Making connections: a multi-disciplinary analysis of domestic homicide, mental health homicide and adult practice reviews

Amanda Lea Robinson (School of Social Sciences, Cardiff University, Cardiff, UK)
Alyson Rees (School of Social Sciences, Cardiff University, Cardiff, UK)
Roxanna Dehaghani (School of Law and Politics, Cardiff University, Cardiff, UK)

The Journal of Adult Protection

ISSN: 1466-8203

Article publication date: 19 February 2019



The purpose of this paper is to “read across” a sample of domestic homicide reviews (DHRs), mental health homicide reviews (MHHRs) and adult practice reviews (APR) to identify the cross-cutting themes.


The study involved a qualitative comparative analysis of 20 Welsh reviews: 10 DHRs, 6 APRs and 4 MHHRs. Each review was triple coded by a multi-disciplinary team of researchers (representing criminology, social work and law).


Five overarching themes were identified from this diverse sample of cases: crossing boundaries, including transitions between services and geographical boundaries; hoodwinking, where there was manipulation of the presentation of self; faulty assessment, which was not always holistic and only based on certain aspects of behaviour; tunnel vision, resulting from the initial underpinning narrative rarely being challenged; and knowledge, with certain types being privileged over others, especially professional over that of families and para-professionals.

Research limitations/implications

Further research into death reviews should adopt a comparative, multi-disciplinary approach.

Practical implications

The research highlights the possibility for duplication across the different types of reviews. Further, it suggests that review processes could be streamlined.


Five cross-cutting themes have been developed through the very first study “reading across” three types of reviews (DHRs, APRs and MHHRs). Findings suggest the need for streamlining review processes and highlight the importance of adopting a multi-disciplinary perspective when researching death reviews.



Robinson, A.L., Rees, A. and Dehaghani, R. (2019), "Making connections: a multi-disciplinary analysis of domestic homicide, mental health homicide and adult practice reviews", The Journal of Adult Protection, Vol. 21 No. 1, pp. 16-26.



Emerald Publishing Limited

Copyright © 2019, Emerald Publishing Limited


Both the volume and type of death reviews taking place in the UK have grown significantly since the first child death scandal of Dennis O’Neill who was “boarded out”[1] in 1945 (Hopkins, 2007) and the subsequent Monckton four day inquiry and report (Home Office, 1945). The introduction of new statutory requirements in the 1990s and 2000s has greatly contributed to an “inquiry culture” that has prompted death reviews becoming “much more a feature of public life” (Nash and Williams, 2008, p. 134). Consequently, they routinely feature in the professions of many in both statutory and non-statutory agencies (e.g. via training which draws from reviews, or submitting evidence or otherwise participating in an official inquiry).

The aim of conducting a review is to “generate professional and organisational learning and promote improvement in future inter-agency adult protection practice” (Welsh Government, 2016). For example, domestic homicide reviews (DHRs)[2], mental health homicide reviews (MHHR)[3] and adult practice reviews (APR)[4] are all underpinned by a desire to “learn the lessons” from tragic and potentially avoidable deaths. Indeed, the expectation that learning must follow from these events means that some will trigger multiple reviews under current statutory obligations. Yet, despite the significant level of resources invested in producing these various types of reviews, presently it is unclear the extent to which their findings have added to the sum of professional knowledge (Salter, 2003) or if their recommendations are simply a placebo, giving the impression of change “on the ground” (Elliott and McGuiness, 2002).

Existing publications synthesising findings from a number of reviews tend to restrict their sample to a single type. For example, there are a few analyses of DHRs (Home Office, 2016), of APRs (Stevens et al., 2017) of MHHRs (Health Inspectorate Wales, 2016) and of serious case reviews (Sidebotham et al., 2016)[5]. These publications are produced with a focus on a particular type of review, and are thus derived from and speak to a particular disciplinary and practice-based audience (e.g. criminology and criminal justice for DHRs, social work for APRs and psychology and psychiatry for MHHRs). This reinforces the notion of different professional groups inhabiting different “planets”, with their own separate histories, culture, and laws (Hester, 2011), and reinforces divides rather than helping professionals make connections across teams, settings and disciplines. The current study aims to address this gap in knowledge, by “reading across” different types of reviews to uncover learning that can be considered fundamentally relevant to all professional practice (whether it is in the context of domestic abuse, vulnerable adults and/or mental health). As this is the first study to provide a multi-disciplinary thematic analysis across more than one type of review, the results provide a preliminary foundation to inform future research and practice in this area.


This research project was commissioned by the National Independent Safeguarding Board (NISB) (via Welsh Government) and took place during January–March 2018. The overall approach to this study is qualitative, involving thematic coding of reviews complemented by focus group discussions with practitioners working in different agencies and geographical locations within Wales.


The NISB provided a sample of 20 reviews to be triple coded by the research team: 10 DHRs, 6 APRs and 4 MHHRs (for an overview table of the sample, see appendix 1 in Robinson et al., 2018). This represents a proportion of the total number of such reviews carried out in Wales: approximately 20 DHRs, 15 APRs and 10 MHHRs over the period 2008–2018[6]. Each review in the sample was identified and retrieved by the NISB by using their professional networks, with the aim of facilitating the consideration of a broad array of cases and circumstances across Wales. Each of the 20 reviews dealt with a unique case and all but one involved the death of an individual (the other involved a serious sexual assault). Two of the reviews involved multiple deaths (these were domestic homicides of partners along with other family members). Of the ten DHRs, nine involved female intimate partners killed by males, and one involved a son killing his father. In two DHRs, the perpetrator attempted suicide (and in one was successful) following the homicide. Three of the four MHHRs involved males killing females (two were strangers and one was an acquaintance or possibly a new intimate partner); the fourth involved a male killing a male acquaintance. One of the MHHRs also involved the death of the perpetrator whilst in police custody. The APRs involved two elderly people dying in care, one middle-aged man dying in the community and two younger people dying (one ended his life by suicide in prison). The one non-fatal case in the sample involved an APR into a serious sexual assault.

Coding framework

The research team established a method and framework to identify key themes. This involved an initial reading and discussion of two reviews to develop of a coding framework. An Excel worksheet was created for each researcher to note findings relating to each of the following categories: characteristics of the abuse; agency performance (police, probation, health, mental health, adult safeguarding, children’s safeguarding and other); multi-agency partnership working; new learning/valuable insights; key recommendations made in the review; comments on its quality; and any other comments from a particular disciplinary perspective.

All reviews were triple coded by the multi-disciplinary research team. Weekly team meetings over a four-week period provided the space to discuss batches of reviews. Upon the completion of the coding, all coding worksheets were combined into a single Excel database, with one worksheet designated for each review. Each worksheet contained the codes from every team member, in order to evaluate their similarity and points of divergence.

The coding exercise generated a group of five cross-cutting themes and these provided the structure for the focus group discussion. These five themes were significant features in all three types of review; they were not specific per se to issues of domestic abuse, vulnerable adults, or mental health (for an overview table of the themes for each review, see appendix 2 in Robinson et al., 2018). Thus, these are high-level themes that go beyond particular operational boundaries or substantive issues. These five themes were subject to a validity check through discussion with practitioners in two focus groups.

Focus groups

The NISB identified suitable participants for the focus groups, with one focus group held in North Wales (Wrexham) and the other in South Wales (Cardiff). Each focus group included twelve participants and lasted two hours. The topics for discussion were guided by the cases: the researchers selected exemplar cases for each theme and then discussed, with the participants, whether the theme resonated with them.

Prior to the focus group, a brief survey was sent to the 30 individuals who were registered to attend; 22 responses were received. All participants had some level of knowledge and experience with either DRHs, APRs or MHHRs. The survey was compiled so as to gauge the respondents’ level of experience with the different types of review, whether they believed there to be commonalities between different types of review (and what these commonalities were), and how learning could be enhanced through the reviews or dissemination of these reviews. The survey was compiled through discussion amongst the three researchers, in line with the research questions. Respondents were unanimous in their belief that these types of reviews would, generally speaking, tend to identify similar failings and missed opportunities for intervention. For example:

From my experience there are often common themes across reviews e.g. working in silos, not sharing information, inaccurate risk assessments, full history of case not considered or used to inform risk assessment although known to some or all agencies.

(No. 11)

Yes, because the reasons for why things go wrong are generally similar but are very difficult to change.

(No. 13)

These perceptions, expressed so consistently and prior to the focus groups taking place, reinforce the results of the coding exercise, which found more similarity than difference across reviews. Whilst the focus groups were not recorded, notes on flip charts were taken at the time and then immediately consolidated afterwards into a written account of the key themes. This information was then supplemented by an opportunity for all participants to provide feedback to the research team via a short online survey that was disseminated afterwards.


The sample was a convenience sample provided by the NISB; it does not necessarily provide a representative sample of Welsh DHRs, APRs and/or MHHRs. However, reviews were chosen with a view to ensuring a wide geographic spread of cases within Wales, and to illustrate the diverse range of issues that tend to be found in such reviews. Further research is necessary to substantiate the findings presented here.


The sections below discuss the five cross-cutting themes identified from the coding exercise and confirmed by the focus group discussion.

Theme 1: crossing boundaries

The room for error seemed to increase when boundaries were “crossed” or where there was a transition between one type of service user to another, from one service to another or from one geographic area to another. When boundaries were crossed, individuals were often seen as someone else’s responsibility and fell out of sight and/or were deemed to pose a lower risk, or to be experiencing decreased vulnerability. Additionally, transitions could result in information being lost. This theme appeared in the following reviews: DHRs1, 2, 3, 4, 5, 8 and 9; APRs2, 3, 4, 5 and 6; and MHHRs3 and 4.

Coding the reviews revealed many types of transitions or boundaries being crossed, often several within a single review. Children, for example, were considered less vulnerable when crossing the boundary into adulthood. Indeed, the transition from child to adult services was not necessarily well-managed and actually resulted in an increased vulnerability because service provision was being changed and because the individual (now an adult) was deemed to be able to cope as they were now no longer “vulnerable” (their “vulnerability” status was tied to their being a child) (see Brown, 2015). In APR3, a former looked after child was convicted of serious sexual offences against a minor. He was in a secure estate and within that secure estate was moved three times. After the third move, after his 18th birthday, he hung himself. There was little recognition that he was still vulnerable after turning 18. His movements within the secure estate were not particularly well-managed and information did not always follow him on his moves within the secure estate. Indeed, mental health concerns were raised but not shared. What was particularly worrying about this case was that a copy of his psychiatric report was not received until after his death.

In DHRs1, 5, 8, 9 (and also APRs4 and 6), the crossing of boundaries was particularly salient when individuals were moving from one geographical area to another[7]. Indeed, it seemed that perpetrators could evade their past by moving to a new location. What is particularly problematic here was that information did not follow the perpetrator or the victim, was not shared across borders, and/or was not readily accessible. In DHR8, three generations were killed in a house fire: the grandmother, her 17-year-old daughter (a mother), and a young baby. In this particular case, the perpetrator had crossed geographic boundaries: he had previously lived in England and had moved to Wales; his previous history of fire-setting and threats of arson in England was not known in Wales. This case also contained a second boundary crossing: the 17-year-old mother was not recognised as a vulnerable child and the only point that the social services stepped in was when the baby was born. Their intervention was based primarily on concern for the baby rather than the mother (herself a 17-year-old child).

Transitioning from one type of service user to another also challenged agency responses. In MHHR4, the perpetrator had recently been released from prison into a hotel; however, he had not been provided with medication or appointments for his mental health problems. Similarly, in MHHR3, the patient was discharged from a psychiatric hospital and was without accommodation. His transition into the community served to directly increase the likelihood of recurring problems. This was also evident in APR5: the deceased’s condition had improved whilst in hospital (where support was readily available) but deteriorated quickly when moving back into the community. There seemed to be a lack of recognition of how being transferred into the community could increase an individual’s vulnerability. As Preston-Shoot (2018, p. 84) noted, discharge from hospital is a “pivotal moment”.

Theme 2: hoodwinking

Hoodwinking refers to individuals disguising or manipulating their presentation of self. Numerous reviews illustrated how individuals would attempt to appear more benign or better able than they actually were (DHRs2, 4, 5, 7, 8; APRs2, 3; MHHRs2, 4). This was especially evident in DHRs, where abuse was often minimised by perpetrators as well as professionals (e.g. recording explicit disclosures of abuse as “marital /relationship difficulties”). When the professional and perpetrator knew each other socially, this further blurred the picture and could be viewed as collusion. For example, in DHR2, the perpetrator’s abusive behaviour was regularly recorded as “marital difficulties”. The perpetrator was viewed as vulnerable because of mental health and addiction problems, and he seemed to present these difficulties as a way of masking his abuse. Meanwhile, his partner (the victim) was seen as his carer. In addition, he attended the same social club as his GP, which may have been a further disguising factor.

Some perpetrators used their difficulties to frame themselves as victims or patients and deflect attention away from their abusive behaviour. One perpetrator (DHR5) told the victim as well as his previous partners that he suffered from post-traumatic stress disorder from his military service to garner sympathy, although this had never been diagnosed. In this case, the perpetrator’s partial disclosure of some of his previous abuse further increased the trust of his victims; he was described as “hiding in plain sight”. Interestingly, in none of the DHRs, even though coercive and abusive behaviour was prevalent, were any of the perpetrators challenged by a professional about this behaviour. This may reflect a lack of confidence amongst practitioners in recognising and dealing with perpetrators (e.g. DHR4 noted the need to up-skill practitioners). It sometimes seemed as though perpetrators were able to coerce professionals in the same way as their victims. For example, in DHR8, the perpetrator was allowed to stay in the hospital chapel, whilst his partner was in the maternity ward.

In DHR5, the use of online dating sites was highlighted as a means to hoodwink, and a source of particular risk, as the perpetrator was able to create an enhanced impression for the dating site and had access across geographical boundaries to a range of different women of different ages who knew nothing of his past history. This issue resonated with professionals in both focus groups who had experience of working with this situation. Surveillance of social media dating sites is particularly difficult, and it was recommended that warnings about them need to be issued more forcibly.

There was also evidence of hoodwinking in the form of “disguised compliance” (NSPCC, 2014); for example, this involved perpetrators appearing to comply with taking their medication, when in fact they were not (MHHRs2, 4). The term disguised compliance was first used by Reder et al. (1993) and has since gained attention through featuring prominently in Serious Case Reviews (Brandon et al., 2010). As a term devised by social workers, it has been used ubiquitously despite being critiqued for assuming a sophistication in individuals that is seldom present (Hart, 2017). Others have noted that the term may be used to signify social workers being ill-equipped to differentiate between those who are engaging and those who are resistant to intervention (Littlechild, 2013). The current study identified a need for more effective ways to monitor compliance in order to detect non-compliance and the possibility of hoodwinking. In MHHR2, the perpetrator told professionals that receiving medication via injections was making him feel unwell, and he was therefore prescribed oral medication, which he chose not to take. This issue resonated with the focus group participants, especially those working in mental health. Similarly, some mental health patients were seen to be adept at masking their symptoms so that they could avoid detention or further surveillance (MHHRs2, 4). There is a need for professionals to confidently identify and challenge disguised compliance.

Finally, young people (both victims in DHRs8 and 10, and APR2) were seen to disguise their vulnerability by presenting as more mature and able than they really were (Brown, 2015). This meant that professionals attributed them with more agency and ability than they truly possessed and, consequently, less protection was forthcoming. This masking of vulnerability reinforces the need for professional curiosity and challenge (see also Theme 4).

Theme 3: faulty assessment

The assessments conducted by practitioners tended to focus on particular aspects of behaviour, neglecting others, thereby reducing the overall accuracy of the assessment. Furthermore, the clinical picture or the assessment could be blurred or obfuscated by multiple factors. This occurred in the following reviews: DHRs2, 3, 4, 5, 6, 7 and 10; APRs1, 3, 4, 5 and 6; and MHHRs1, 2, 3 and 4.

The presentation of more than one problem (such as mental health, substance abuse and violent behaviour) could result in the individual being wrongly assessed. There was no evidence of perpetrators being actively worked with regarding their abuse: assessments and, therefore, interventions, focussed solely on alcohol or drugs, or mental health. Similarly, faulty assessments of victims were evident in failures of GPs to enquire as to the root cause of their mental health problems (e.g. DHR2). Those who had both mental health problems and abused substances often were not recognised as being mentally ill: the substance abuse was viewed as the cause of the problem rather than a possible means for the individual to deal with the underlying issue (DHR3, MHHRs1 and 3). Thus, the likelihood of faulty assessments increased when practitioners focussed on incidents rather than identifying patterns of behaviour, combined with a poor recognition and management of the full “toxic trio” (Domestic Violence Death Review Team, 2017; Forrester and Harwin, 2008; Taylor and Lazenbatt, 2016).

Assessments also failed to take account of how best to respond to someone who was disengaged or chaotic; frequently, such individuals were discharged from services when they failed to engage as opposed to when their condition actually improved. Mental health services experienced difficulty assessing “aloof” patients and those who rejected their diagnoses. In both MHHRs2 and 4, the perpetrators experienced difficulties in managing their medication and, whilst this should have resulted in a more rigorous response, their issues led to a decrease or removal of services. Through reading the reviews, it appeared that a failure to engage should actually trigger a new assessment and/or greater service involvement rather than case closure (Preston-Shoot, 2018).

Discharge from services also occurred where the individual appeared to be “doing well”. Not only might an individual be discharged from their current service, they also would be assessed as not needing any services, as was the case in APR5 and MHHR2. It seemed that there was a “rule of optimism” (Kettle and Jackson, 2017) whereby it was assumed (or hoped) that the individual was able to cope with their issues and therefore not to be in need of further help, despite previous histories suggesting that relapse was highly unlikely (e.g. DHR6).

Some individuals were assessed as at risk (i.e. vulnerable) rather than posing a risk (i.e. harmful). This was particularly evident in APR4, where the risk posed by a vulnerable adult (who was assessed as lacking capacity) was not considered, despite his history of engaging in sexually harmful behaviour. This adult later committed a serious sexual assault against another resident. There appears to be a tension between the recognition of vulnerability and a recognition of risk (Brown, 2015): whilst it may be difficult to conceptualise risk and vulnerability in tandem, practitioners must be cognisant that an individual could simultaneously present a risk to others and be at risk themselves. Similarly, in DHR3, the perpetrator’s risk of suicide was foregrounded to such an extent that there was no consideration given to the vulnerability of or risk posed to his grandparents, who were providing his care.

Theme 4: tunnel vision

There was a tendency for practitioners to focus solely or predominantly on certain aspects of an individual’s vulnerability or risk, and to exclude or fail to recognise other aspects. This theme appeared in the following reviews: DHRs2, 3, 5, 6, 7, 8 and 10; APRs2, 4 and 5; and MHHRs1 and 3.

Tunnel vision meant that a narrative was constructed and practice would be shaped to fit this particular narrative (Findlay, 2012). In MHHR3, the patient was diagnosed by the psychiatrist as “malingering”; he was seen to be manipulating the situation to remain in hospital, rather than genuinely suffering from psychosis. Although evidence continued to challenge the “malingering” diagnosis, this was never re-evaluated by other professionals. This was also the case in MHHR1 (where the perpetrator was only diagnosed with schizophrenia after the death of the victim) and where a lack of consensus amongst professionals resulted in a view of him as primarily suffering from substance misuse rather than psychosis (see also MHHR3, and Theme 3).

Tunnel vision also was apparent in the lack of recognition that someone’s situation or condition could change over time. The abuse that a victim encounters, for example, does not remain static over time but can escalate and/or manifest in different ways. Abuse was downplayed as merely criminal damage and therefore not seen in the broader context of coercive, controlling abuse (DHR8), or was trivialised as “play-fighting” (DHR10). Physical and mental health could deteriorate over time (DHR6, APR5 and MHHR1). In APR5, the deterioration occurred after release from hospital, whilst in hospital, he had been doing well but upon release his situation and health rapidly deteriorated. Finally, those who have addictions, whilst potentially on the road to recovery during assessment, can relapse.

A problematic consequence of tunnel vision is that the range of options open to the individual tends to narrow rather than to broaden. Cases could become “stuck”; tunnel vision reinforced a particular view of the person, which resulted in a particular set of options being tried. When these did not work, practitioners did not “step outside of the tunnel” to re-evaluate their options and reflect on what type of approach had gone well in the past (i.e. taking a strengths-based approach) and therefore how they might adapt their practice so that it was more palatable or acceptable for the individual. Indeed, focus group participants felt that, due to limited time and resources, there was a tendency to pigeonhole individuals, particularly where there is a volume of contact. In such instances, the approach was to assume that the same problem had emerged yet again, without fully appreciating the ways in which it might be different. Practitioners recognised the need to “step back” in order to be able to effectively consider and evaluate the whole case but felt that there was a tendency to try to identify and deal with the immediate problem, or what was perceived as the immediate problem.

Theme 5: knowledge

This theme is positioned last, as it ran through many of the reviews (DHRs2, 3, 6, 7, 8, 10; APRs2, 4, 5, 6; MHHRs1, 2, 3) and contributes to the other four themes already discussed. It therefore underpins and is central to the findings. First, from reading all of the reviews, it was evident that some sources of knowledge were privileged and therefore dominant (Preston-Shoot, 2018). Professional knowledge took precedence over personal knowledge. This was particularly the case for medical knowledge, where much time was spent searching for a diagnosis (see Theme 3: faulty assessment), and once decided upon by psychiatrists (DHR6 and MHHRs1, 3, 4) was not challenged or reviewed (see Theme 4: tunnel vision). In particular, the view of the psychiatrist was revered despite this often being the person who had limited information and/or had spent the least amount of time with the individual (typical appointments were quite short). In several cases, “locum” doctors (those who temporarily fulfil the duties of another) were key decision makers, although, due to their role, they were inherently less knowledgeable of the full background history (DHR2; APR2; MHHR4).

In contrast, the views of families or para-professionals were not often drawn upon or were seen as less credible in contributing to assessments of risk, even though they may see the individuals concerned on a daily basis and in their homes (Winter and Cree, 2015) and therefore may be far more attuned and alert to changes in condition and presentation. In MHHR1, the para-professionals took the client to the GP on numerous occasions highlighting their concerns, and whilst this information was fed “upwards” to mental health professionals, information about assessment and treatment was not sent back down to those working with the individual “on the ground”. Para-professionals (including third sector workers) and family members were not invited to decision-making meetings. Clients were often de-coupled from their families and seen in isolation (Featherstone et al., 2014), despite the family being the lynchpin in providing professionals with information about the client (MHHR2, DHR3). Families often highlighted deterioration and increased risk, and, for example, advised against release from hospital (DHR6; MHHR2) but often were not listened to. Focus group participants noted that families could be seen as part of the problem or as a “nuisance”, as was the case in APR6. It was notable, however, especially in the process of coding DHRs, that family members and the information they could provide was seen as central, when it had not been during the course of the case. Furthermore, in none of the cases were any children seen alone as has been highlighted for many years as best practice in child protection (Munro, 2011). Thus, the knowledge they could have contributed was lost and they were “invisible” in the review process (Ferguson, 2017).


The reading and analysis of the three different types of death reviews is both unique and innovative and has not been undertaken before. Thematic coding of this diverse sample of reviews allowed for an “aerial” view to be taken to determine patterns and cross-cutting themes that could not necessarily have been gleaned from analysis of a single type of review. Another distinctive aspect of this study was to have a research team of professionals from three different disciplines – criminology, social work and law – to code and analyse the data. Each evaluated the reviews from their own disciplinary perspective and thus applied a different lens to understand the features of the case. This helped the team to avoid “silo thinking” and the privileging of one particular discipline over another, and facilitated the corroboration of findings through triangulation. Our research reinforces the importance of taking a multi-disciplinary and comparative methodological approach to the study of death reviews. Future research should build on this by adding further disciplinary perspectives (e.g. psychology, psychiatry, mental health or medical) and additional types of reviews (e.g. child practice reviews, serious further offence reports from probation, etc.).

Our research identified very similar themes from each of the review documents, and many of the same themes also emerged across the sample of reviews, regardless of whether the type was a DHR, APR or MHHR. Although the examples are diverse, the common thread drawing them together is that they involved agencies responding to vulnerable people in a way that may be reasonably expected to have been better and should be improved upon. However, each type of review has its own disciplinary and practice-based audience that reinforces a particular conceptualisation of “vulnerability” (Brown, 2015). Thus, the tendency is to identify “lessons to be learned” which appear to be bespoke to particular issues (e.g. domestic violence, learning disabilities, lack of capacity, mental health, etc.) rather than seeing the common challenges across disciplinary and agency practices and devising more substantive and innovative ways to rise above them.

The methods of this study demonstrate some of the benefits of working across professional disciplines. We as researchers learnt from each other and reflected on the differing conceptual ideas often discussed in one field, but rarely considered in another. For example, disguised compliance is often used in social work (NSPCC, 2014), but not as regularly in law; hoodwinking is understood in criminology (Croall, 2001) but rarely used in other disciplines; tunnel vision is used in law and psychology (Findlay, 2012) but less in social work; and the privileging of certain forms of knowledge is understood within the discipline of law (Roy, 2014) but as not readily conceptualised or recognised in others. A one-dimensional approach was avoided through this healthy cross-fertilisation of ideas, which also strengthened the learning from this study.

Our research suggests several implications for policy and practice. First, the similarity of the key themes identified across reviews, corroborated by the discussions in the focus groups, provides evidence to suggest that having separate reviewing processes may not be the most efficient and productive way to promote multi-agency and multi-local authority learning from these tragic events. As previously discussed, each review is commissioned and held separately, as specific to that context, situation, team or setting. This arrangement does not encourage or facilitate the spread of knowledge across local authorities and disciplines. Multiple, separate reviewing processes inhibits the learning and “reading across” these incidents. This reinforces the divides between professional groups, almost as if they inhabit different “planets” (Hester, 2011), rather than allowing practitioners to make connections across their different practice settings.

Second, the duplication of evidence gathering, where single incidents trigger numerous reviews (e.g. both MHHR and DHR) would seem to be unwieldy, unfair to the family and not in the spirit of multi-professional, inter-agency working. This was highlighted as a concern in both focus groups and in survey feedback: for example, “I feel that the reviews work well but the issue is the impact of multiple reviews on families and other agencies in the duplication of work” (No. 7). These separate processes could be seen to be potentially deepening the silos in which people work and are expensive and time consuming.

Third, the heterogeneity of reviews in terms of their structure, mechanisms of governance, case identification processes and inclusion criteria have been noted (Bugeja et al., 2017). We also found that the quality and scope of the reports differed markedly, consistent with past research[8]. The reviews look back over a range of differing periods, from 2 months prior to death (APR6) to 20 years (DHRs1, 2; MHHR4). Some were of far better quality in terms of their level of detail and analysis than others and writers of reviews would benefit from guidelines, training, a consistent standard and benchmarking. Unpredictable variability both within and across types of reviews was highlighted as a barrier to learning in the focus groups. Establishing a single type of death review might help to alleviate this problem. In addition, many participants expressed a desire for a more centralised, proactive, structured approach to facilitate learning from reviews:

The ideal situation would be if an overarching body could take ownership of collating reviews, extracting and putting the learning in to themes, disseminating the learning and ensuring that this was being acted on.

(No. 11)

I think the findings need to be collated centrally and fed back, so that we can all learn from them, not just the services involved.

(No. 4)

To raise the profile when these are published, not only for professionals but the wider communities. To ensure clear access to learning experiences for all those who may be involved in similar situations.

(No. 19)

In conclusion, there is a need for working together across disciplines both when working with vulnerable individuals in practice but also when reviewing what went wrong. If the aim is for multi-disciplinary working to best support vulnerable individuals as enshrined in the Social Services and Well-being Act (2014), then it is important that this continues and is demonstrated when professionals look back and reflect on lessons learned. This requires a confident, consistent and open approach where each profession can learn from the other. A streamlined, joined-up approach will lessen the risk of loss of information across disciplines, tunnel vision and the privileging of certain forms of knowledge occurring within the reviewing process. This shared approach should extend to all professionals, and the resulting learning materials should be held in a central repository, readily available in a variety of accessible formats.



Boarding out was the practice of placing abandoned or neglected children in the long-term care of a family for a weekly allowance, as an alternative to the workhouse or orphanage.


DHRs were established on a statutory basis under the Domestic Violence, Crime and Victims Act 2004 to review the circumstances in which the death of a person aged 16 or over has resulted from violence, abuse or neglect from a person to whom he or she was related or with whom he or she was or had been in an intimate personal relationship, or a member of the same household as him/herself.


MHHRs are commissioned and carried out by Health Inspectorate Wales (HIW) after homicides are committed by individuals known to mental health services in Wales.


APRs are commissioned by regional Safeguarding Boards and take place after an “adult at risk” has died; sustained potentially life threatening injury; or sustained serious and permanent impairment of health (Welsh Government, 2016). Under Part 7 of the Social Services and Well-being Act (2014), an “adult at risk” is defined as a person who: is experiencing or is at risk of abuse or neglect; has needs for care and support (whether or not the authority is meeting any of those needs); and, as a result of those needs, is unable to protect himself or herself against the abuse or neglect or the risk of it.


One recent report compares adult practice reviews and child practice reviews (Pachu and Jackson, 2018); however, both of these are delivered in accordance with the Social Services and Well-being Act (2014) and thus share a focus on social work and social care.


These are estimations based on various sources (e.g. Pachu and Jackson, 2018; HIW website) and in discussion with key stakeholders (e.g. Welsh Government and the National Independent Safeguarding Board). Due to the lack of a central repository for death reviews, it is difficult to establish conclusively the total number of different types of reviews that have taken place over a specific time period.


This may present difficulties for the Domestic Violence Disclosure Scheme (otherwise known as Clare’s Law).


Quality can also be impacted upon issues with funding, national data and/or legislation (Bugeja et al., 2017; Mazzola et al., 2013; Vincent, 2013).


Brandon, M., Bailey, S. and Belderson, P. (2010), “Building on the learning from serious case reviews: a two year analysis of serious case reviews 2007-2009”, Department for Education, Research Report No. DFE-RR040, London.

Brown, K. (2015), Vulnerability and Young People: Care and Social Control in Policy and Practice, Policy Press, Bristol.

Bugeja, L., Dawson, M., McIntyre, S. and Poon, J. (2017), “Domestic/family violence death reviews: an international comparison”, in Dawson, M. (Ed.), Domestic Homicides and Death Reviews: An International Perspective, Palgrave, London, pp. 3-26.

Croall, H. (2001), Understanding White Collar Crime, Open University Press, Buckingham.

Domestic Violence Death Review Team (2017), “New South Wales Domestic Violence Death Review Team report 2015-17”, Domestic Violence Death Review Team, Sydney.

Elliott, D. and McGuiness, M. (2002), “Public inquiry: panacea or placebo”, Journal of Contingencies and Crisis Management, Vol. 10 No. 1, pp. 14-25.

Featherstone, B., White, S. and Morris, K. (2014), Re-imaging Child Protection: Towards Humane Social Work with Families, Policy Press, Bristol.

Ferguson, H. (2017), “How children become invisible in child protection work: findings from research into day-to-day social work practice”, British Journal of Social Work, Vol. 47 No. 4, pp. 1007-23.

Findlay, K.A. (2012), “Tunnel vision”, in Cutler, B.L. (Ed.), Conviction of the Innocent: Lessons from Psychological Research, American Psychological Association, Washington, DC, pp. 303-23.

Forrester, D. and Harwin, J. (2008), Parents who Misuse Drugs and Alcohol in Social Work and Child Protection, Wiley-Blackwell, London.

Hart, P. (2017), “Disguised compliance – or undisguised nonsense?”, Family Law Week, available at: (accessed 7 February 2019).

Health Inspectorate Wales (2016), “Independent external reviews of homicides: an evaluation of reviews undertaken by healthcare inspectorate wales since 2007”, available at: (accessed 7 February 2019).

Hester, M. (2011), “The three planet model: towards an understanding of contradictions in approaches to women and children’s safety in contexts of domestic violence”, British Journal of Social Work, Vol. 41 No. 5, pp. 837-53.

Home Office (1945), “Report by Sir Walter Monckton KC on the circumstances which led to the boading out of Nennis and Terence O’Neill at Bank Farm, Minsterley, and the steps taken to supervise their welfare”, Presented to the Secretary of State for the Home Department to Parliament by Command of His Majesty May 1945, His Majesty’s Stationery Office, London.

Home Office (2016), “Domestic homicide reviews: key findings from research”, London, available at: (accessed 7 February 2019).

Hopkins, G. (2007), “What have we learned? Child death Scandals since 1944”, Community Care, 10 January, available at: (accessed 7 February 2019).

Kettle, M. and Jackson, S. (2017), “Revisiting the rule of optimism”, British Journal of Social Work, Vol. 47 No. 6, pp. 1624-40.

Littlechild, B. (2013), “Working with resistant parents in child protection: recognising and responding to the risks”, paper presented at the 2012 world Congress IASSW/IFSW/ICD, Stockholm, July 2012.

Mazzola, F., Mohiddin, A., Ward, M. and Holdsworth, G. (2013), “How useful are child death reviews: a local area’s perspective”, BMC Research Notes, Vol. 6 No. 6, pp. 295-301.

Munro, E. (2011), “The Munro review of child protection: final report”, A Child Centred System, Department for Education.

Nash, M. and Williams, A. (2008), The Anatomy of Serious Further Offending, Oxford University Press, Oxford.

NSPCC (2014), “Disguised compliance: learning from case reviews”, available at:

Pachu, D. and Jackson, C. (2018), Learning from Reviews, Public Health Wales, Cardiff, available at:

Preston-Shoot, M. (2018), “Learning from safeguarding adult reviews on self-neglect: addressing the challenge of change”, Journal of Adult Protection, Vol. 20 No. 2, pp. 78-92.

Reder, P., Duncan, S. and Gray, M. (1993), Beyond Blame: Child Abuse Tragedies Revisited, Routledge, East Sussex.

Robinson, A., Rees, A. and Dehaghani, R. (2018), Findings from a Thematic Analysis of Reviews into Adult Deaths in Wales, Cardiff University, Cardiff, available at:

Roy, S. (2014), “Privileging (Some forms of) interdisciplinarity and interpretation: methods in comparative law”, Oxford University Press and New York School of Law, Vol. 12 No. 3, pp. 786-807.

Salter, M. (2003), “Serious incident inquiries: a survival kit for psychiatrists”, Psychiatric Bulletin, Vol. 27 No. 7, pp. 245-7.

Sidebotham, P., Brandon, M., Bailey, S., Belderson, P., Garstang, J., Harrison, E., Retzer, A. and Sorensen, P. (2016), Pathways to Harm, Pathways to Protection: A Triennial Analysis of Serious Case Reviews 2011-2014, Department for Education, London.

Social Services and Well-being Act (2014), available at:;; (accessed 6 February 2019).

Stevens, M., Martineau, S., Norrie, C. and Manthorpe, J. (2017), Helping or Hindering in Adult Safeguarding: An Investigation into Practice, Social Care Workforce Research Unit, London.

Taylor, J. and Lazenbatt, A. (2016), Child Maltreatment and High Risk Families, Dunedin Academic Press, Edinburgh.

Vincent, S. (2013), “Child death review processes: a six‐country comparison”, Child Abuse Review, Vol. 23 No. 2, pp. 116-29.

Welsh Government (2016), “Working together to safeguard people: volume 3 – adult practice reviews”, available at: (accessed 7 February 2019).

Winter, K. and Cree, V. (2015), “Social work home visits to children and families in the UK: a Foucauldian perspective”, British Journal of Social Work, Vol. 46 No. 5, pp. 1175-90.

Further reading

Mental Health Act (2005), available at: (accessed 7 February 2019).

Corresponding author

Amanda Lea Robinson can be contacted at:

About the authors

Amanda Lea Robinson is based at the School of Social Sciences, Cardiff University, Cardiff, UK.

Alyson Rees is based at the School of Social Sciences, Cardiff University, Cardiff, UK.

Roxanna Dehaghani is based at the School of Law and Politics, Cardiff University, Cardiff, UK.