Abstract In their article ‘A marriage made in hell: Child protection meets early intervention’, Featherstone et al. (2014) question the value of early intervention in preventing or addressing early signs of child maltreatment. In this article, we summarise and critique their main contentions. Among the issues we cover are the difference between intervention and support, the tension between fidelity and flexibility, the relative value of randomised controlled trials, the evidence of ‘what works’, the use of neuroscience, the place of innovation and the role of wider socio-economic factors. We are sympathetic to many of the points raised by Featherstone et al. but argue that they misrepresent early intervention, provide insufficient empirical support for their case and ignore evidence that runs counter to their views. We outline an alternative vision for child protection that addresses many of the concerns expressed while incorporating high-quality evidence on early intervention. Child, early intervention, evidence-based, family, protection, support Introduction Early intervention and child protection is a ‘marriage made in hell’, ‘an unholy alliance’ (p. 1735), according to Featherstone et al. (2014). This trenchant, award-winning (the 2015 British Association of Social Workers (BASW) Kay McDougall British Journal of Social Work Prize) and, for some we suspect, compelling critique calls into question the value of early intervention in preventing or addressing early signs of child maltreatment. We want to offer a different perspective and thereby promote further reflection. We start with some definitions. By early intervention in the context of child protection, we are referring to activity to prevent and/or reduce child maltreatment (physical, emotional and sexual abuse or neglect) and/or factors associated with it. The intervention may: (i) operate at a universal level for all children and/or their parents, such as to enhance parents’ knowledge of child development or children’s self-protection skills (promotion or universal prevention); or (ii) target children and families on the basis of elevated risk, such as economic deprivation, parental depression, substance misuse or family conflict (selective prevention); or (iii) focus on early signs of abuse, such as to reduce harsh discipline and the use of coercive parenting strategies (indicated prevention). It is not about addressing the physical or psychological consequences of maltreatment (what we might call treatment or therapy). The activity may entail practices or programmes, and it can be delivered in social work but also in other areas of children’s services, including health, education and youth justice. By defining early intervention in this way, we explicitly want to distance ourselves from some of what Featherstone et al. (2014) call ‘early intervention’, notably ‘early removal’ (p. 1736)—where children whose parents are deemed not to be providing adequate care are removed from their families or adopted. Further, in their article, ‘early help’ is contrasted with and deemed preferable to ‘early intervention’ (pp. 1742–3), but we think early intervention often is early help. We should be upfront about our stance. While we share many of the concerns raised by Featherstone et al., we think they paint an inaccurate and overly negative picture of evidence-based early intervention. If accepted uncritically, this will thwart efforts to promote interventions that have been tested using rigorous scientific methods and found to be effective in preventing and intervening early to address child maltreatment. We believe that evidence-based early intervention is essential to effective child protection and should be used more widely, and hope to demonstrate why in this article. We do not think the approach is a panacea, and nor do we make a case for any particular evidence-based programme (EBP). Our analysis of Featherstone et al.’s thesis is that it amounts to twelve main contentions. In what follows, we take each one in turn, summarising them and offering some critical but, we hope, constructive reflections. Intervention or support? Their first contention is that early intervention involves telling parents what to do, not helping them: it is about ‘intervention’, not ‘support’ (p. 1737). Thus, families are not supported or listened to (p. 1745) but rather have things delivered or done to them in a muscular fashion and as part of a government programme of ‘behaviour change’ (p. 1740). This sits in contrast with the concept of ‘support’, which is hands-on and practical; for example, it includes providing meals and childcare, and helping with washing (pp. 1736–7). In our experience, many evidence-based early intervention programmes, including those designed to prevent or reduce maltreatment, are supportive and involve working alongside families—indeed, failure to do this would render them ineffective. Building a strong ‘therapeutic alliance’ between the practitioner and family is integral to the more effective home-visiting programmes (Olds and Kitzman, 1990; Barnes et al., 2008). Far from being didactic or prescriptive, they often involve helping families to discover and practise their own solutions to problems. Moreover, while practical support is undoubtedly welcome for some families, it is unlikely to be sufficient for addressing many of the risk factors known to affect outcomes. Research strongly points to four family-related risk factors for child maltreatment, namely substance misuse, maternal depression, poor parenting skills and intimate partner violence (MacMillan et al., 2009; Mustillo et al., 2011). Interventions that address one or more of these are, we argue, more likely to help prevent or reduce maltreatment than those that do not. Risk and need In light of such comments, it is perhaps easy to see why Featherstone et al. maintain that early intervention is concerned with lessening risk—or ‘risk aversion’ (p. 1744)—rather than with meeting needs (p. 1738). This reflects a decline of earlier moral imperatives of mutual obligation (p. 1738). Support, by contrast, focuses on children and families’ strengths as well as their vulnerabilities (pp. 1735, 1737, 1747). While we recognise the current preoccupation in children’s services with risk aversion, we see this less as a defining feature of early intervention and more as a reflection of the prevailing policy climate and organisational culture, fuelled by media coverage of child protection ‘scandals’ (de Haan and Connolly, 2014). From a conceptual perspective, we also understand need to be a way of describing or summarising a cluster of risk and protective factors (Little et al., 2004). Thus, to contrast them in the way that is implied is unhelpful: the process of addressing specified risk factors helps to meet need. Moreover, early intervention is often additionally about boosting protective factors: it is not purely about reducing risk. Again, this helps to meet need. Evidence-based early intervention parenting programmes see as their core objective the need to build on families’ and communities’ strengths and the protective factors that are already part of children’s home environment. Of course, there are different ways of conceptualising need (Bradshaw, 1972), and we acknowledge that here we are referring essentially to ‘normative’ (or expert-defined) need. However, even if ‘felt’ (or subjective) need is the focus, there is ample evidence from satisfaction surveys, interviews and focus groups with participants that evidence-based early intervention programmes often scratch where people itch, even if there is more to do in this respect. This may reflect, in part, the tendency to use ‘goal-setting’ as a core modality—in other words, asking users what they want to get out of the intervention. Relationships A related criticism made by Featherstone et al. is that early intervention is delivered by ‘disembodied experts’ (p. 1745) who seek to intervene and solve problems rather than, as would be the case in a support-orientated model, to ‘listen, challenge and support a process of discovery and transformation’ (p. 1746). Early intervention is characterised as involving experts who assess and treat families in a disconnected and robotic fashion, spending more time recording data than properly engaging with people (pp. 1742, 1745). A more relational approach, the authors suggest, entails workers—described as ‘agents of hope’ (p. 1737)—building trust, applying professional judgement and delivering support with families, captured in the concept of ‘co-production’ (pp. 1743–5). In response, we would note that evidence-based early intervention programmes invariably require practitioners to operate in a very supportive and relational way. They emphasise building a strong therapeutic alliance, with home-visiting programmes and parent-training groups requiring that practitioners spend considerable time with families and build trust through activities that involve sharing, active listening, acknowledgement and validation, and praise and encouragement (Olds and Kitzman, 1990; Schmidt et al., 2014). This can be achieved even when time with the family is short and focused (Dishion et al., 2014). Many EBPs also require practitioners to have a core profession, expecting that the values and ethics they adhere to and the training they have received will benefit the delivery of the intervention. Further, little of the increased data-recording burden on child-protection practitioners in recent years has obviously been tied directly to evidence-based early intervention; rather, it has been a government requirement. Thus, the two should not be conflated. That said, we think that recording relevant and high-quality ‘real-time’ data on implementation and outcomes, and acting on what these show, can benefit practitioners, managers and, crucially, users. Specifically, as well as helping to plan interventions with families, and ensuring that staff are suitably trained and supported, it can help to improve child outcomes (Durlak and DuPre, 2008). We would also argue that relationships, while necessary, are unlikely to be sufficient. It is what happens in those relationships that matters (Rubin, 2011). For example, a strong therapeutic alliance may help a practitioner to win the trust of the family they are working with but, if they are to change harmful behaviours, they need to deploy techniques to achieve this behaviour change, such as motivational interviewing and cognitive behavioural therapy. Time Building relationships takes time, but Featherstone et al. maintain that early intervention implies a sense of urgency: it is ‘now or never’ (p. 1736). It requires rapid improvement within set time limits (p. 1739), so interventions are short-term, time-limited and discrete (p. 1745). This contrasts with the long-term and responsive support that is needed for ‘human flourishing’ (pp. 1744–5). We regret that early intervention has been hijacked in some quarters to support an agenda whereby marginalised families must quickly prove themselves adequate carers or lose their children to forced adoptions, and reject an individualised, moral discourse of child protection and failed parenting. We also do not doubt that intensive work over a sustained period is sometimes necessary, particularly for the most disadvantaged families. This does, however, point to the need for even earlier help. Numerous commentators point out that intervention is often ‘too little, too late’, as, by the time children reach the notice of social services, problems are complex and entrenched (Allen, 2011). Early intervention does not apologise for its ‘sense of urgency’; providing support to families before difficulties become embedded and cyclical is at the heart of the approach. Indeed, families may welcome early support that is construed not as seeking to ‘fix’ a failing or fault but rather as promoting their well-being (De Haan and Connolly, 2014). More fundamentally, it is simply not true that all early interventions are short or that there is no flexibility over duration. Some, notably home visiting, often last a significant length of time (several months, even a year or more), and many could be described as providing ‘responsive support’. With many programmes, practitioners can use their discretion to tailor delivery to participants’ needs—for example, meeting the pace of the group and not being prescriptive about covering all content in certain weeks. Moreover, there is nothing intrinsically wrong with short; indeed, there are good reasons not to intervene with families more than is necessary. From a rights perspective, families are entitled to have no more ‘intrusion’ into their lives than is necessary and, from an economic perspective, it makes no sense to spend more than is necessary to achieve the desired outcomes. While it is often hard at the outset to know how much support families will need, and over what period, a time boundary does let families know how much help they will be given and how long they need to commit to a programme, enabling them to make informed judgements about whether the service is right for them. Fidelity and flexibility The perceived rigidity of programme duration is part of a wider argument made by Featherstone et al., namely that early intervention demands fidelity to the model, meaning that delivery must be the same everywhere, regardless of context (p. 1740). It emphasises ‘standardising’ or ‘manualising’ the intervention (p. 1740), such as regarding features such as timing and format, resulting in rule-books and a reliance on jargon (p. 1746). By contrast, responsive support places greater emphasis on practitioners’ reflective learning (pp. 1744–5), allows users a choice about what support they receive and from whom, uses plain English and generally exhibits more human (presumably less robotic) qualities (p. 1746). In response, we have no hesitation in claiming that fidelity is positively associated with outcomes: there is substantial empirical evidence that, when interventions are delivered as intended, they generally produce better results (Axford et al., forthcoming a). It is also known that practitioners naturally ‘fiddle’ with interventions—for example, changing the sequence, amending content or shortening the length. EBPs are essentially vehicles for delivering a logic model or theory of change, and manuals are a method for packaging that logic model to help practitioners to deliver it consistently. At the same time, fidelity does not mean ‘no adaptation’. Indeed, there is an appreciation in the early intervention field that change is needed to ensure fit to the context, to secure buy-in from practitioners and, ultimately, to facilitate intervention scale-up. In short, there is flexibility. But adaptation should arguably be planned, undertaken in consultation with the developer and focus on peripheral rather than core components: unless there is good reason to believe the underlying mechanisms do not apply in the new context, the theory of change needs to be preserved (Bumbarger and Perkins, 2008). Further, implementing with fidelity does not mean practice without reflection or professional judgement. Many manualised interventions require practitioners to reflect formally (both individually and with peers) and make judgements about the pace of delivery, considering the needs of their client families. The notion that robotic adherence to the model is the core requirement flies in the face of evidence from many early intervention evaluations that skilled professionals are more likely than para-professional and non-professional colleagues to replicate positive outcomes in real-world delivery (e.g. Layzer et al., 2001). Poverty and inequality Featherstone et al. also take a broader, sociological, perspective, suggesting that early intervention overlooks the pernicious effects of poverty and inequality, which are an inevitable by-product of neo-liberalism (p. 1737). This is reflected in a tendency to design discrete and targeted interventions that focus on proximal risk (individual behaviour, family environment) (p. 1740) rather than recognising and addressing the context of adversity and associated distal risk (poor neighbourhoods, insecure jobs) (p. 1737). We agree that most evidence-based early interventions focus on factors at the individual child and family level, and to some extent the school level, although it is worth noting that not all are targeted: many are universal. While there is arguably scope for innovation to address factors at the neighbourhood and economic levels, these are more likely to require policy intervention. We also acknowledge the danger of focusing on discrete interventions only and thereby overlooking the need for social and economic policies to address poverty and inequality, both of which are associated with poor outcomes for children and families. But it need not be either/or: we advocate creating a fairer society as much as we support evidence-based early intervention. Early intervention programmes, including family support approaches, will always be swimming upstream in the absence of attention to structural issues (Rowlands, 2010). Moreover, if it is possible to address risk and protective factors now that will result in better outcomes, we should do so. There is some evidence, for example, that parenting mediates the effect of poverty on early childhood outcomes such as behaviour and cognitive and emotional well-being (Hölscher, 2008; Kiernan and Huerta, 2008). It is feasible with the right intervention to improve parenting skills relatively quickly, and for this effect to last, while lifting families out of poverty often takes much longer (not least for political reasons). If children with higher levels of difficulty benefit the most, which is the case with some interventions, then evidence-based early intervention can also promote greater equality. Ecology A connected claim made by Featherstone et al. is that early intervention focuses on the child and is individualised (p. 1744), implying that it does not attend sufficiently to the environmental systems with which children interact—family, services, even prevailing ideologies. They contrast this with a family support approach, which, they suggest, focuses on the family and is more collective, holistic, family-minded and community-based (pp. 1742, 1747). While we agree that relatively few evidence-based early intervention programmes target community and economic factors, most of them do target factors besides the individual child—in particular, issues in the family and the child’s school environment. Nevertheless, early intervention programmes are clearly not the solution; they can, and in our view must, co-exist with policies and practices to reorganise society, the economy and physical space in such a way as to better support child well-being. Complex and large-scale social problems, such as child abuse and neglect, require a range of stakeholders pursuing a common agenda through a series mutually reinforcing activities in order to achieve ‘collective impact’ (Hanleybrown et al., 2012). Methods such as Communities that Care, which has been found to enhance community-level protective factors (Kim et al., 2015), are promising in this respect, although they tend to focus predominantly on programmes. We like the more ambitious vision of Shonkoff and Fisher (2013), who call for ‘a fully integrated, inter-generational strategy that is grounded in developmental science, aligned at the program, community, and policy levels, and committed to the pursuit of breakthrough outcomes in lifelong learning, behavior, and health’ (p. 1646). Randomised controlled trials Of course, views about the best way to prevent child maltreatment might be regarded as conjecture in the absence of evidence of impact. On this issue, Featherstone et al. claim that early intervention rests on a particular methodological view of ‘what works’ (p. 1740). They do not spell out what this means but the implication is that it refers to the use of comparison group studies (especially randomised controlled trials (RCTs)) to determine programme effectiveness and, by extension, systematic reviews or meta-analyses of such studies. The advantages of RCTs are well rehearsed: in particular, they control for extraneous factors, such that differences in outcome between intervention and control groups at follow-up can reasonably be attributed to the intervention (Torgerson and Torgerson, 2008). Without them, there is a danger of over- or under-estimating impact, even suggesting a benefit when there is actually harm. Indeed, uncontrolled studies often yield exaggerated effect sizes due to ‘regression to the mean’ or temporal effects–things just ‘getting better’ (Oliver et al., 2010). In advocating the use of RCTs, we do not deny that there are other important types of evaluation, or valid questions about services besides whether they affect outcomes (Little et al., 2005). Moreover, RCTs are not all of equal quality, hence the need for standards of evidence and rigorous review processes, and randomisation cannot neutralise all bias, since—as in any evaluation—what is measured is inevitably contested, both politically and morally. We are also not naïve enough to think that effects in one setting will necessarily be replicated elsewhere: context clearly matters (Mayo-Wilson et al., 2013). This is why we and others trial established EBPs locally (e.g. Little et al., 2012) and warn against using ‘what works’ lists in a deterministic way. Nevertheless, where ethical and practical, RCTs arguably remain the best means of measuring whether a given intervention improves target outcomes. Further, narrative or statistical syntheses of such studies invariably provide a helpful overview of what works in a specified subject area, often accounting for variability in context and identifying moderators of impact and factors that predict greater effectiveness. Impact These debates about method are pertinent to the inference of Featherstone et al. that early intervention over-claims for its impact. They cite the Munro (2011) report’s caution about the ‘grander claims’ (p. 1743) of parenting programme franchises, contrasting this with the growing evidence internationally for a family support approach (p. 1746). The authors do not provide any empirical evidence to support the latter claim. Indeed, we would argue that the lack of specificity regarding such activity and inconsistency in how it is measured mean that it is impossible to appraise such evidence under one banner. Besides, we suspect that some of it is what we would regard as early intervention. Even so, we agree that proponents and developers (and, we would add, evaluators) of individual interventions sometimes exaggerate impact (Axford and Morpeth, 2013). This is an artefact of how and by whom evaluations are conducted, and how the results get reported. Studies may devise measures that favour the intervention group, for example, or omit from analyses participants who would skew the results negatively—perhaps those who did not receive the entire programme. There is also evidence that, when the programme developer is involved in an evaluation, the results tend to be more positive (known as ‘developer bias’) (Eisner, 2009). Further, papers often ignore or bury equivocal or negative results while cherry-picking positive findings—a particular problem given that evaluators do not always disclose a potential financial conflict of interest (Eisner et al., 2015). Prevention science takes such problems seriously, hence the growing use of standards of evidence against which to appraise programmes (e.g. Gottfredson et al., 2015) and guidelines to encourage the transparent and accurate reporting of trials (Schulz et al., 2010). Such approaches are designed to identify and reduce bias. The need for more evaluations that are independent of the developer is also widely acknowledged. All this said, we believe that the tendency to exaggerate impact is as great, if not greater, lower down the ‘hierarchy of evidence’ than is typically inhabited by evidence-based early intervention. As we note below, meta-analyses of RCTs in the area of child-maltreatment prevention often conclude that effects are small to modest. The increasing application of standards of evidence by online clearinghouses of EBPs further tempers the claims made for their success. For example, the Blueprints for Healthy Youth Development project (www.blueprintsprograms.com) has only approved approximately 5 per cent of the 1,300 or so programmes it has assessed (all of which, by definition, had at least one experimental or quasi-experimental study). Meanwhile, it is often difficult to find negative or null results in non-experimental evaluations of children’s services interventions, implying that ‘everything works’. In short, the science used to determine the impact of early intervention is inherently conservative. Further, while no one could claim that evidence-based early interventions offer a panacea for preventing and reducing child maltreatment, systematic reviews of experimental and quasi-experimental studies provide evidence of their promise. A range of different types of programmes—including home visiting, parent skills training, school-based curricula and family-focused provision—have been found to have a positive impact either on maltreatment itself or, more commonly, on associated risk and protective factors (for an overview, see Berry et al., 2013; Axford et al., forthcoming b; see also Euser et al., 2015; Chen and Chan, 2016). The effects are admittedly often small or modest, with variations by intervention type and level of intervention (i.e. universal or targeted), indicating the need for further innovation. Review authors also consistently call for more rigorous impact evaluations, including the use of more direct measures of abuse and neglect. It seems ironic to us that, at a time when many scientists are saying it is often hard to know ‘what works’ in child protection owing to the methodological weaknesses of many studies, others should appear to criticise more exacting tests of impact. Neuroscience One of the most vigorous contentions of Featherstone et al. is that early intervention abuses neuroscience (pp. 1736, 1739). Specifically, it rests on fallacious assumptions, implying that early adverse experiences cause irreversible damage, whereas the infant brain is actually remarkably resilient (p. 1739). In our experience, this claim is vastly exaggerated. Interventions that educate parents about the neurobiology underlying their child’s development tend to be premised not on avoiding or reducing damage, but rather on promoting and reinforcing parenting behaviour known to be positive for babies’ developing brains, such as attunement to baby’s cues, responsive communication and sensitive interaction (Glaser, 2014). Moreover, while we agree that that neuroscience has sometimes been abused in our field, by no means all early intervention is informed by neuroscience (at least explicitly): hitherto, it has mostly been based on an analyses of risk and protective factors derived from epidemiological and longitudinal research. To suggest otherwise is to overstate the importance of neuroscience in early intervention. That said, as findings emerge from neuroscience, its relevance to our field is growing rapidly. It helps to explain how risk gets into the body, and why some children are resilient. For example, Shonkoff and Fisher (2013) argue that providing child-focused enrichment and parenting education can be helpful but that they might have a ‘relatively limited impact on outcomes for children whose exposure to significant adversity might be producing neurobiological disruptions that make it more difficult to benefit fully from enhanced learning opportunities’ (p. 1641). Such insights can usefully inform innovation to reinforce care-giver and community capacity and resources to buffer children against exposure to toxic stress. This might include improving parents’ care for themselves and their children by boosting their cognitive flexibility (e.g. ability to set goals and control impulses), strengthening family economic stability, and reducing domestic and neighbourhood violence. Innovation It is surprising, then, to find Featherstone et al. also contending that evidence-based early intervention programmes stifle diversity and innovation, which are ‘frowned upon’ (p. 1745). This resonates with their depiction of early intervention as being preoccupied with standardisation as opposed to flexibility. Our first comment in response is that EBPs are themselves the product of innovation. They are tried-and-tested innovations, often started by practitioners and clinicians. Moreover, not all innovations are good, and some simply do badly what existing EBPs already do well. A rigorous process of designing, testing and refining innovations is needed to identify the best (Craig et al., 2008), accepting that some approaches or iterations deservedly fall by the wayside. The difference between new innovations and established programmes is that the latter have progressed further in that process. We would also add that there is much scope for innovation within and beyond existing early intervention programmes, not least in relation to neuroscience (see above) and how best to serve so-called ‘hard-to-reach’ families (see below). Indeed, innovation is arguably the norm with EBPs. Developers continue to innovate, making interventions longer or shorter, adapting content for different socio-demographic or need groups, and changing the form or nature of training. Variants of the original catch-all model are also increasingly common (e.g. Webster-Stratton and Reid, 2010). In short, to suggest that evidence-based early intervention is somehow at loggerheads with innovation is a false distinction: it is both-and, not either-or. That said, more could be done to use existing interventions as a platform for innovation—for instance, through the distillation and sharing of common theories of change, or identifying and disseminating common elements or ‘kernels’ of effective programmes (Embry and Biglan, 2008). The underclass The final assertion by Featherstone et al. is that early intervention perpetuates the idea of a feckless underclass (p. 1739). The approach involves targeting families considered to be ‘risky’, ‘hard-to-reach’, ‘undesirable’, ‘deviant’ or ‘very troubled’, especially in the early years (pp. 1739–40, 1744), and treating them as self-excluding and exhibiting learned helplessness; rather than being seen as needing support, they are deemed to require ‘reconstruction’ as active citizens (p. 1738). Interventions in the UK deemed to exemplify this ‘moralising’ and ‘authoritarian’ approach to families (p. 1744) include Family Nurse Partnership (FNP) and Sure Start (p. 1741). We recognise that recent UK governments have promulgated a moral underclass discourse (e.g. Levitas, 1998) but it is not a perspective we subscribe to and nor do we consider it integral to the concept of early intervention. As already indicated, by no means all early intervention is targeted: much is universal. It seeks to promote good things for everyone (‘promotion’), or work with everyone to prevent bad things (‘universal prevention’) (O’Connell et al., 2009). In child protection, there are numerous examples of interventions involving mass media components or school-based curricula. These typically focus on the knowledge, attitudes, skills and intentions of parents, children and bystanders. Other interventions are targeted according to elevated risk (‘selective prevention’) or early signs of problems (‘indicated prevention’) but this is based on evidence that, without intervention, such children are disproportionately likely to have poor outcomes in the future. In child protection, parents might be targeted on the basis of elevated risk of abusing or neglecting children (e.g. previous abuse, substance misuse, psychiatric disorder, violent offending). Notwithstanding the potential stigma of targeted services, it seems sensible to us to target interventions at those who need them and stand to gain the most. Moreover, as De Haan and Connolly (2014) suggest, ‘stigma could be defused through professionals embodying and promoting the notion that attending to needs shows strength’ (p. 89). Further, an important feature of early intervention programmes is often that the people who deliver them need to treat participants with great respect, acknowledging their own knowledge and expertise. For instance, a recent study found that FNP was positively received by fathers because of the quality time that nurses invested in developing relationships with them and their ‘skilled, therapeutically orientated, holistic approach’ (Ferguson and Gates, 2015, p. 104). Such practice could hardly be described as authoritarian or moralising. Indeed, Featherstone et al. maintain elsewhere that Sure Start ‘was in a tradition of helpful, negotiated support’ (p. 1740). They also, ironically, give more credit than is deserved to the efforts of early intervention programmes to target and serve so-called ‘hard-to-reach’ families. If only it were true. There is good empirical evidence that some groups are disproportionately likely not to use services, but the problem is arguably as much, if not more, of a problem of early intervention often proving ‘hard to access’ (Axford et al., 2012). Discussion Featherstone et al.’s article is part of a long-standing radical tradition in social work and child welfare that commends practical action and attacks neo-liberalism (e.g. Holman, 1983; Garrett, 2009), and this debate has echoes of earlier altercations about the nature and role of evidence and the art or science of practice—including in the pages of this Journal (e.g. Sheldon, 2001; Webb, 2001). We are sympathetic to many of the concerns expressed by Featherstone et al. but dispute much of their reasoning and draw different conclusions. We argue that they set up straw men, make false distinctions and provide insufficient evidence to support their case while ignoring evidence that runs counter to their views. To start with, they imply that early intervention fails to do things that in our experience are actually quite common: building supportive relationships with families to protect children, often over a long period; adapting services so that they are responsive to families’ needs and contexts; and promoting strengths as well as minimising risks. Next, they unfairly attribute to early intervention features that few of its advocates would recognise or defend: a requirement to collect excessive data; the lazy use of or excessive reliance on neuroscience; a stifling of innovation; the perpetuation of a pernicious moral underclass discourse; and a denial of the need for structural reform. Then they fail to acknowledge the significant strengths of early intervention or explain how their preferred approach matches these or provides credible alternatives: robust evidence of impact, including for the worst-off; tackling known contributors to abuse and neglect; and doing so in an efficient manner. It is clear that all is not well with child protection. A review of trends in child maltreatment in developed countries since the inception of modern child protection services found no consistent evidence of a decrease in rates of abuse and neglect (Gilbert et al., 2012). Traditional service responses are typically embroiled in ‘fighting fires’—working with families already caught up in a pattern of abuse and neglect, and trying to prevent its reoccurrence or long-term impairment. Something needs to change. Evidence-based early intervention programmes are no silver bullet. Their effects are often small, fidelity in the real world is elusive and few if any have been scaled. Yet we contend that they are part of the silver ‘buckshot’ needed to address the problem. A public health approach affords one possibility for incorporating high-quality evidence on early intervention while also addressing many of the concerns expressed by Featherstone et al. Public health models involve protecting all children in the population from risks to their health and development, including those at high risk (Berry, 2008; Barlow and Calam, 2011; Berry et al., 2013; Woodman and Gilbert, 2013). The hypothesis for its application to child protection is that encouraging more parents in the population to use warm and responsive strategies in caring for their children, and limiting their use of overtly hostile or aggressive tactics, creates a social contagion effect, setting the boundaries for what is considered ‘normative’ practice within a community and reducing abuse at the high end of the distribution (cf. Rose, 2008). What would this alternative vision for child protection entail? First, it would mean gaining a more robust estimate of the extent of the problem in the population and the factors (at all levels or in all contexts) that drive the problem. Large-scale epidemiological and aetiological data as well as local routine health and social care data are needed to do this. The data, not ideology, drive policy and programme decisions. Second, public systems would employ more early intervention activities that promote secure and safe relationships, attachment and positive parenting (Axford et al., 2015; Barlow et al., 2016). The use of universal proportionate (UP) intervention models is advocated, where input is both preventative and selective (Marmot, 2010). Families are not required to demonstrate difficulty in order to receive assistance in the first place; instead, a universal layer of intervention works in a preventative fashion to reduce the need for later more intensive treatment. The scale and intensity/dose of the intervention are determined on the basis of need, and families are offered more than one opportunity to receive assistance. Interventions can be tailored according to both neighbourhood and individual need, based on appropriate data (Lowe, 2007). Few UP models have been tested with child protection outcomes in mind, but one trial is examining the effectiveness of a UP parent support model for reducing maternal depression and child social-emotional developmental difficulty (see ISRCTN11079129). Third, adopting a public health approach would mean greater collaboration between—and integration of—existing services and professionals, working together to promote child welfare. This approach locates the responsibility and activity for safeguarding children with all early years, education and health professionals; it is not solely the business of children’s social care. Tiered early intervention programmes are best delivered by multi-agency teams. For example, health workers, working with young mothers of newborn babies, would develop relationships with families early on to educate them about child development before difficulties arise, with social care and mental health professionals also providing input into these services to ensure that risks are appropriately assessed and crises are avoided. How does this translate into day-to-day practice for frontline staff? It means family support workers, health visitors and other children’s services professionals co-delivering evidence-based parent support and family resilience programmes as part of their role or remit. The people delivering these interventions need appropriate expertise and training, as well as supervision in relation to managing cases and assessing risk. Attention must be paid to implementation fidelity and the quality of intervention delivery, and agreement on a common set of outcome monitoring instruments would promote inter-agency collaboration, data sharing and evaluation. A public health approach advocates testing and monitoring the effectiveness of any intervention introduced, including policy change. Woodman and Gilbert (2013) point to the distinct lack of robust studies of services provided in the child protection field, despite the consequences and costs of abuse and neglect. The public health approach claim is that improving—in small ways—the interaction between parents and children in the general population has the potential to impact those at the extreme end of the continuum, where poor parenting and interaction have become abusive/harmful (Berry, 2008; Woodman and Gilbert, 2013). Indeed, there is evidence to suggest that parenting programmes that adopt a strengths-based approach—also advocated by Featherstone et al.—often demonstrate the greatest impact for those with higher starting levels of difficulty (Holzer et al., 2006). While the emphasis in early intervention is often on work with individual children and families, macro-level policies can and would also set a tone that promotes children’s welfare, such as banning the use of physical punishment or using mass media to communicate messages about child development or parenting issues. We offer these thoughts in the spirit of intellectual inquiry and a desire to help all children flourish, and hope that it stimulates further reflection and discussion. Acknowledgements Vashti Berry's time is supported by the National Institute for Health Research (NIHR) Collaboration for Leadership in Applied Health Research and Care South West Peninsula. The views expressed are those of the author(s) and not necessarily those of the NHS, the NIHR or the Department of Health. We are grateful to Louise Morpeth and Roger Bullock for comments on an earlier draft of this article, and to Bethany Cuffe-Fuller for help with references. References Allen G. ( 2011) Early Intervention: The Next Steps , London, Cabinet Office. Axford N., Morpeth L. ( 2013) ‘ Evidence-based programs in children’s services: A critical appraisal’, Children and Youth Services Review , 35( 1), pp. 268– 77. Google Scholar CrossRef Search ADS Axford N., Barlow J., Coad J., Schrader-McMillan A., Sonthalia S., Toft A., Wrigley Z., Goodwin A., Ohlson C., Bjornstad G. ( 2015) The Best Start at Home: What Works to Improve the Quality of Parent–Child Interactions from Conception to Age 5 Years? A Rapid Review of Interventions , London, Early Intervention Foundation. Axford N., Bywater T., Blower S., Berry V., Baker V., Morpeth L. ( forthcoming a) ‘Critical factors in the implementation of parenting programmes: Fidelity, adaptation and promoting quality’, in Dixon L., Perkins D., Craig L. A., Hamilton-Giachritsis C. (eds), What Works in Child Protection: An Evidence-Based Approach to Assessment and Intervention in Care Proceedings , Chichester, Wiley-Blackwell. Axford N., Lehtonen M., Tobin K., Kaoukji D., Berry V. ( 2012) ‘ Engaging parents in parenting programs: Lessons from research and practice’, Children and Youth Services Review , 34( 10), pp. 2061– 71. Google Scholar CrossRef Search ADS Axford N., Lowther K., Timmons L., Brook L., Webb L., Sonthalia S. ( forthcoming b) Rapid Review on Safeguarding to Inform the Healthy Child Programme 5–19 , London, Public Health England. Barlow J., Calam R. ( 2011) ‘ A public health approach to safeguarding in the 21st century’, Child Abuse Review , 20( 4), pp. 238– 55. Google Scholar CrossRef Search ADS Barlow J., Schrader-McMillan A., Axford N., Wrigley Z., Sonthalia S., Wilkinson T., Rawsthorn M., Toft A., Coad J. ( 2016) ‘ Attachment and attachment-related outcomes in pre-school children: A review of recent evidence’, Child and Adolescent Mental Health , 21( 1), pp. 11– 20. Google Scholar CrossRef Search ADS Barnes J., Ball M., Meadows P., McLeish J., Belsky J. and the FNP Implementation Research Team ( 2008) Nurse–Family Partnership programme: First Year Pilot Sites—Implementation in England Pregnancy and the Post-Partum Period , Research Report DCSF-RW051, London, DCSF and Department of Health. Berry V. ( 2008) ‘The relative contribution of family conflict to children’s health and development’, unpublished Ph.D. thesis, University of Bath. Berry V., Blower S., Axford N., Little M., Kaoukji D. ( 2013) ‘The prevention of child maltreatment and neglect’, in Conte J. (ed.), Child Abuse and Neglect Worldwide , Santa Barbara, CA, Praeger. Bradshaw J. R. ( 1972) ‘The taxonomy of social need’, in McLachlan G. (ed.), Problems and Progress in Medical Care , Oxford, Oxford University Press. Bumbarger B., Perkins D. ( 2008) ‘ After randomised trials: Issues related to dissemination of evidence‐based interventions’, Journal of Children’s Services , 3( 2), pp. 55– 64. Google Scholar CrossRef Search ADS Chen M., Chan K. L. ( 2016) ‘ Effects of parenting programs on child maltreatment prevention: A meta-analysis’, Trauma Violence Abuse , 17( 1), pp. 88– 104. Google Scholar CrossRef Search ADS PubMed Craig P., Dieppe P., Macintyre S., Michie S, Nazareth I., Petticrew M. ( 2008) ‘ Developing and evaluating complex interventions: The new Medical Research Council guidance’, BMJ , 337, p. a1655. Google Scholar CrossRef Search ADS PubMed De Haan I., Connolly M. ( 2014) ‘ Another Pandora’s box? Some pros and cons of predictive risk modelling’, Children and Youth Services Review , 47( 1), pp. 86– 91. Google Scholar CrossRef Search ADS Dishion T. J., Brennan L. M., Shaw D. S., McEachern A. D., Wilson M. N., Jo B. ( 2014) ‘ Prevention of problem behavior through annual check-ups in early childhood: Intervention effects from home to elementary school’, Journal of Abnormal Child Psychology , 42( 3), pp. 343– 54. Google Scholar CrossRef Search ADS PubMed Durlak J. A., DuPre E. P. ( 2008) ‘ Implementation matters: A review of research on the influence of implementation on program outcomes and the factors affecting implementation’, American Journal of Community Psychology , 41(3–4), pp. 327– 50. Google Scholar CrossRef Search ADS Eisner M. ( 2009) ‘ No effects in independent prevention trials: Can we reject the cynical view?’, Journal of Experimental Criminology , 5( 2), pp. 163– 83. Google Scholar CrossRef Search ADS Eisner M., Humphreys D. K., Wilson P., Gardner F. ( 2015) ‘ Disclosure of financial conflicts of interests in interventions to improve child psychosocial health: A cross-sectional study’, PLoS ONE , 10( 11), p. e0142803. Google Scholar CrossRef Search ADS PubMed Embry D. D., Biglan A. ( 2008) ‘ Evidence-based kernels: Fundamental units of behavioral influence’, Clinical Child and Family Psychology Review , 11( 3), pp. 75– 113. Google Scholar CrossRef Search ADS PubMed Euser S., Alink L. R. A., Stoltenborgh M., Bakermans-Kranenburg M. J., van IJzendoorn M. H. ( 2015) ‘ A gloomy picture: A meta-analysis of randomized controlled trials reveals disappointing effectiveness of programmes aiming at preventing child maltreatment’, BMC Public Health , 15, p. 1068. Google Scholar CrossRef Search ADS PubMed Featherstone B., Morris K., White S. ( 2014) ‘ A marriage made in hell: Early intervention meets child protection’, British Journal of Social Work , 44( 7), pp. 1735– 49. Google Scholar CrossRef Search ADS Ferguson H., Gates P. ( 2015) ‘ Early intervention and holistic, relationship-based practice with fathers: Evidence from the work of the Family Nurse Partnership’, Child & Family Social Work , 20( 1), pp. 96– 105. Google Scholar CrossRef Search ADS Garrett P. M. ( 2009). Transforming Children’s Services? Social Work, Neoliberalism and the ‘Modern’ World , Maidenhead, Open University Press/McGraw-Hill. Gilbert R., Fluke J., O’Donnell M., Gonzalez-Izquieredo A., Brownell M., Gulliver P., Janson S., Sidebotham P. ( 2012) ‘ Child maltreatment: Variation in trends and policies in six developed countries’, Lancet , 379, pp. 758– 72. Google Scholar CrossRef Search ADS PubMed Glaser D. ( 2014) ‘ Neurodevelopment in the first three years: Implications for child development, professional practice and policy’, Journal of Children’s Services , 9( 2), pp. 154– 64. Google Scholar CrossRef Search ADS Gottfredson D. C., Cook T. D., Gardner F. E. M., Gorman-Smith D., Howe G. W., Sandler I. W., Zafft K. M. ( 2015) ‘ Standards of evidence for efficacy, effectiveness, and scale-up research in prevention science: Next generation’, Prevention Science , 16( 7), pp. 893– 926. Google Scholar CrossRef Search ADS PubMed Hanleybrown F., Kania J., Kramer M. ( 2012) Channelling Change: Making Collective Impact Work, Stanford Social Innovation Review, available online at http://ssir.org/articles/entry/channeling_change_making_collective_impact_work (accessed 8 February 2016). Holman R. ( 1983) Resourceful Friends: Skills in Community Social Work , London, The Children’s Society. Hölscher P. ( 2008) ‘Children and young people’s experiences of poverty and social exclusion’, in Ridge T., Wright S. (eds), Understanding Inequality, Poverty and Wealth , Bristol, The Policy Press. Holzer P. J., Higgins J. R., Bromfield L. M., Higgins D. J. ( 2006) ‘ The effectiveness of parent education and home visiting child maltreatment prevention programs’, Child Abuse Prevention Issues , 24, pp. 1– 24. Kiernan K. E., Huerta M. C. ( 2008) ‘ Economic deprivation, maternal depression, parenting and children’s cognitive and emotional development in early childhood’, British Journal of Sociology , 59( 4), pp. 783– 806. Google Scholar CrossRef Search ADS PubMed Kim B. K. E., Gloppen K. M., Rhew I. C., Oesterle S. J., Hawkins J. D. ( 2015) ‘ Effects of the Communities that Care prevention system on youth reports of protective factors’, Prevention Science , 16( 5), pp. 652– 62. Google Scholar CrossRef Search ADS PubMed Layzer J. I., Goodson B. D., Bernstein L., Price C. ( 2001) National Evaluation of Family Support Programs. Final Report. Volume A: The Meta-analysis , Cambridge, MA, Abt Associates. Levitas R. ( 1998) The Inclusive Society? Social Exclusion and New Labour , Basingstoke, Macmillan. Little M., Axford N., Morpeth L. ( 2004) ‘ Research review: Risk and protection in the context of services for children in need’, Child and Family Social Work , 9( 1), pp. 105– 17. Google Scholar CrossRef Search ADS Little M., Axford N., Morpeth L., Weyts A. ( 2005) ‘ Evaluating children’s services: Recent conceptual and methodological developments’, British Journal of Social Work , 35( 1), pp. 73– 88. Google Scholar CrossRef Search ADS Little M., Berry V., Morpeth L., Blower S., Axford N., Taylor R., Bywater T., Lehtonen M., Tobin K. ( 2012) ‘ The impact of three evidence-based programmes delivered in public systems in Birmingham, UK’, International Journal of Conflict and Violence , 6( 2), pp. 260– 72. Lowe R. ( 2007) Facing the Future: A Review of the Role of Health Visitors , London, Queen’s Nursing Institute. MacMillan H. L., Wathen C. N., Barlow J., Fergusson D. M., Leventhal J. M., Taussig H. N. ( 2009) ‘ Interventions to prevent child maltreatment and associated impairment’, The Lancet , 373, pp. 250– 66. Google Scholar CrossRef Search ADS Marmot M. ( 2010) Fair Society, Healthy Lives: The Marmot Review, available online at www.marmotreview.org/ (accessed 17 January 2017). Mayo-Wilson E., Grant S., Hopewell S., Macdonald G., Moher D., Montgomery P. ( 2013) ‘ Developing a reporting guideline for social and psychological intervention trials’, Trials , 14, p. 242. Google Scholar CrossRef Search ADS PubMed Munro E. ( 2011) The Munro Review of Child Protection: Final Report: A Child-Centred System , London, TSO. Mustillo S. A., Dorsey S., Conover K., Burns B. J. ( 2011) ‘ Parental depression and child outcomes: The mediating effects of abuse and neglect’, Journal of Marriage and Family , 73( 1), pp. 164– 80. Google Scholar CrossRef Search ADS O’Connell M. E., Boat T., Warner K. E. (eds) ( 2009) Preventing Mental, Emotional, and Behavioural Disorders among Young People: Progress and Possibilities , Washington, DC, The National Academies Press. Olds D. L., Kitzman H. ( 1990) ‘ Can home visitation improve the health of women and children at environmental risk?’, Pediatrics , 86( 1), pp. 1108– 16. Oliver S., Bagnall A. M., Thomas J., Shepherd J., Sowden A., White I., Dinnes J., Rees R., Colquitt J., Oliver K., Garrett Z. ( 2010) ‘ Randomised controlled trials for policy interventions: A review of reviews and meta-regression’, Health Technology Assessment , 14( 16), pp. 1– 165. Google Scholar CrossRef Search ADS PubMed Rose G. ( 2008) Rose’s Strategy of Preventive Medicine , Oxford, Oxford University Press. Google Scholar CrossRef Search ADS Rowlands J. ( 2010) ‘ Services are not enough: Child well-being in a very unequal society’, Journal of Children’s Services , 5( 3), pp. 80– 8. Google Scholar CrossRef Search ADS Rubin A. ( 2011) ‘ Teaching EBP in social work: Retrospective and prospective’, Journal of Social Work , 11( 1), pp. 64– 79. Google Scholar CrossRef Search ADS Schmidt F., Chomycz S., Houlding C., Kruse A., Franks J. ( 2014) ‘ The association between therapeutic alliance and treatment outcomes in a group Triple P intervention’, Journal of Child and Family Studies , 23( 8), pp. 1337– 50. Google Scholar CrossRef Search ADS Schulz K. F., Altman D. G., Moher D. for the CONSORT Group ( 2010) ‘ CONSORT 2010 Statement: Updated guidelines for reporting parallel group randomised trials’, BMJ , 340, p. c332. Google Scholar CrossRef Search ADS PubMed Sheldon B. ( 2001) ‘ The validity of evidence‐based practice in social work: A reply to Stephen Webb’, British Journal of Social Work , 31( 5), pp. 801– 9. Google Scholar CrossRef Search ADS Shonkoff J. P., Fisher P. A. ( 2013) ‘ Rethinking evidence-based practice and two-generation programs to create the future of early childhood policy’, Development and Psychopathology , 25(4, Pt 2), pp. 1635– 53. Google Scholar CrossRef Search ADS Torgerson D., Torgerson C. ( 2008) ‘ Invited editorial: Randomised controlled trials in children’s services’, Journal of Children’s Services , 3( 1), pp. 2– 8. Google Scholar CrossRef Search ADS Webb S. A. ( 2001) ‘ Some considerations on the validity of evidence-based practice in social work’, British Journal of Social Work , 31( 1), pp. 57– 79. Google Scholar CrossRef Search ADS Webster-Stratton C., Reid M. J. ( 2010) ‘ Adapting The Incredible Years, an evidence-based parenting programme, for families involved in the child welfare system’, Journal of Children’s Services , 5( 1), pp. 25– 42. Google Scholar CrossRef Search ADS Woodman J., Gilbert R. ( 2013) ‘Child maltreatment: Moving towards a public health approach’, in BMA Board of Science (ed.), Growing Up in the UK: Ensuring a Healthy Future for Our Children , London, British Medical Association. © The Author 2017. Published by Oxford University Press on behalf of The British Association of Social Workers. All rights reserved.
The British Journal of Social Work – Oxford University Press
Published: Jan 1, 2018
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 12 million articles from more than
10,000 peer-reviewed journals.
All for just $49/month
Read as many articles as you need. Full articles with original layout, charts and figures. Read online, from anywhere.
Keep up with your field with Personalized Recommendations and Follow Journals to get automatic updates.
It’s easy to organize your research with our built-in tools.
Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.
All the latest content is available, no embargo periods.
“Hi guys, I cannot tell you how much I love this resource. Incredible. I really believe you've hit the nail on the head with this site in regards to solving the research-purchase issue.”Daniel C.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud
“I must say, @deepdyve is a fabulous solution to the independent researcher's problem of #access to #information.”@deepthiw
“My last article couldn't be possible without the platform @deepdyve that makes journal papers cheaper.”@JoseServera