Translating risk: how social workers’ epistemological assumptions shape the way they share knowledge

Abstract

Social workers are at the heart of drives to improve child health and wellbeing, with knowledge sharing between them and other professionals viewed as a way to reduce the uncertainty associated with this area of risk work. We aim to fill a significant gap in the literature by examining how social workers assess, interpret, filter and share knowledge relating to risk and uncertainty – what we call the translation of risk – within their profession. Based on data from a qualitative study with social workers in England between 2012 and 2013, we identify two main approaches social workers employ. We conceptualise them as 1) reluctant translating, and 2) dynamic translating. Our analysis shows that epistemic assumptions such as how social workers conceptualise the fact/value separation; how they view what we call ‘grey evidence’; and how they understand the relationship between objectivity and subjectivity, underpin how social workers translate risk. We add a new dimension to the literature on risk by arguing that we need to pay attention to the epistemological values that underpin ‘client-facing’ risk work. Thus, we aid understanding of not only how knowledge is shared in particular ways, but also why this is the case. We identify reasons why some social workers include valuable ‘grey evidence’ and prioritise adequacy over accuracy in their translations of risk. We highlight, however, that through an over-emphasis on accuracy and boundaries, evidence-based practice might end up driving out ‘grey evidence’ and inadvertently hampering effective decision-making, judgement and knowledge sharing on risk.

Keywords:

• Risk
• uncertainty
• social work
• evidence
• epistemology

Introduction

A central aspect of what health and social care experts ‘do’ involves assessing, interpreting, filtering and sharing knowledge relating to risk and uncertainty. We call this the translation of risk. In relation to child health and wellbeing in England, the translation of risk is all the more important because information sharing between health and social care agencies is framed as a key way of addressing the uncertainty related to professional judgements about which children are considered to be at risk of harm (Lonne et al., 2009; Thompson, 2016). Although researchers have explored the social, moral, psychological and organisational factors involved in multi-agency knowledge sharing (for example, Reder & Duncan, 2003; Taylor & White, 2001; Thompson, 2016; White et al., 2015), less is known about how social workers share knowledge relating to risk and uncertainty within their profession. We take a social constructionist approach to knowledge and draw on the sociology of risk and uncertainty and translation studies research to provide new insights into the complexities of knowledge sharing within social work. We argue that in addition to acknowledging risk, trust, morality, hope, emotion and intuition (Brown & Calnan, 2013; Douglas, 1990; Lupton, 2013; Zinn, 2008; 2016) as potential aspects of the response to uncertainty, we need to pay attention to the epistemological values that underpin the translation of risk by health and social care workers. We do this by exploring qualitative data collected via semi-structured interviews (n = 34) and research diaries (n = 12) among social workers working with children and families between 2012 and 2013.

Social work and the evidence-based practice response to uncertainty

Child and family social work is at the heart of attempts to improve child health and wellbeing in England. Although the definition of child health and wellbeing has broadened in England in the last 20 years, the language of child protection dominates (Parton, 2016). Hilgartner (1992, p. 41) uses the term risk object to describe how ‘things, situations or activities … are seen as a source of danger’. Child abuse can be seen in these terms; our definition of what constitutes harm to children is constantly changing and is heavily influenced by the medical or disease model (Hacking, 1991). Presently, there are four prominent categories of child abuse: physical, sexual and emotional abuse and neglect (Devine & Parker, 2015). Child protection policy is reliant on the assumption that the greatest risk to a child’s health and wellbeing comes from parent/carer intentional child abuse: in this context, the blame for harm to children is placed on individuals and institutions rather than wider social, political and cultural factors (Featherstone et al., 2016), and children are viewed as individual units separate from their family (Featherstone et al., 2014).

Child and family social workers operate in so-called ‘client-facing’ contexts, where they make decisions about children and their families’ lives based on their assessment of the extent to which a child may be at risk of harm. Social workers are thus ‘in the business of risk work’ (Stanley, 2018, p. 104). Here, our emphasis is on the everyday interpretation of risk rules and guidance as applied to the individual level (Brown & Gale, 2018a), that is, how social workers identify and classify risks and, importantly, account for the ways in which they have done so (Dixon-Woods et al., 2009; Horlick-Jones, 2005). Shaping how such risk work is carried out is the evidence-based practice approach to knowledge, which is based on an individualised, risk-orientated response to uncertainty (Kemshall, 2016; Webb, 2006). Underlying evidence-based practice is a technical-rational approach to knowledge where fact and value are separated, leading to only some forms of knowledge being viewed as legitimate (Webb, 2001). Social workers are thus expected to appraise and apply evidence (in the form of research) to the situations they come across when working with children and their families, whilst limiting less tangible knowledge such as gut feeling (Nevo & Slonim-Nevo, 2011). At the same time, social workers face the fear, blame and emotional politics that shape the media and wider society’s understanding of what it is to be a social worker (Stanford, 2010; Warner, 2015). Such understandings are usually based on media reports of tragic child deaths, with subsequent public inquiries into their deaths shaping policy and generally lacking reference to the social, cultural and political conditions in which they took place (Lonne et al., 2009).

At the heart of the evidence-based practice approach is a drive to improve knowledge sharing between professionals in order to reduce risk and uncertainty in relation to children’s health and wellbeing (Thompson, 2016). In this context – what Brown and Gale (2018a) call the communicative aspect of client-facing risk work – the effective translation of risk emerges as central to the assessment of the health and wellbeing of children and their families. We use the term translation of risk to describe the interpretation, assessment, filtering and sharing of knowledge relating to risk and uncertainty social workers engage in. To clarify, the effective translation of risk between experts is of course important irrespective of the dominance of the evidence-based practice approach. As we outline below, it has become even more central within an evidence-based practice context because sharing knowledge, in and of itself, has emerged as a response to child health and wellbeing.

Knowledge sharing and child health and wellbeing

Linking child health and wellbeing to effective knowledge sharing between health and social care agencies is not new; recommendations for multi-agency working in England can be traced back to 1950 (Thompson, 2016). Since the death of seven-year-old Maria Colwell in 1973 and the emphasis from the subsequent committee inquiry on a lack of cooperation between professionals, multiple public inquiries into non-accidental child deaths continue to conclude that a lack of multi-agency knowledge sharing contributes to failures in these one-off, tragic events (Thompson, 2016). The act of knowledge sharing has thus become central to health and social care with children in England and many other Anglophone countries to the extent that it has been described as a ‘moral and political imperative’ (Thompson, 2016, p. 11).

In this context, there have been attempts to formalise the mandate for knowledge sharing through legislation (primarily the Children Acts 1989 and 2004) and the Working Together guidance (Department for Education, 2018) for health and social care professionals on how to cooperate to protect children at risk of abuse (Thompson, 2016). Further, Munro has included ‘good information sharing’ in a set of nine ‘risk principles’ for child protection, the benefits of which include ‘reduction in crises through taking earlier effective action’ (Munro, 2019, p. 129). Taking these developments into consideration, Carson and Bain (2008) assertion that whether or not to share knowledge with other health and social care professionals can be described as a risk decision in and of itself remains relevant. This is particularly important when it is acknowledged that there is no such thing as a ‘single truth’ about a child and their family, nor that a fixed, ‘full picture’ of a child and their family’s life can ever be reached (Taylor & White, 2001; Thompson, 2013).

Which forms of knowledge social workers privilege over others is, of course, shaped by the risk-orientated, evidence-based practice context outlined above. Forms of knowledge identified as being more likely to be hidden include moral judgements (Taylor & White, 2001); emotion (Thompson, 2016); and what Bessant and Broadley (2016) call ‘uncomfortable knowledge’, where professionals know that ethical principles are not being followed. These forms of knowledge – particularly emotion – are thus ‘filtered out’ of multi-agency knowledge sharing processes (Lees, 2017; Thompson, 2016) and social work documents (Huuskonen & Vakkari, 2015). Due to the focus on the complexities of multi-agency work in practice, policy and research, there is very little research on knowledge sharing within child and family social work – social worker to social worker, within the same team. For example, although Helm (2017) comments on communication within a social work team or agency, Helm’s focus is social work decision-making or ‘sense-making’ more generally. Another exception is the work of Huuskonen and Vakkari (2015) which focuses on the written recording of information in social work settings.

Further, in the literature on multi-agency knowledge sharing in health and social care, there tends to be an assumption that communication problems occur only between different agencies or types of workers (such as GPs, health visitors and so on) rather than within such groups. While these assumptions have been challenged by work on information filtering in child and family social work outlined above, there remains a tendency to subtly reinforce a distinction between what Lees (2017) terms ‘facts’ and ‘feelings’ and suggest that all social workers act in similar ways. The resulting gaps in research which we attempt to address in this article has two dimensions: 1) the complexities involved in sharing knowledge within child and family social work; and 2) the way in which social workers’ epistemological assumptions inform how they share knowledge with colleagues about risk.

Methods

To explore how social workers share knowledge about the messy, complex reality of the lives of children and their families, the first author conducted 34 semi-structured interviews with social worker, and 12 of these 34 social workers completed research diaries reflecting on their day-to-day experiences. Data were collected between 2012 and 2013. Although there have been legislative changes to social work in England since this time, there remains an underlying risk-orientated approach to uncertainty, as well as an emphasis on knowledge sharing as a response to child health and wellbeing in and of itself. Our findings, therefore, remain relevant.

Social workers were recruited from across England rather than from a particular organisation, locality or institution. Defining social work is hard – defining who is ‘doing’ social work is, perhaps, even harder. Social worker is a legally protected title, and child and family social workers are most closely associated with statutory child protection, although child and family social work, in fact, encompasses a spectrum of ‘child protection’ and ‘family support’. Social workers and family support workers who held a ‘caseload’ of children and their families, or who managed colleagues who did, or both, thus took part in the semi-structured interviews. Of the 34 interview participants, eight were family support workers (two of whom had a social work qualification) and 26 social workers. Of the 12 who completed a research diary, four were family support workers (one of whom had a social work qualification) and eight were social workers. As such, we did not reproduce the artificial separation between child protection and family support. For ease and clarity, we refer to this group as ‘social workers’ throughout. Length of experience and level of seniority were prioritised when combining purposive and convenience sampling and, later, snowball sampling. Of the 34 participants, 11 professional acquaintances were contacted and recruited by the first author (who has a social work background), the rest by snowball sampling.

The interview schedule was developed to explore social worker accounts of knowledge sharing, both orally and in written form, within their own teams or agencies. Two interview schedules were prepared to reflect the differences between those with and without managerial responsibilities. Both schedules consisted of three sections: before meeting a child and their family for the first time, during and after meeting the child and their family and discussions with others. Questions focussed on the mode of knowledge sharing (telephone, in-person, and so on), how that information was interpreted and acted upon, and decision-making relating to sharing knowledge with others. Social workers were also asked to define terms that are often taken-for-granted (evidence or gut feeling, for example). Asking them to think of their own (anonymous) examples of practice rather than providing ‘exemplars’ for discussion helped promote engagement with tacit knowledge and understandings (Keddell, 2011).

The study is an exploration of social worker accounts of risk work, rather than an observation of it. This has limitations as well as benefits. Limitations include not knowing whether the approaches social workers described were used in practice and relying on memory for accounts of past events (Foddy, 1993). An advantage of exploring accounts is that social workers could reflect in a space away from their workplace and were thus reassured regarding confidentiality, something many of the social workers highlighted as important. Further, this strategy enabled exploration of the meanings behind actions, rather than the actions themselves. This was crucial, because the question of why, in relation to knowledge sharing, as well as what, was a priority. The social workers knew that the interviewer (the first author) had experience of social work with children and families. This meant the interviewer occupied ‘the space between’ (Dwyer & Buckle, 2009) the insider/outsider dichotomy. Whether the interviewer was viewed as an insider or outsider, and the related power dynamics, were constructed in new ways in each interview.

Research diaries provided examples of how social workers shared knowledge in writing without the need to access confidential information (neither interviews nor research diaries referred to any identifying information about children and their families). The research diaries asked social workers to record what happened and their thoughts on three scenarios for the same child (or children) and their family: 1) after completing a home visit, 2) following a discussion with a colleague or manager, and 3) after writing a report/written piece of work. Although only 12 out of the 34 social works who took part completed the diaries, the diaries proved to be a fascinating source of data and were helpful in terms of crosschecking, contrasting and comparing data from the interviews.

Data were analysed using a form of Timmermans and Tavory (2012) abductive analysis. This model of data analysis is employed in an attempt to overcome the limitations of both deduction and induction by constructing analysis ‘from a grounded theory foundation to foster theoretical innovation’ (Timmermans & Tavory, 2012, p. 169). Practical strategies from grounded theory were used, including coding, memo writing and constant comparing (Hodkinson, 2008). Emphasis was placed on both identifying the ‘surprises’ in the data and alternative casing, which presume the researchers have knowledge of a wide range of literature from different fields (Timmermans & Tavory, 2012).

The study fully adhered to university research ethics guidelines and the Code of Ethics of the British Sociological Association. No confidential information about children and their families was accessed during the study, and pseudonyms are used to refer to social workers. All potentially identifiable data were removed at the transcription stage.

Findings

In the course of their work, social workers visit children and their families at home. They often visit families on their own, choosing whether to write notes during the visit, straight afterwards (in their car, for example) or at a later time in the office. A record of any visit (usually referred to as a case note) must be recorded on a local database, to which colleagues within the same team have access. This is not a straightforward process. During and after home visits, social workers must assess and translate the risks they identify into a formal social work language – language that is evidence-based, risk-focused and can be picked up and understood by colleagues with whom they work. Such translations are carried out in case of notes, in formal reports (for example, court documents), in supervision meetings or in their conversations with other social workers. It is worth noting that translations of risk in reports are usually read and acted on by other social work colleagues who may never meet the family nor the social worker who has written the report. It is expected that risks be translated in a way that is comprehensible to those who only read the reports on the database, perhaps years after the event. Multiple translations are, therefore, necessary and need to be spatially and temporally accessible. Such translations of risk are also expected to be able to stand cross-examinations in court – should this ever be necessary. In other words, the translation of risk is central to what social workers do, and involves a transformation of informal interactions within a family home into a formal account which will remain on the database. Our research identified two main approaches social workers used when translating risk knowledges to their colleagues. We coined these two approaches as ‘reluctant translating’ and ‘dynamic translating’.

Reluctant translating

As stated above, we use the term translation to refer to the strategies involved in sharing knowledge about risk with their colleagues – what to share, what not to share and how to share it. Our research revealed that some of the social workers viewed knowledge sharing as straightforward and a means to an end. These social workers did not view their own knowledge sharing as particularly strategic and wanted to avoid making decisions about what to share and how to share it – or did not think this was necessary. Therefore, we coined this type of translation as ‘reluctant’ and social workers who deployed it as ‘reluctant translators’. Of the 34 social workers, 15 reported using what we term ‘reluctant’ translation strategies. They had varying lengths of experience, from a matter of months to over 15 years, although only one held managerial responsibilities. It was thus the nature, rather than the length of their experience that appeared important in relation to the way they shared knowledge. Social workers who were in this group reported employing a variety of strategies and tactics to overcome problems and difficulties associated with the translation of risk that they identified.

First and foremost, the social workers in this group tried to record ‘everything’. That is, they saw it as valuable (and possible) to share all the knowledge they received about the children and families they worked with. This was viewed as particularly important in relation to written recording of home visits, which could be read by other colleagues. For example:

I write notes when I’m on my visits and generally I key point everything. I document everything. Whatever’s written in my book goes in … So, some records can be quite long. But I think it’s only fair, otherwise you could leave something out that was important and something you could come back to, a thing you thought wasn’t important is, and is the key, the key to everything.

(Olivia, interview)

As this was the first visit to the family, the case note was very comprehensive, listing everything I’d seen and heard.

(Stuart, research diary)

For social workers in this group, recording and reporting ‘everything’ was both possible and desirable. This was either because they thought it was ‘too early’ to know what knowledge to exclude or include or because that was their strategy regardless of how long they had known a child and their family. Some of the social workers in this group mentioned the possibility that their case notes might be used in court in future, with Ava reporting that she would include the same knowledge in the same way, regardless of who was receiving it:

It [recorded information] stays the same. Because I think if ever I had to go to court for whatever reason it’s in black and white there in my [case notes]. And [it’s] exactly how they’ve said it, how they’ve done it.

(Ava, interview)

What connected all reluctant translators was the idea they could share everything, ‘exactly’ as it happened, and that this was desirable – at least in some circumstances.

Social workers were required or instructed to translate the knowledge they had acquired in some of the reports they had to write, translating what they saw and identified in their visit into a formal ‘social worker language’. In such cases, reluctant translators reported that they relied on their managers to ‘pick out’ what was important as they often found it difficult to carry out this complex task. . Despite this acknowledgement that managers engaged in the task of identifying what was important, reluctant translators not only aimed to simply share ‘the full picture’ but also expected other social workers to share everything as well. For example:

It’s difficult because when you’re reading somebody else’s case notes you don’t get that full picture of what the worker knows about that family … because you don’t have time to write it in the way that – and it’s about how workers write that story for that family and each worker’s different and they kind of hold it as maybe a feeling and an understanding because they’ve got a much more broad understanding of that, and if you sat down and spoke to them about [it] but you don’t very often get that luxury.

(Jemima, interview)

This ambivalence – expecting themselves and others to share ‘everything’, yet relying on managers to pick out what was important – was a key feature of reluctant translating.

A closer analysis of those who were reluctant translators, in fact, revealed that their hesitancy to prioritise and sift – to translate – was in fact underpinned by two interrelated values they held. Firstly, they privileged facts and objectivity, and secondly they made a strict distinction between facts and values. For them, privileging facts and objectivity meant not only refraining from colouring facts with views and prejudices, but also ensuring that they communicated all the facts, without leaving any behind. They were reluctant to translate values and emotions as this would take away from the facts of the matter. Their readiness and willingness to share ‘everything’ was linked to these two epistemological priorities and values.

The aim to share everything, however, created various tensions for them in the work place. For example, it led to a conflict arising between what they aspired to, and what they could deliver. There was on the one hand the ideals they held and on the other hand, the multiple, often conflicting day-to-day demands of their role that required them to sift when translating. It also led to tensions arising between what they thought was the ideal form of sharing and what others were willing to, or could, deliver. Interviews with them revealed that they were frustrated with these contradictions. They were annoyed about both their own limitations and the limitations of other colleagues – ‘you don’t get that full picture’.

Although the social workers in this group reported using similar strategies for sharing knowledge, there were some differences in the extent to which they separated fact and value, with the social workers in this group either making a straightforward binary divide between fact and value, or acknowledging the uncertainty inherent in their work yet still privileging fact over value, particularly in more formal knowledge sharing spheres (such as report writing). The important point is that all the reluctant social workers viewed fact and value as separate, which underpinned the knowledge sharing strategies they all reported employing.

Dynamic translating

Our research identified that 19 of the social workers took a radically different approach to translation. We coined their approach as ‘dynamic translating’ and the social workers who employed it ‘dynamic translators’. Again, these social workers had been practising for varying amounts of time, from less than three years to over 15 years, but tended to hold more senior roles than reluctant translators. Eleven out of 12 social workers with managerial responsibilities reported using this approach. The level of seniority rather than the length of their professional experience thus appeared significant in terms of how they shared knowledge. This second group took a ‘less is more’ approach, viewing this as a more effective way of assessing and translating the complex interactions and events they identified in the lives of children and families. Whilst reluctant translators viewed sharing ‘everything’ as a worthy aim, and saw sifting and sorting as impediments to the objective sharing of facts, dynamic translators viewed sharing too much as undesirable. When asked about the process they undertook when sharing knowledge, many of them described using an explicit filtering process when translating the risk they identified into social work language both in their reports and in their conversations with other social workers. As Simon stated:

It’s about what’s appropriate, it’s about what’s relevant … and that’s about your filtering.

(Simon, interview)

This process involved not only choosing how to share knowledge, but also to share less often in order to add weight to their professional judgement. This was especially important when sharing knowledge with managers:

So, I’m not one to just panic or ring up at the slightest thing; I’ll try and get as much information and sort of use my own judgement to some extent … you can’t go running to them every five minutes for every little thing …

(Gwyneth, interview)

In fact, they viewed it as problematic if they shared too much, with Gilly describing this as her ‘weakness’. This skill – deciding when it was best to share, and when not to – was important also because managers were not always available.

We also identified that, in specific contexts, dynamic translators held a desire not to ‘tell a story’ but to develop an analytical approach to understanding a child and their family’s life. This, they argued, was how risks needed be translated to other social workers. Dynamic social workers used the term ‘narrative’ or ‘story’ in a negative sense to reflect what some of their colleagues did. They viewed the inclusion of superfluous detail not only as unnecessary but also more importantly as a hindrance to the effective translation of what was happening to a child and their family:

The best [case notes] we would like not to be [a narrative]. But in reality, case notes still tend to be a narrative about what has happened.

(Emily, interview)

Right. What are the issues; I don’t want the story.

(Lesley, interview)

This does not mean that dynamic translators rejected the use of narrative in social work entirely. Rather, this group were of the view that, in contexts where, for example, social work managers needed to understand a situation quickly in order to make a decision, the narrative approach was unhelpful – Lesley did not want the whole ‘story’, for example, and for Emily, the best case notes were not narrative accounts of events. Dynamic translators thus conveyed how central an effective translation approach was to helping them understand the knowledge they received, as well as the confidence to challenge pre-existing filters. For example:

Sometimes [when] meeting with your managers they’ll say, ‘there’s not so much a concern with this [child and their family]’ … sometimes you feel that some of them have made a decision before you’ve even gone out, like ‘oh this needs a quick assessment, can you close it?’ type of thing till you go out and start assessing it and think right, it doesn’t.

(Naomi, interview)

A closer analysis of dynamic translators revealed that their willingness and ability to translate were underpinned by being able to include or exclude subjectivity and analysis ‘if and when necessary’. This allowed them to reconcile more effectively the complex demands of the day-to-day of challenge of converting complex, uncertain situations they came across into formal social work language and thus reconciling their work with the formal rules and guidelines they were expected to follow. Thus, the tension reluctant translators felt between the demands of the job and what they deemed as ideal was reduced for dynamic translators. For example, Wendy acknowledged the subjectivity and inherently value-laden nature of her work, but explained that she was able to translate this knowledge into something more ‘suitable’ for more formal settings:

You can use feelings and we actually use feelings in court reports and things like that [for example] when we went in [to a home], we had a sense that somebody had just been in there, the children looked nervous, there’s all that you can use and evidence your gut feeling, but when a social worker’s doing an assessment you’ve got to prove that your gut feeling is right or wrong … Because I think it can be quite powerful if it’s used correctly. It’s all the wording of how you use it, if you use it factually then it can be quite powerful. If a social worker has used it in a value base, then it’ll just say that you’re using your own values to make that fit, but if you are saying these were my feelings and this is why it made me feel like that, then you’re factually putting why you felt like that.

(Wendy, interview)

This was also evident in the research diaries social workers completed as part of the study. In her diary, Naomi referred to ‘a feeling’ or what can be seen as a more tacit form of knowledge she developed in her visits and translations. She was comfortable explicitly acknowledging such informalities in a more formal context (the research diary):

I sensed from [the mother’s] presentation that she had something to hide and [I] therefore insisted we went into the home to check on the welfare of the children.

(Naomi, research diary)

As we show in the quotations above, Naomi and Wendy had slightly different approaches to incorporating gut feelings into more formal settings, with Naomi being more explicit about her use of feelings and Wendy preferring to translate feelings into something more ‘factual’. The rest of the dynamic translators took one or other of these approaches, meaning there were slight differences within this ‘group’. What we would like to emphasise here, however, is that dynamic translators, in contrast with reluctant translators, shared such knowledge in their translations.

Overall, dynamic translators had an explicit, ‘less is more’ approach to the translation of risk compared with reluctant translators; they were of the view that this added weight to their assessments and translations. Dynamic translators also attempted to retain an element of uncertainty and subjectivity, albeit in slightly different ways.

Discussion

Below, we examine some of the conceptual interventions the research makes in terms of rethinking how experts (in this case social workers) translate risk and also outline the limitations of the research. We examine the broader consequences of the research on the literature on risk, and especially on the translation of risk between experts. Although we are careful not to make unwarranted generalisations from our sample, we argue that the three analytical points we offer are portable to other settings (Polit & Beck, 2010) where both evidence-based practice is heavily promoted and the translation of risk between professionals is a core part of their risk work.

The nature of the data means that we cannot be sure that the strategies social workers described using when sharing knowledge were used in practice. The number of completed research diaries was relatively low, although the resulting data led to important insights due to the methodological triangulation between knowledge shared in different contexts. The strengths of the research include the way in which interviews and research diaries enabled reflection on not only how knowledge is shared, but why, thus aiding a significant contribution to the literature. We did not anticipate that epistemic assumptions would underpin risk translations, and, in part, the ability to ask questions without focusing on the work of a particular organisation or area of the country gave social workers the space to reflect deeply on their work.

Epistemic assumptions underpin risk work

Our research underlines that as well as moral judgement, hope, trust and emotion, the epistemic assumptions of experts also underpin risk work. Our research also identifies that different epistemic assumptions lie beneath different risk translation strategies. In other words, our research not only identifies how different social workers translate risks differently, but also reveals, through interviews and research diaries, the reasons for those differences. In this way, we extend Brown and Gale’s conceptualisation of the communicative aspect of client facing risk work (Brown & Gale, 2018a) and their model of risk work (Brown & Gale, 2018b), highlighting key tensions and differences in terms of not only how risk knowledge is shared, but why it is shared in particular ways.

Differences in how risks are translated overlap with the specific epistemic values and priorities social workers hold, a surprising finding that we had not anticipated. Reluctant translators wanted to translate ‘everything’ (everything defined as ‘all the facts’) and privileged facts over values. They made a sharp distinction between fact and value, privileged objectivity over subjectivity and had what we deem as a ‘firm evidence’ focus. Reluctant translators acknowledged that their managers were able to prioritise or filter knowledge provided to them, but (understandably) did not have the confidence to do this themselves. There was thus a striking difference between what they expected from themselves and others (share ‘everything’) and reliance on their managers to pick out the salient points from their discussions. These priorities should be understood in relation to reluctant translators’ relative junior position, and the broader context in which they were working. Instead of empowering social workers, the technical-rational approach to knowledge promoted by the evidence-based practice model seemed to minimise the capturing and translation of ‘grey evidence’. We coin the term grey evidence to refer to a type of evidence that includes uncertainty, subjectivity and ambiguity, and hence is less explicit. The uncertainty and subjectivity which are often excluded in standard, ‘firm’ evidence is included here, purposefully and selectively. Grey evidence can emerge from tacit knowledge which, following Lam’s (2000) typology, is embodied and embedded in organisational routines and norms, and which is inherently social, rather than simply an individual experience of a ‘gut feeling’ (Collins, 2010; Mitchell, 2016). As such grey evidence is difficult to verify, confirm and substantiate. Yet it is evidence that is imbued and infused in rich descriptions and is thus highly valuable.

Additionally, the fear and blame associated with the media and wider public’s understanding of what it is to be a social worker are not conducive to learning and reporting ‘grey evidence’. That reluctant translators attempted to ‘share everything’ could be read as a form of defensive practice via record keeping identified by Warner (2006), where social workers write lengthy case notes, for example, to defend themselves against an imagined future where they may be blamed for harm to a child. Whilst it is possible that some social workers were sharing defensively with future courts in mind, our findings suggest that there were differences between social workers which needed to be explained: that social workers’ sharing practices depended primarily on their underlying epistemic assumptions, thus furthering existing understanding of how social workers share knowledge.

Such epistemic assumptions were not the forefront of reluctant translators’ minds. Rather, it was through detailed assessment of the data that we could identify how the epistemic assumptions they held were the basis for the translation strategies they described. Dynamic translators, on the other hand, had explicit translation strategies. They were of the view they could not translate ‘everything’, and they acknowledged the act of translation. Further, they expressed a level of frustration with narrative accounts they received in case of notes, for example, thus highlighting that attempts to ‘tell a story’ or include ‘everything’ hindered rather than helped them understand the knowledge they received. In their translations, dynamic translators were able to retain subjectivity, uncertainty, gut feeling and used rich descriptions – what we call ‘grey evidence’. As they did not make sharp distinctions between fact and value, they could deal with blurry lines and articulate grey evidence in their translations. In other words, they did not try to ‘bury’ grey evidence.

It is important to recognise that just because dynamic translators attempted to retain the subjectivity and uncertainty of children and their families’ lives when translating risk, it did not mean that they achieved this on all occasions. Rather, it was their stated aim. Within a risk-orientated approach to uncertainty, evidence-based practice reinforces a hierarchy of knowledge (Gray et al., 2009) and demands a distinction between objectivity and subjectivity, facts and values, and evidence and gut feeling. This means that the ability of social workers as experts to retain and translate uncertainty and grey evidence become more limited. We argue that a possible consequence of evidence-based practice is that an important element of social work expertise – the ability to retain and translate grey evidence – is at risk of being erased through the prioritisation of a narrow definition of evidence and objectivity. Ironically then, the dominance of evidence-based practice can potentially erase another type of important evidence – grey evidence. We argue that the retention and translation of ‘grey evidence’ is extremely important within the same profession, never mind its lack causing communication problems between different agencies or different types of workers, such as GPs, health visitors and so on. We argue that the systemic filtering out of a form of knowledge used by professionals – that is the erasure of ‘grey evidence’ from their practice – has the potential to distort judgement and hamper good decision-making about child health and wellbeing.

Excluding knowledge: through loss, there can be gain

Traditionally, there has been a tendency to view the act of knowledge sharing, particularly in relation to the assessment of child health and wellbeing, as inherently positive. In more recent years, it has been identified that emotion is more often excluded from formal social work spheres in order to better fit existing professional knowledge systems. This body of literature has focused on what is ‘lost’ in this process. As translation scholars show, translation is in fact always a site of gain and discovery (Bassnett & Trivedi, 1999; Mukherjee, 1994) as well as one of loss and erasure (Venuti, 1995). Whilst it is important to recognise losses and gains, we wish to go beyond this simple dichotomy and argue that, at least when it comes to the translation of risk in social work, loss can, in fact, become a gain.

As discussed in the findings section, reluctant translators attempted to follow the epistemic goal of sharing ‘all the facts’. This was understandable, considering their relative junior position, the emphasis on technical-rational knowledge in the evidence-based practice model, and the fear and blame associated with child and family social work by the media and broader public. Dynamic translators, on the other hand, described the positive consequences of filtering out ‘the story’ and being clear and concise when translating risk. Moreover, they viewed the inability to translate and ‘being wordy’ as problematic. We recognise that for dynamic translators, there was an understanding that in translating risk, certain losses in fact lead to gains. In our study, dynamic translators tended to be more senior than reluctant translators. Our findings thus partially reflect findings in Huuskonen and Vakkari (2015) research, which shows that more experienced social workers tend to write more filtered case notes. In this Finnish study, it was the level of seniority that was associated with providing dynamic translators with an ability to make decisions about what knowledge to exclude and include, and how and when to share it – in other words, how to translate risk. Our research nuances these arguments by highlighting that it is not just seniority but also different epistemic values which allowed dynamic translators to carry out translations which enabled gains to be achieved through loss – that is, through a process of exclusion.

Prioritising adequacy over accuracy

Social workers perform their role in a social and political context that privileges sharing. In practice however, there exists a subtle friction between this imperative and the practical needs of social work. On the one hand, there is the pressure to share knowledge, with the associated idea that this is inherently positive and straightforward, and on the other, the practical impossibility of sharing and translating ‘everything’, thus rendering decision-making about what to filter in or out inevitable. As discussed above, reluctant translators respond to this by aiming for accuracy in their translations. Dynamic translators, however, respond to this friction differently. According to our analysis, they translate not through the imperative of accuracy but by prioritising adequacy. In other words, here we see primarily senior social workers employing the epistemic value of adequacy over accuracy (Mitchell, 2016), a value, which has previously been, defended in relation to other knowledge communities (Demir, 2011; Demir & Murtagh, 2013).

Crucially, our analysis shows that the ability to prioritise adequacy over accuracy does not lie solely with the individual; it is dependent on the social, cultural and political system in which knowledge sharing takes place. Reluctant translators’ training, the requirements of evidence-based practice and the wider expectations and pressures on social work do not encourage practitioners to develop professional skills to capture and translate ‘grey evidence’. Neither were dynamic translators always confident that the system they operated in enabled them to use the knowledge sharing strategies they had developed. They translated ‘grey evidence’ but reported that at times this was difficult and against the grain. The demand for accuracy, in other words, creates a paradox for professionals who translate risk: whilst the demand leads social workers to desperately seek and long for accuracy, the demand, in fact, ends up frustrating the sharing and translation of risk and increases the chances of misunderstanding and miscommunication within a professional community of experts (in this case social workers), never mind when between experts from different communities (in this case multi-agency work).

Conclusion

Despite much work on the translation of risk in multi-agency work relating to child health and wellbeing, there is a lacuna of research on how risk is translated within a professional community of experts. Our findings and analysis aim to fill this significant gap in the literature on risk and social work by focusing on child and family social workers – a group of professionals who are at the heart of drives to improve child health and wellbeing. Based on findings from a qualitative study with social workers in England, we examined how social workers translate risk within their community. We identified two main approaches that they employ, namely reluctant translating and dynamic translating. Social workers identified as reluctant translators aim to share ‘everything’ and also stick to ‘the facts’. This approach, we identify, is based on their privileging of facts over values as well as a strict distinction they hold between facts and values that brought about (understandable) personal frustrations and tensions with others in the workplace. Dynamic translators, however, had an explicit, ‘less is more’ approach to the translation of risk; they were of the view that this added weight to their assessments and translations. Dynamic translators also attempted to retain an element of uncertainty, subjectivity and what we term ‘grey evidence’ when translating risk.

Our analysis demonstrated that different epistemic values and priorities underpin how social workers translate to risk. We identified that translations of risk by social workers are shaped by how they view three central epistemological points, namely grey evidence, the relationship between objectivity and subjectivity and the fact and value distinction.

This identification of the epistemological basis of risk work in this article is important on many other levels and invites a rethinking of risk research. Firstly, it significantly extends and enriches the literature on risk by making epistemology central to risk assessments and translations. Risk researchers have identified hope, morality, trust, emotion and intuition (Brown & Calnan, 2013; Douglas, 1990; Lupton, 2013; Zinn, 2008; 2016) as aspects of client-facing risk work but have not unpacked how epistemic values underpin risk assessments and translations. We give this previously missed yet central factor its due place. Secondly, there is an expectation, driven by evidence-based practice, that sharing with colleagues is inherently positive. Our findings showed that more senior social workers in fact thought that when it comes to the translation of risk ‘less can be more’. By showing that sharing ‘too much’ can be as problematic as ‘not sharing enough’, we raise food for thought for evidence-based practice advocates. We thus raise a new central concern for the field on risk and evidence. Thirdly, we identify a major paradox at the centre of evidence-based practice. We argue that through an emphasis on accuracy and boundaries, evidence-based practice might end up driving out ‘grey evidence’ which, as data show, only some social workers try to capture and translate through their prioritisation of adequacy over accuracy. The driving out of adequacy and the loss of grey evidence has the potential to hamper professional decision-making and judgement further. By pushing for the minimisation of grey evidence, and demanding accuracy, evidence-based practice might ironically end up losing valuable (grey) evidence – evidence which senior social workers report as important. More crucially, it might prevent social workers from exercising effective professional judgement, which is needed in order to improve the lives of children and their families.

Last but not least, our research has consequences for translation in multi-agency settings or for moving ‘bench to bedside’. Whilst our research focused on the translation of risk within a community, we argue that our findings and analysis could also be helpful in terms of rethinking the sharing of knowledge between different knowledge communities and professional groups in multi-agency work. If different epistemological values and priorities lead to variations on how experts translate risk within their own community, they also need to be uncovered in multi-agency settings. An examination of epistemic values are also needed in order to understand the complexities involved in translating health and social care research into practice and moving from bench to bedside. Thus, we argue that which epistemic values different expert communities prioritise need to be uncovered if we want to enhance our understanding of the failures and difficulties of translation across the borders of different professional groups, never mind within a group which our research identified.

Disclosure statement

No potential conflict of interest was reported by the authors.