method research", "focus groups", "interviews as topic". These terms were integrated with keywords searched in title and abstract, including again terms related to the methodology (like "qualitative study" or "mixed methods", and others) and terms related to the topic (like "point of view", "perception", and similar). All the possible terms were used in MeSH and in title and abstract in order to assure identification of all the relevant papers from the searched databases.

We included studies published in the period 2001–2015 in English, French, Spanish, Portuguese, Italian and Polish, which were the languages that at least two researchers in the group could read. Studies were excluded if they did not have frailty as a focus, if they were conducted in specialised communities, e.g. homeless, if they focused on specific diseases, end of life care, neglect or abuse of older adults, or organizational interventions.

Screening

Seven members of the review team (MB, BD, CH, DK, RN, SR, RS) screened records by title and abstract against inclusion criteria. All records were screened independently by two people to check for consistency; disagreements were resolved by discussion within the pair and consultation with another team member if required. Full text papers were further screened, each by two people, to make final decisions about inclusion.

We did not find any paper in Italian, Polish or Spanish. The papers in Portuguese could be read by the Potuguese (JA, EC), UK (RN) and Italian members (BD).

Quality appraisal

Included papers were appraised using the Critical Appraisal Skills Programme checklist (CASP) for qualitative research (http://www.casp-uk.net/; S2 Text). The evaluators (JA, MB, EC, BD, CH, DK, RN, SR, RS) conducted the quality appraisal independently in pairs and any differences concerning inclusion were resolved by discussion within the pair or through consultation with another team member.

Synthesis

The synthesis was informed by meta-ethnography [25–27]. We chose meta-ethnography because we aimed to not only aggregate findings from papers but also to develop a conceptual understanding of what the original studies say. This higher order, synthesised conceptualisation of the evidence base will be used to contribute to best practice guidelines for frailty management and prevention to be developed in the FOCUS Project [28,29]. In developing the synthesis, we referred to the seven phases described by Noblit and Hare [25,27,30]. After having formed specific research questions (Phase 1, Getting started) and defined the material through search strategy, acquisition and selection of papers, and quality appraisal (Phase 2, Deciding what is relevant), all papers were repeatedly read by the authors. This was a continuous process which lasted until the end of the synthesis (Phase 3, Careful reading and re-reading), enabling development of a good understanding of all papers. Following Toye’s (2014) [25] conceptualisation of first- and second-order constructs, we tabulated extracted data as first- and second-order constructs. We extracted constructs verbatim using the theme titles (or subheadings) used by authors. Where theme titles were absent we created descriptors to code the content. If we experienced difficulties in identifying an appropriate descriptor, we consulted the discussion section of papers to help consolidate the authors’ interpretations.

A grid of concepts was then developed (Phase 4, Determining how the studies are related), which served to prepare the reciprocal constant translation of studies. This process was shared with six team members (MB, BD, DK, CH, SR, RS) reviewing all theme clusters to ensure coherence and representation of all included studies. The large initial grid of concepts was refined in order to achieve a manageable and useful synthesis (Phase 5).

The lines-of-arguments synthesis (Phase 6) involved interpretative work to identify major themes, within the synthesized findings. The set of synthetic themes (third-order constructs) were derived from results of the reciprocal translational analysis and discussions within the team to reach consensus.

This interpretative phase of the synthesis was led by three members (BD, CH and RS) and agreed by the full team. The final set of themes was used to develop a conceptual model of frailty. This was achieved through discussion within the analytic team to arrive at a shared understanding of relationships between themes. The reasoning behind each connection was analysed. Themes were moved around into new positions to test out their meaning and inter-relationships. After varying the patterns, a model was developed which depicted a new, more consistent and integrated conceptualisation of frailty, taking into account multiple meanings, and perspectives from different stakeholders.

First-order constructs can be considered in the middle between the raw data and second-order constructs, or even second-order constructs themselves. First-order constructs identified and selected by the authors according to their second-order constructs (Toye et al., 2014) [25] are presented in the results of the synthesis. When these were not given, or given to a very limited extent, we explained the second-order constructs without first-order extracts.

Included studies

A total of 3414 records were identified as potentially relevant and this number was reduced to 1964 after removal of duplicates. Limiting the papers to those published since 2001 resulted in a total of 1566 papers. Titles and abstracts of these papers were read to identify the relevant papers and this process led to the exclusion of 1485 papers. Eighty-one papers were read in full text and appraised for quality. Among these, 36 were excluded: 27 did not address frailty or did not include frail older people or their caregivers and professionals (including seven studies addressing implementation and organization issues); five used methodologies other than qualitative (including one sociological study using sociological categories and conceptualizations); four were very poor quality (did not give information about analysis, gave very limited information about methods, or presentation of findings was not supported by the data) (Fig 1). S1 Table illustrates the main chracteristics of the 45 included studies [9,31–74].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. PRISMA flow diagram.

https://doi.org/10.1371/journal.pone.0180127.g001

Quality appraisal revealed that none was judged as poor. The majority had clear statements of research aims, employed qualitative methods appropriately, had clear conclusions and were judged to make a valuable contribution to the evidence base. Studies performed less well in their reflections about potential impact of relationship between researchers and participants, and lacked full consideration of ethical concerns.

The majority of included studies were conducted in a limited number of Western countries, with a focus on Northern Europe. Most studies recruited frail older adults as their participants (n = 28) either solely or with another stakeholder group; ten papers included informal caregivers. The largest practitioner group was nurses (n = 7), followed by social workers (n = 3), occupational therapists (n = 2), GPs (n = 1), and physiotherapists (n = 1). Only one paper recruited non-frail older adults in terms of examining their beliefs about frailty and frailty prevention. Six papers included more than one stakeholder group, and three included more than two groups. Eleven studies used focus groups, either as the only method or in combination with individual interviews or questionnaire administration, and the others used individual interviews. In one case, interviews were done in pairs of caregiver and professional matched as caring the same frail older person [47]. The most used method of analysis was grounded theory (15 papers), then content analysis (10 papers), phenomenology (four papers), and thematic analysis (three papers), with the remaining papers using other methods.

Synthesis

The synthesis produced four synthetic themes (third-order constructs): Uncertainty about malleability of frailty; Strategies to prevent or to respond to frailty; The capacity to care and person and family-centred service provision; Power and choice. Each theme is presented with example quotations from study participants. The contribution of each paper to the constructs is displayed in Table 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Contribution of included studies to themes.

https://doi.org/10.1371/journal.pone.0180127.t001

Strengths and limitations

In order to ensure methodological quality of our review, the searching and screening process was based on pre-specified criteria, and the assessment of methodological quality of each paper was made by pairs of independent reviewers. Disagreements between reviewers about inclusion were resolved through discussion. The use of quality appraisal tools did not entail any paper exclusion. There are doubts about the meaning of quality appraisal in meta-synthesis, since it is based on methodological and formal aspects and can neglect the more insightful dimension of the paper [82]. This can be a serious problem, anyway we wanted to have an idea of how the included papers performed in face of some criteria.

The synthesis process was shared between team members and decisions about themes were discussed in order to demonstrate transparency between original studies and our synthesis of them.

However, we could not manage all difficulties. The study was restricted to literature existing in three electronic databases. Also, we chose to include only papers in the languages known by the group members, thus excluding papers that could have provided important insights. Connected to this, we acknowledge that the cultural representation of stakeholders in the reviewed studies was limited.

Publication bias could affect our results. It was shown that publication bias in qualitative research can cause a large number of studies presented in conference settings or other contexts remain in a drawer [83]. Looking for these reports and other grey literature could help address this bias. We limited our search to published papers given the high number retrieved. Although not including grey literature could be viewed as a limitation of this work, its inclusion would have made it unmanageable.

The reviewed studies were based on different epistemological models and different methodological designs, with their internal validity dependent on the framing adopted for the research. This necessarily raises the question of the possibility to gather them in a comprehensive synthesis. However, this difficulty is inherent to synthesis of qualitative (as well as quantitative) studies, which are all individually different [84]. To this regard, we agree with Finfgeld (2003) [85] who has argued that the combination of findings from multiple approaches can enhance the ‘truth value’ of the synthesis. Meta-ethnography is a widespread methodology employed to synthesise different and numerous studies, illustrating its suitability to a review data-set like this one [23,25].

Other points are related more to the limits of the included papers. We cannot rule out the possibility of selection biases in the included studies affecting the results of our review. The included studies also presented some methodological weaknesses. Few authors provided information about strategies used to control bias of subjectivity and discussion about saturation of data was frequently omitted. The impact of taking part in the research on the participants was only rarely considered, although some acknowledged that the process itself had useful outcomes [56].

Moreover, the definition of frailty used to select participants was not universally given, and among those who did there was huge variation—from use of a full geriatric assessment, use of a frailty phenotype, to selecting people who received support from a local service or who had been admitted to hospital on more than one occasion in a set period. Nevertheless, this should not have an impact on the findings given that our aim was to examine understandings, attitudes and values towards frailty and interventions and care for frailty.