Towards a contemporary social care ‘prevention narrative’ of principled complexity: An integrative literature review

Prevention has become increasingly central in social care policy and commissioning strate gies within the United Kingdom (UK). Commonly there is reliance on understandings bor rowed from the sphere of public health, leaning on a prevention discourse characterised by the 'upstream and downstream' metaphor. Whilst framing both structural factors and re sponses to individual circumstances, the public health approach nonetheless suggests lin earity in a cause and effect relationship. Social care and illness follow many trajectories and this conceptualisation of prevention may limit its effectiveness and scope in social care. Undertaken as part of a commissioned evaluation of the Social Services and Wellbeing Act (2014) Wales, a systematic integrative review was conducted to establish the key current debates within prevention work, and how prevention is conceptually framed, im plemented and evaluated within the social care context. The databases Scopus, ASSIA, CINAHL and Social Care Online were initially searched in September 2019 resulting in 52 documents being incorporated for analysis. A further


| INTRODUC TI ON
Prevention in social care is not new (McCave & Rishel, 2011;Ruth et al., 2015). Early pioneers were motivated to address the material and social conditions that shaped lives and opportunities for individuals, families and communities, as well as ameliorating individual and social hardships. Jane Addams' (USA) work for community development, education and wider social and legislative reform; Alice Salomon's (Germany) focus on internationalism and a holistic approach to social work knowledge; Octavia Hill's (UK) innovations in housing reform for the poor are examples. Throughout the 20th century, the focus on prevention in social care scholarship in western contexts fluctuated (McCave & Rishel, 2011;Rapoport, 1961;Wittman, 1961). McCave and Rishel (2011;229) (Ruth et al., 2015;126). Although there was a rise in prevention-focused papers over this decade, the North American researchers found this was only 9% of all papers published in the selected journals (Ruth, et al., 2015). Set against the complexity of social issues and inequalities, the authors called for a widespread dialogue on prevention (Ruth, et al., 2015;132).
Consideration of complexity in social issues is not a new preoccupation. Rittel and Webber's (1973) framing of 10 properties of 'wicked social problems' attempts to gain a measure of complexity in the realm of social issues. They suggest social problems 'have no definitive formulation', 'multiple causes', can be 'explained in many ways' and 'have no stopping rules'. Unlike 'tame problems' that may be technically complicated but have boundaries and solutions, like building a desalination plant, Rittel and Webber argue that social problems are comprised of multiple dynamic components and defined from ideological perspectives. Capacity to isolate cause and effect is problematic, which is a longstanding philosophical debate.
In Aristotle's frame, as Crane and Farkas (2011;369) write, causes can be viewed as 'giving an account of why something is the way it is'. In Aristotle's own words, a cause is '…that from which (as immanent material) a thing comes into being' (Aristotle, in Crane & Farkas, 2011;380). In discussing causation, the philosopher David Hume distinguishes between Relations of Ideas and Matters of Fact (Hume, in Crane & Farkas, 2011;382) and theorises that it is difficult to discern what the mind sees as 'conjoined objects' with an impression of causation and what is a relationship of cause and effect.
The constitution of relations between cause and effect is a key difference between a public health preventative paradigm and prevention in social care. In the 1960s, Wittman, in crafting a conceptual framework for prevention, contrasts the nature of an illness trajectory with the more fluid nature of social issues dealt with by social work: There are visible difficulties in the adaptation of prevention as it is known in other fields. In public health there is physical intervention, made possible through knowledge of causation or the agents of transmission of a specific illness. In social work there is less that is concrete to work with in terms of illness. (1961; 21) Rapoport (1961;3) in a comprehensive analysis of prevention in public health and social work begins with the view that …the concept of prevention, borrowed largely from the public health model, is often used in a distorted and confusing manner in the social work framework.
She argues, public health employs a 'unifying notion of prevention', whereas social work is based on many concepts and practices, operates in 'complex systems' about presenting issues in the here and now, as well as what is yet to happen and often with incomplete knowledge of causation. The latter, she writes, is one of the social work's '…built-in professional stresses ' (1961; 8). Rapoport contends that shoehorning social work into a model from public health will not iron out confusion unless there is a more precise definition based on social work's purpose, knowledge bases and models. She writes: …social work has major responsibility for amelioration and control, and a vital role in all levels of prevention.
Prevention should be more strictly defined to sharpen professional practice and give impetus to greater activity in the area of primary prevention, which involves the imaginative application of all social work methods in anticipating problems and need. (1961; 12) We report findings from an integrative review of the literature on preventative social care with the following research questions and objectives: What is known about this topic?
• Prevention in social care is commonly tied to threetiered, linear public health preventative narratives.
• Drivers for prevention have been contested, including values-based logic and reduced dependence or council expenditure.
• The conceptualisation and enactment of preventative work in social care varies significantly between localities.

What this paper adds?
• There is a continued reliance on linear, cause-effect models for prevention in social care and limited accounts of the complexity associated with everyday life.
• Developments in complexity theory within social and healthcare sciences offer new perspectives on how prevention is conceptualised and enacted.
• Prevention work will arguably benefit by integrating guiding principles embracing the complexity of service users' and carer's lives.
• What are key debates in prevention work in the context of contemporary social care?
• How, and in what ways, is prevention work conceptually framed in the social care context?
• How are preventative interventions in social care being implemented and evaluated?
In addressing these questions, we confirm the recurrent nature of issues outlined by both Rapoport and Wittman in the 1960s, highlighting a continued reliance on the discourse of the public health paradigm, but alongside other emergent prevention narratives.
Following Rapoport (1961) and Lundberg (2020), the latter writing in the context of public health, we explore the potential for a reconfiguration of how prevention is conceptually bounded and understood in the realm of social care that offers greater consideration of the complexity underpinning it. In framing our discussion, we use a definition of social care as inclusive of social service and welfare supports to meet human needs that are provided by the state, market and households, and are both formal and informal.

| MATERIAL S AND ME THODS
Initially undertaken as part of a commissioned evaluation of the Social Services and Wellbeing Act (2014) Wales, an integrative review of the literature was conducted, allowing for a range of methodological approaches to be included, as well as both academic and grey literature (Llewellyn et al., 2020;Whittemore & Knafl, 2005).
The guidelines and framework of PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-analysis) were followed to identify the set of included documents and to guide the review process (Moher et al., 2009).

| Study selection criteria
Searches were initially conducted within the Scopus, ASSIA and CINAHL databases during September 2019, with limitations in place to exclude references on the following criteria: published before August 2014; not being written in English; being drawn from non-Western countries; and non-journal articles. For Social Care Online, though the search was also initially conducted in September 2019, the same limitations were not repeated based on the database containing legislation, government documents, practice and guidance, research briefings and more alongside journal articles (Social Care Online, 2020). As such, grey literature such as books and reports were included within the search with no prescribed date range in the first instance so as to capture documents that paved the way to the current situation. Any articles meeting these criteria were selected as part of the initial data extraction by the research team.
Study leads reviewed the list of reference abstracts in conjunction with the research team for topic relevance, identifying any further exclusions based on this or the previously outlined criteria. A significant proportion of articles were excluded due to covering prevention of hospital admissions or related issues. Additionally, at this stage, it was decided to restrict Social Care Online documents to be published from 2000 onwards based on diminished relevance. Study leads conducted a further snowball search within key articles to identify any additional reading of relevance to prevention within the context of social care. Articles deemed relevant were marked for inclusion, incorporating those identified through the snowball search. Any articles felt to have limited relevance were put to one side and discussed with the wider team before a consensus decision was made.
Finally, a refresh of these searches was conducted in March 2021, identifying a further set of 14 relevant documents once duplicates and exclusions had been applied. These were analysed as a separate supplementary exercise by study leads to assess whether new information had emerged since the initial searches were performed.

| Literature search terms
Combinations of search terms were entered into the online data- Further detail on search terms can be found in Table 1.

| Data extraction and analysis
Documents selected for inclusion were extracted full-text for thematic analysis by the research team. This process sought to identify any patterns in how prevention was discussed, methodologies used, primary research setting and topic. Several strategies were used to improve the rigour of analysis including multiple iterations of search terms in the initial phases, familiarisation and close individual readings of returned documents / articles and regular research team meetings to discuss the categorisation and interpretation of findings (Ritchie & Spencer, 1994). In practice, this led to the analysis itself undergoing several iterations as primary themes and sub-themes were developed, re-visited and re-developed. After each stage, the research team re-applied the latest iteration of the conceptual framework to those documents already analysed to assess whether new approaches may be beneficial (Whittemore & Knafl, 2005).
This also saw the emergence of hitherto unexpected concepts and research areas, including those of social enterprises and community businesses that ultimately became central to contemporary understandings of the topic. The emergent conceptual framework integrated the key components of the research questions: how prevention is discussed, framed and conceptualised; the manner in which studies of prevention activity have sought to implement and evaluate themselves; and the similarities and differences associated with such processes.

| FINDING S
We commence our findings with a brief description of the returned sample of literature including document type and primary focus.
Following this, we will outline the core themes emergent from the analysis: the drivers for prevention activity; the varying concepts and definitions of prevention; and how the literature shows this to be actualised in the social care context.

| Sample description
From a total of 505 references identified through database and snowball searches, 66 documents and articles were included for review. Figure 1 PRISMA flowchart outlines the selection process: After eliminating duplicates and reviewing against the inclusion criteria, as well as ascertaining topic relevance by scanning titles, 158 articles were extracted for full-text review by study leads.
These were assessed by study leads for relevance based on title and abstract with any exclusions noted for discussion with the wider team. This provided a total of 66 articles deemed to fit the criteria and with relevance to the topic. Table 2 provides the characteristics of the documents analysed.
The final document set predominantly comprised projectspecific research reports (n = 29) and journal articles (n = 24), with the remainder being categorised as guidance or strategy documents (n = 10), briefings (n = 2) or process evaluations (n = 1). Initial analysis also categorised the documents by theme and primary focus, highlighting a range of different, often overlapping, topics being discussed under the umbrella of preventative social care. These primarily included community development (n = 23), community businesses, social enterprises and the wider voluntary/community sector (n = 15), and documents focussing specifically on adult, children, family or older people services (n = 20).

| Drivers for prevention activity
The Cambridge online dictionary defines 'prevention' as an act 'to stop something from happening or someone from doing something'.
Historically, the drivers for preventative policies and practices in social care have been diverse and contested, influenced by ideologies Non-developed and non-Western countries • "prevention" AND "social care and support" • prevention AND social care and support • prevention AND "social care" • "social enterprise development in the UK" • "social enterprise development" • "role of the third sector in prevention and social care" • role of the third sector in prevention and social care • "community development in social care" OR "community development in social services" • community development in social care OR community development in social services • "community based approaches to supporting wellbeing" • community-based approaches to supporting wellbeing • "information and assistance and advice services in local government" • information and assistance and advice services in local government • "family strengthening programmes" • family strengthening programmes • "planning for preventative services in social care" • planning for preventative services in social care and values about what it is that is to be stopped from happening, different theories and knowledge about causation, and how to frame and act on solutions (Clark, 2019;Curry, 2006;Gough, 2013). The contemporary landscape is no different, most recently highlighted by an evaluation of the Social Services and Well-being Act in Wales (Llewellyn et al., 2020). Ambiguity and prevention in social care have remained related (Curry, 2006;Llewellyn et al., 2020;Marczak, Wistow, et al., 2019).
A contemporary preventative agenda is an emphasis on prevention to reduce state expenditure by stopping current and future demands for high-cost services. This agenda is based on arguments about unsustainable social care expenditure (Bown et al., 2017;Curry, 2006;Kerslake, 2011;Wavehill Social and Economic Research, 2019). For instance, Kerslake argues the 'primary goal of any prevention strategy has to be the reduction of future demand' (2011; 14). An example is preventative strategies to reduce falls in the older population which can stop falls-related hospital admissions and associated expenditure (Curry, 2006). A discourse of prevention to save future expenditure can seep into an agenda about budget cuts and the transfer of provision of care and support to the household and community sector. Marczak, Wistow, et al. (2019), in their study of prevention in social care in England, find that economic, cost-saving discourses are prominent in practice.
There are counter viewpoints that prevention in social care needs investment. A review of the implementation of the Care Act 2014 in England identified financial pressures as the clearest driver for prevention activity, but also acknowledged that such pressures were common barriers to sustainability and success (Tew et al., 2019).  Young et al., 2014) and integrated services and programmes across functional areas. An example of a preventative programme using an explicit human rights discourse is a comprehensive child protection model proposed by Young et al. (2014) to combine protective work with the provision of supports and services using community development principles. One of the implications from a reading of debates in the literature reviewed is the need to name and reconcile competing motivations and agendas for prevention in the context of political processes and contestation about the allocation of scarce resources.

| Prevention concepts and definitions
Mirroring this contestation over the drivers for prevention activity, sampled articles offered various viewpoints on how prevention itself is conceived, defined and delivered. Many of these were reliant on pre-existing public health narratives, commonly drawn from a threetiered approach to prevention, that conceive it as a linear, layered and interlinked pathway (Wavehill Social and Economic Research, 2019).
Within this, as both Gough (2013)

| Prevention, social care and context
As suggested by Marczak, Wistow, et al. (2019) and confirmed by other papers in this review, how prevention activity is actualised in the social care context is varied, encompassing person-centred, family-focused and community-led approaches, as well as diverse local conditions, policies, decision-making and funding constraints (Miller & Whitehead, 2015;Richards et al., 2018aRichards et al., , 2018bRichards et al., , 2018c.
Numerous authors highlight how prevention approaches will necessarily vary based on the identified needs, the organisational contexts of implementation and the nature of interagency collaboration. There are, for instance, a range of social care interventions aiming to respond to needs: information or advice services,

| General reflections
The review indicated broad comparability in many drivers and definitions of prevention. Discourse of cost-saving initiatives predominated, alongside a leaning towards a public health paradigm of preventative spheres or levels. Yet the overarching debate on how prevention is, and should be, defined in social care remains contested. This is exemplified by Allen and Glasby's (2010;33) (Lundberg, 2020). To address this issue, work there has focussed on unpicking the current narrative so as to establish, at a theoretical level, where the narrative misaligns with evidenced reality.
Adherence to medication intended to prevent future illnesses occurring is a case in point. Research in this field has highlighted the importance of understanding structural factors as to whether people take their medicine as prescribed (e.g., prescription problems, unclear guidance), as well as individual decision-making and motivational issues (Jackson et al., 2014;Pound et al., 2005). Notably, adherence issues are reportedly more pronounced when conditions are asymptomatic and degenerative, where lived experience of the illness may lead people to believe the medication is unnecessary (Miller, 1997).
Understanding the dynamics of how prevention activity embeds into day-to-day realities for social care will require the translation of theoretical approaches encompassing both structural determinants and interventions, as well as individual agency. Stagner and Lansing (2009) integrated further consideration of the latter in conceptualising prevention in terms of temporality or past and future horizons. This is enhanced by findings from Warin et al. (2015) who noted that prevention agendas concerned with a long-term future time horizon, such as within public health obesity initiatives, may be too distinct from the everyday realities of those they are targeting (2015; 309).
For a prevention narrative to successfully conflate the inputs of both organisational activity and individual response, it first requires a theoretical underpinning better encapsulating both aspects, as well as how they interact together.
Developments in complexity theory within the social sciences may offer helpful insights in this regard. Much like in the realm of prevention, sociological thinking has also wrestled with issues of structure and agency, grand narratives and contradictory particularities. This ongoing debate, as Walby (2007)  When considering prevention activity, the notion of influencing future events is self-evidently crucial. Indeed, the existing prevention narrative, by encouraging activity upstream to avoid it being required downstream, implicitly aligns with this. However, the linearity of this model reduces some of the evidenced complexity associated with how prevention initiatives play out, as well as how different areas of social care may conceive, commission, plan, design and deliver such programmes. This reductionism, therefore, has potential to extend into the presumed purpose of prevention activity as well. Within social care, prevention can extend to issues of child abuse and neglect, maternal and child health outcomes, homelessness, unnecessary admissions to hospital, promotion of well-being, strengthening resilient communities and reducing the need for formal social care services. Each of these can be interpreted with the principles of addressing social inequalities, moral principles and values, or, as reported by various research, cost-saving for the social sector (Curry, 2006;Gough, 2013). Certainly, as Marczak, Wistow, et al. (2019) report, the blurring of these discourses together has the potential to result in ambiguity in both conception and decisionmaking (2019; 210).
As conflicting as it may seem, integrating elements of complexity into how prevention is conceptualised, planned and commissioned may ultimately benefit it with greater clarity. Efforts in this direction are becoming more noticeable in the realm of health care, including the development of the Angel Taxonomy in Wales (Rutter et al., 2019). As Lundberg (2020) (2007) argues that complexity theory, when conceived as a 'set of theoretical and conceptual tools' as opposed to a singular, holistic 'theory', offers a framework by which the age-old issues of structure and agency can be navigated (2007; 456). In order to develop how this may operate in practice, we will demonstrate the complexity associated with two common social service functions mentioned in our findings: (a) support for older people and (b) support for unpaid carers.

| Support for older people
In the context of an ageing population, preventative support for older people has become increasingly central to social and health care services (Wistow & Lewis, 1997). When specifically considering the social care needs of older people, this initially resulted in a prevention model encompassing (a) prevention or delay of the need for care in higher cost, more intensive settings and (b) promotion of quality of life of older people and their engagement with the community (Wistow & Lewis, 1997). Alongside this, there has been an ongoing values-based policy emphasis on issues of 'choice' and 'independence' into older age within Western contexts, apparently incorporating both elements of the model. As Wistow et al. (2003) observe, the perceived overlap between the quality of life in ageing with the potential for cost-saving has resulted in evidenced-based commissioning strategies encouraging older people to remain living in their own homes. Yet the authors also note, that many of these strategies offer high value but not necessarily at low cost (2003; 2).
Thus, the common discursive policy driver of cost reduction, in and of itself, cannot be perceived as a singular, linear narrative -prevention work for older people may result in cost savings, but also requires investment.
Beyond this, though, the pivotal notion of 'independence' itself requires unpicking. Central to much of the prevention policy discourse for older people over the last 30 years, its presence has resulted in a diverse array of programmes including falls prevention, active ageing, re-ablement and adapted housing in order to enable older people to remain living in their own homes (Care Inspectorate Wales, 2020). While there is evidential cause for such initiatives to be in place, as Wistow et al. (2003) friends and neighbours (2003; 4). To conceptually demarcate this new perspective, they suggest a shift away from the conventionally adopted discourse of 'independence' to one of 'interdependence' where older people are seen as both individuals, but simultaneously as individuals within their own specific social context (ibid; 5).
Ultimately, the practical implications of these for policy suggested by the authors resulted in another three-tiered model of prevention activity: individual, community and government (ibid).
However, by conceptualising the interdependence between these layers of prevention work, some necessary elements of complexity were introduced. Though the emphasis remained largely on the upstream to downstream narrative, the mechanisms by which one stream affected the other were shown to be increasingly interrelated and muddied. This point was furthered by their suggestion that in order to address inequalities embedded into successful ageing, there was the need to address pre-existing inequalities between 'individuals and groups over their life course as well as between different age groups' (ibid; 5).
To some degree, this mirrors the recent assertions of Lundberg

| Prevention and unpaid carers
The sheer volume of people who are unpaid carers and the diver- there is an emphasis on support for adults who need care and support for carers, who meet certain eligibility requirements (Fernandez et al., 2020). This is also the case in Wales with provisions in the be complex issues with an aura of insolvability and defeatism (e.g., Peters, 2017). We share such concerns and believe returning to Walby's (2007) idea of complexity acting not as a universal, holistic theory but more of a lens or conceptual toolkit will aid their navigation. Taking this as our starting point we would suggest the principles embedded within a 'principled complexity' view of prevention would be as follows: • Social issues requiring prevention activity are often complex, messy and interrelated; as such they may need multiple, interlinked and dynamic solutions.
• 'Communities' may share many characteristics but differ in many other ways. Attempting to understand the diverse characteristics of a 'community' and the social groups and individuals comprising it, is a prerequisite to effective prevention activity.
• Individuals engaged in prevention activity should be considered within their own specific social contexts which, in turn, each offer their own influence.
• Prevention activity may result in future cost-savings, but also requires ongoing investment and engagement to ensure the consistency of activities on the targeted problems, as well as the emergence of others.
• Understandings of causation emerge in the undertaking of preventative activity and to be alert to this emerging knowledge requires a critically reflective approach.
• Certain life events may increase the likelihood of prevention activity being required but do not universally determine the lived realities of communities and individuals.
• What 'works' for one community or individual may not 'work' for others; ongoing engagement with those targeted for prevention may identify the factors provoking this disconnect.
• 'What' we choose to do as prevention activity is equally as important as 'how' we choose to do it.
Though by no means a definitive list, we hope that by conceptualising prevention outside of existing linear narratives and shifting from the policy discourse of cost-saving we may begin the development of a more nuanced approach in line with Rapoport's call, over 60 years ago, for a sharpened and imaginative preventative practice.
Within this, the aim should be to demarcate social care prevention from metaphors and parables associated with the public health model and embrace the fluidity, non-linearity and dynamics of social life.

CO N FLI C T O F I NTE R E S T
All authors have no conflict of interest to declare.

DATA AVA I L A B I L I T Y S TAT E M E N T
Data sharing is not applicable to this article as no new data were created or analysed in this study.