Abstract Risk-assessment models are widely used in casework involving vulnerable children and families. Internationally, there are a number of different kinds of models with great variation in regard to the characteristics of factors that harm children. Lists of factors have been made but most of them give very little advice on how the factors should be weighted. This paper will address the use of risk-assessment models in six different Danish municipalities. The paper presents a comparative analysis and discussion of differences and similarities between three models: the Integrated Children’s System (ICS), the Signs of Safety (SoS) model and models developed by the municipalities themselves (MM). The analysis will answer the following two key questions: (i) to which risk and protective factors do the caseworkers give most weight in the risk assessment? and (ii) does each of the different models ensure a holistic assessment (which is required by law in Denmark)? The study contributes to the discussion about the use of risk-assessment models and whether these newer models have actually made a difference in what is assessed when looking at families at risk. Child protection, children and families, comparative social work, methods, social work and IT Introduction Previous research has discussed the use of risk-assessment models in casework involving vulnerable children and families. In an English context especially, the Integrated Children’s System (ICS) has been scrutinised with mixed conclusions regarding whether the model is fit for its purpose (Cleaver et al., 2008; Holmes et al., 2009; Shaw and Clayden, 2009; Shaw et al., 2009a). Nevertheless, the Danish government recommends the ICS model and requires that all municipalities should have the ICS information technology (IT) system or equivalent to ensure a systematic and standardised approach as well as uniformity and a holistic view of the children and families at risk (Servicestyrelsen, 2011). However, research has shown variation between different models (Munro et al., 2013). There is also variation in what information the social workers collect and on which they thereby base their assessments even when applying the same model (Cleaver et al., 2008; Holmes et al., 2009; Shaw et al., 2009a). This article will address the risk assessment of vulnerable children and families, and especially focus on the use of three different types of risk-assessment model. The purpose of this article is to reflect critically on the use of different risk-assessment models and whether the use of these models gives different outcomes regarding risk and protective factors and the achievement of a holistic approach. It is a comparative study using a mixed-methods design. I am using six Danish municipalities as my research case and comparing three different risk-assessment models. These will be introduced at the end of this section. As stated in the beginning, risk assessment in child protection has been discussed for many years in research as well as on the practice and policy level. Whereas social workers historically based their assessments mainly on knowledge, experience and intuition, new ways of assessing risk in a more standardised way have been introduced (Baird et al., 1999; Egelund and Sundell, 2001). The literature have pointed at different explanations for these changes, whereas one explanation is the appearance of the management strategy New Public Management (NPM) in the public sector (Munro, 2010). Another explanation is found partly in the uncertainty of positive progress in child protection, such as for looked after children (Vinnerljung, 1996; Levin, 1998). At the same time, social workers in Denmark, the UK and the USA (English and Pecora, 1994; Rossi et al., 1999; Egelund and Sundell, 2001; Munro, 2008) experienced criticism of the lack of substantial (good-enough) risk assessments—if any risk assessment at all—not enough participation of children and/or the families in the process, no clear description of the aims with the interventions, not enough explicit use of knowledge of children at risk and about which interventions work, and lack of methodology in regard to doing the risk assessments, which also meant lack of uniformity in the case management (Den Digitale Taskforce, 2004). The criticism and the discussions about uncertainty of positive progress are also linked to the discussions of evidence-based practice in social work (Høybye-Mortensen, 2011). The final explanation is the increase in the financial expenses used on interventions for children at risk and for looked after children. From 1993 to 2001 in Denmark, the cost of placements increased by nearly 40 per cent, even though the percentage of children placed in out-of-home care was the same (1 per cent) (Andersen and Jensen, 2016). These increases created financial pressure on the Danish municipalities that finance the expenses. This financial increase legitimised a discussion about whether the money was spent sensibly (Egelund and Hestbæk, 2003). These three different explanations show an area that has been under a lot of focus and political pressure as well as media coverage. The consequence, internationally and in Denmark, is increasing regulation from the government and standardisation of child-protection services (Hood et al., 2001; Høybye-Mortensen, 2011). The changes have also led to the use of different risk-assessment models. However, such risk-assessment models may fail to give the social workers the support they need. Critics, for instance, discuss the inaccuracy of risk-assessment models (Camasso and Jagannathan, 2000; Egelund and Sundell, 2001). I will now briefly describe the context of my study as well as the models I investigate. Danish context Some of the essential characteristics of the Danish way of assessing risk are the intentions to balance the focus between resources and problems (Kildedal and Nordstoga, 2011, p. 26), and that the social workers must apply a holistic view in the risk assessment, which means assess ‘all the way around’ a child. This also includes looking at factors concerning the surroundings of the child, such as the local society or the family network (Folketinget, 2010–11, item 277). However, the law consists of six sub-points: development and behaviour, family, school, health, leisure time and friendship, and other relevant circumstances (Law on Social Services § 50, stk. 2), on which, if relevant, the social worker must base their assessment. The Child’s reform was launched in 2011, changing the daily practice of child protection in Denmark. The changes have especially concerned the use of methods in the risk-assessment process. From being characterised by methodically freedom, all Danish municipalities are after the reform obliged to digitise the case handling, which has resulted in a standardisation of casework method and a centralised intention of uniformity across the Danish municipalities. As described in the introduction section, the Danish government recommends the ICS as one of the mutual risk-assessment models for the municipalities to use (Servicestyrelsen, 2011). However, the smaller municipalities do not have the financial resources to develop new models or new IT systems—why ninety out of ninety-eight Danish municipalities have implemented the ICS. My empirical data were collected in 2013. At that point, two out of the six participating municipalities used the ICS. One used the SoS model and three used models of their own (in the following, these will be called the municipality model (MM)). These three MMs are historically how the municipalities used to do risk assessments before the new, more standardised approaches entered the field. I analyse these three ‘homemade models’ as one model, as their approaches are alike. In the following, I will briefly introduce the different models with regard to their focus and theoretical background and their developers. Integrated Children’s System (ICS) Focus The ICS is a risk-assessment method developed for the work of dealing with children, young people and families at risk (Bell et al., 2007; Shaw et al., 2009a). The main motivation of the model is to put the child’s needs in focus in the assessment. In assessing the child’s needs, the model has three criteria: (i) the child or the young person’s developmental needs, (ii) the competences of the parents and (iii) the family circumstances, including the surroundings of the family. Theoretical background The ICS is based on a systemic perspective and has its reference point in Bronfenbrenner’s (1979) ecological development model. The model is conceptualised in a triangle where the three different focus areas are placed on each side and are to be understood in a mutual interplay. The triangle is seen as an important tool in the casework. It is intended that this tool will support caseworkers to incorporate essential factors that influence and affect the child’s situation and thereby ensure a holistic approach. Developers The ICS was developed in England in light of many years of development and research work, and can be seen as a progression towards a more evidence-based practice and towards further standardisation and use of manual-based tools in practical social work (Shaw et al., 2009a). Signs of Safety (SoS) Focus One of the primary purposes of the SoS model is to offer a structure or a framework that ensures integration of the family’s as well as the professional’s knowledge. By that, its developers argue for a co-operative partnership between the family and the social worker, where specific mutually understood goals are the focus (Turnell, 1999). The SoS is conceptualised in a map with six practice principles, which also highlights the primary purposes of focusing on safety via looking at the family’s strengths and resources as well as being solution-oriented. Theoretical background The SoS is a safety- and solution-oriented approach (Turnell, 1999). It is also based on looking at resources and strengths within the family or the family’s network as a basis for assessing the safety of the child at risk (Turnell et al., 2013). Developers The SoS was developed in Australia in the 1990s by Andrew Turnell and Steve Edwards, both of whom had been working within practical social work for many years. The model was developed in co-operation with more than 150 practitioners, which makes it very practice-oriented (Turnell et al., 2013). Municipality model (MM) Focus The MM is not one explicit model but can be characterised as a schematic approach which most often is embodied in an electronic form with fixed focus points derived from the six sub-points provided in the Danish Law on Social Services § 50, paragraph 2. As described in the Danish context, the philosophy is that the points in the MM ensure a holistic approach. The points listed in the form explicitly focus on the problems and the resources of the family and the child, not the family’s network. The form will to some degree ensure that all the social workers share the focus in the risk assessment and a common structure. Theoretical background Two out of the three municipalities state that their models belong in general to a systemic narrative approach but there is no theoretical background to the models. Developer Each municipality has developed an approach to the risk assessment which is supported electronically (Den Digitale Taskforce, 2004). Methodology The data used for this study are quantitative and consist of an internet-based survey which was distributed to social workers working with risk assessment of vulnerable children and families in the six municipalities. The survey was produced with numerically responding categories and was pilot-tested before it was sent out. All ethical guidelines are followed in regard to Danish standards and all participants in the study have given a verbal or written consent. My pilot study clarified that the risk factors in my questionnaire had to be similar to the categories in the questionnaire made by the Danish Social Appeals Board. Social workers are required by law to fill out this questionnaire when they place a child in out-of-home care. Making my categories of risk factors similar ensured that the respondents were familiar with the categories and that the categories reflected the categories the social workers usually had to address. I therefore made a compromise between what research within the area of risk factors describes as significant and the more pragmatic way of making sure the respondents knew the categories. However, there was a significant overlap between my list of risk factors and the list made by the Danish Social Appeals Board. The protective factors in my questionnaire are based on international research about protective factors (Rutter, 1990; Rak and Patterson, 1996; Werner and Smith, 2001). In my questionnaire, I divided the risk and protective factors into three categories inspired by the ecological understanding (Bronfenbrenner, 1979) and the questionnaire produced by the Danish Appeals Board. I called the categories ‘factors linked to the child’, ‘factors linked to the family’ and ‘factors linked to the society’. The questionnaire consists of thirty-eight risk factors and nineteen protective factors, and the social workers were asked to tick off as many as they had found relevant for their assessment. When answering the questionnaire, the social workers had to select a specific case in which they had made a risk assessment. The social workers were informed that they had to choose a case where they had made the risk assessment within the last six months. The social workers were asked to answer the questionnaire based on their existing knowledge as being the social worker connected to that specific family—not to gather more information about the family or the risk assessment. In accordance to the questions about the risk and protective factors, the social workers were asked to tick off all the factors they have given weight in the assessment. As all risk assessments are recorded electronically, the social workers had the possibility to reopen the risk assessment at the computer and answer the questionnaire at the same time. To ensure comparability in the data, the case had to have the following characteristics: It must be a first-time risk assessment of the child. The case must involve a child aged between seven and twelve years. The case must be a social case (not a case concerning problems in regard to disability). The case should have resulted in recommendation of one of the following interventions: family therapy (see § 52,3.3 in the Law on Social Services), respite care for the child (see § 52,3.5 in the Law on Social Services) or out-of-home care (see § 52,3.7 in the Law on Social Services). Focusing on cases where the social worker has recommended family therapy, respite care for the child or out-of-home care narrows down the focus to cases where the problems are of a severe degree. This means that cases where social workers have recommended smaller interventions for families with fewer problems are not a part of these data. Choosing to limit the age span of the children from seven to twelve years of age also ensures comparability in the data. Earlier research has shown that risk and protective factors differ, depending on the child’s age. When children are starting school, they are exposed to new challenges and standards. Conversely, the school is also a place where children develop and find new strengths (Werner and Smith, 2001). These strict criteria limited the scope for social workers to ‘cherry-pick’ cases they had specific concerns about, such as particularly severe cases, which could confound the analysis of different risk-assessment models. With the four common characteristics, my aim was also that only the models would differ, not the specific case. This means that the patterns in the findings will be due to the model, not differences regarding the children or families. My questionnaire had fifty-three respondents. One respondent is equivalent to a social worker who has answered one questionnaire based on one case. One case is one family which has been through a risk assessment in the Danish child-protection system. This means that my data consist of assessments of fifty-three families at risk involved with the Danish child-protection system. My analysis is based on these data and discusses the practical use of the models in a comparative view. However, it is important to mention that it is a small data-set so it has not been possible to make any advanced tests or significant analysis. In spite of this, my data do show some clear patterns on which I will elaborate in the following section. Findings and discussion I will present my findings as two topics: (i) the risk and protective factors that are weighted by caseworkers when doing risk assessments and (ii) whether the different models ensure a holistic assessment (as required by the Law on Social Services) in the different cases. Risk and protective factors First, I will elaborate on the relationship between risk and protective factors in the use of the different risk-assessment models. Table 1 gives an overview of the weighting of risk and protective factors compared with each other in regard to the three different models. The number of factors is the average estimation per case and is based on the number of risk and protective factors the social workers have ticked in the questionnaire when asked about which factors they have used in their assessments. Table 1 Risk and protective factors split into the three different models ICS SoS MM Risk factors 6.1 per case 5.3 per case 7.3 per case Protective factors 3.0 per case 5.5 per case 4.3 per case Ratio, Risk/protective 2.0 <1.0 1.7 No. of cases 16 17 20 ICS SoS MM Risk factors 6.1 per case 5.3 per case 7.3 per case Protective factors 3.0 per case 5.5 per case 4.3 per case Ratio, Risk/protective 2.0 <1.0 1.7 No. of cases 16 17 20 As shown in Table 1, the social workers using the ICS ticked off twice as many risk factors as protective factors when asked what they have weighted in their risk assessment of a vulnerable child. This is almost the same regarding social workers using the MM, who ticked off almost twice as many risk factors as protective factors. Linked to the risk assessments, the table shows that risk factors are the main focus for these two models. On the contrary, the social workers using the SoS ticked off more protective factors than risk factors. Relatively, this indicates that the SoS has a very direct focus on protective factors. This finding is interesting especially when recalling that one of the focuses of the ICS was to establish more room for using protective factors in the risk assessments. Protective factors If we look at which particular factors the social workers say they weight when doing their assessment, Figures 1, 2 and 3 show which protective factors the social workers are using in the assessments. The figures include the total numbers of protective factors used in the questionnaire. Figure 1 View largeDownload slide ICS, protective factors (percentages of total numbers of ICS cases on y-axis) Note: the percentage is based on N = 16. Figure 1 View largeDownload slide ICS, protective factors (percentages of total numbers of ICS cases on y-axis) Note: the percentage is based on N = 16. Figure 2 View largeDownload slide SoS, protective factors (percentages of total numbers of SoS cases on y-axis) Note: the percentage is based on N = 17. Figure 2 View largeDownload slide SoS, protective factors (percentages of total numbers of SoS cases on y-axis) Note: the percentage is based on N = 17. Figure 3 View largeDownload slide MM, protective factors (percentages of total numbers of MM cases on y-axis) Note: the percentage is based on N = 20. Figure 3 View largeDownload slide MM, protective factors (percentages of total numbers of MM cases on y-axis) Note: the percentage is based on N = 20. The seven black columns indicate individual factors regarding the child, the nine columns with the stripes indicate factors regarding the family and the three grey indicate factors regarding the society. If all response possibilities were weighted equally by the social workers, then each category would have had equal numbers of responses. The figures show the distribution of how the social workers selected the various categories. It is shown as percentages of the total numbers of cases in regard to the specific model on the y-axis. In the three different figures, the categories are listed in descending order of responses—that means that the categories are the same in the three figures but in different orders. The reason for this order is to show top choices for each model. Figures 1, 2 and 3 clearly show differences in how the social workers working with the three different models say they weight the protective factors. Figures 1 (ICS) and 2 (SoS) show a focus mainly on factors regarding the child and Figure 3 (MM) shows that the main focus is the family. That the child is in focus in Figure 1 might indicate that the intention of the ICS has been achieved. As written in the Danish handbook for this model, the intention is that the ICS should have an explicit focus on the welfare of the child (VIA University College et al., 2012). Focusing on factors which describe conditions in relation to the child can be viewed as being a ‘step on the way’ to having a more child-oriented approach which, with reference to earlier studies (Egelund, 1997), shows a change from having the family at the centre of the assessment. Having the child in focus is also one of the key elements in the latest Danish reform on children at risk which states that the child should be at the centre of the risk assessment (Servicestyrelsen, 2011). When looking at Figure 1, it also becomes obvious that the clear focus on the child means that the rest of the ICS triangle (family and the surroundings) is not as much in focus when looking at protective factors. This leads to an interpretation that factors regarding the family and the society around the family are not the main focus for the assessment of the child. I will elaborate on this in the next section, which concerns the holistic approach. When looking at Figure 2, it becomes clear that the factors are more equally distributed in the SoS model. This indicates first of all that the social workers who answered the questionnaire based on the SoS ticked off more protective factors. This is also shown in Table 1, where the average number of protective factors per case is 5.5 in contrast to the average number of risk factors which is 5.3. It also indicates that the protective factors are more spread, which may show that the social workers using the SoS do not agree on which protective factors they should consider when doing the assessment. On the other hand, the handbook for the SoS model states that the model’s primary focus is to ensure that the family’s understanding of the problem and the solutions are both in focus together with the professional’s (Turnell, 1999). So one could argue that the spread could be caused by the families being more involved in deciding which factors the social workers should weight in the risk assessment. Turning to Figure 3, we find a clear focus on the family factors. This finding is not surprising when we look at how children at risk have been regarded in Denmark historically. Previous research has made it clear that social work with vulnerable children has been affected by a psychiatric way of thinking. The argument is that social workers, in their understanding of children’s problems, have been strongly oriented towards the parents’ deviations. In contrast to this is a lack of orientation towards the ecological as well as the sociological perspective (Egelund and Sundell, 2001). This perception about understanding the children’s problems on a micro or family level is shared by families who have been through the risk-assessment procedure. In an earlier Scandinavian study, parents reported that the risk assessment was especially concentrated on factors related to the family and their life together, the relationship between parents and children, and factors regarding the mother’s mental health (Sagatun, 2011). With these studies in mind, an explanation of Figure 3 could be that the MM retains the more traditional approach to risk assessments, where the social workers look at the parents and the parents own upbringing, etc. Risk factors The pattern which was obvious in the protective factors does not exist when looking at the risk factors. Figures 4,5 and 6 show the twenty most frequently used risk factors, according to my respondents, out of a list of thirty-eight risk factors. Sixteen of the total number of risk factors concerns the child, eighteen the family and four the society. To ensure the figure is illustrative, I have chosen only to elaborate on the twenty most weighted risk factors in each model. Figure 4 View largeDownload slide ICS, risk factors (percentages of total numbers of ICS cases on y-axis) Note: the percentage is based on N = 16. Figure 4 View largeDownload slide ICS, risk factors (percentages of total numbers of ICS cases on y-axis) Note: the percentage is based on N = 16. Figure 5 View largeDownload slide SoS, risk factors (percentages of total numbers of SoS cases on y-axis) Note: the percentage is based on N = 17. Figure 5 View largeDownload slide SoS, risk factors (percentages of total numbers of SoS cases on y-axis) Note: the percentage is based on N = 17. Figure 6 View largeDownload slide MM, risk factors (percentages of total numbers of MM cases on y-axis) Note: the percentage is based on N = 20. Figure 6 View largeDownload slide MM, risk factors (percentages of total numbers of MM cases on y-axis) Note: the percentage is based on N = 20. Figures 4, 5 and 6 show that the social workers focus on the child, as well as the family, independently of the model. In all three models, three out of the five most used risk factors concern the child. For instance, 40–60 per cent of the social workers chose school problems as a weighted factor. School problems concern problems with regard to learning, social difficulties or truancy. The figures also show that five out of ten risk factors are consistent no matter which model the social workers are using. This means that, no matter whether a municipality chooses to use the ICS, the SoS or the MM, the social workers most often indicated that they weight the following five risk factors: (i) insufficient care, (ii) school problems, (iii) severe disharmony in the home, (iv) problems in regard to leisure time, friends and/or network and (v) extrovert problematic behaviour. This finding suggests that the model might not be regulatory for which risk factors the social workers weight in their assessments. As other researchers have argued, this finding indicates that the assessments still depend on aspects such as the social worker’s professional skills, intuition, formal knowledge, practice and emotional wisdom as well as ethics. These aspects are not to be ignored, as it is based on these aspects that social workers choose what information to collect (Munro, 2005, 2008). One could also argue that the reason for this finding is that the social workers might still be sceptical about applying such models, which is why the risk factors they actually assess do not differ depending on the model (Munro, 2008). Holistic assessments As the findings about risk and protective factors show, both the child and the family are the focus of the risk assessments. Having the child and the family in focus is very understandable. However, Danish law (da: Serviceloven) requires that the risk assessments have to be based on an overall consideration, which requires a holistic approach. In the questionnaire, the social workers were asked whether the model they have to apply contributes to a holistic approach. In general, across the three models, the social workers agreed that the models do contribute to a holistic approach—hence thirty-five out of forty-seven of the social workers answered either ‘agree’ or ‘strongly agree’ (see Table 2). Table 2 Distribution of responses between which model is used and the question: To what extent do you find that the model ensures a holistic approach? Strongly agree Agree Neither agree nor disagree Disagree Strongly disagree Total Model ICS Count 3 8 3 1 15 % within model 20 53 20 0 7 100 SoS Count 9 6 2 17 % within model 0 53 35 12 0 100 MM Count 8 7 15 % within model 53 47 0 0 0 100 Total Count 11 24 9 2 1 47 % within model 23 51 19 4 2 100 Strongly agree Agree Neither agree nor disagree Disagree Strongly disagree Total Model ICS Count 3 8 3 1 15 % within model 20 53 20 0 7 100 SoS Count 9 6 2 17 % within model 0 53 35 12 0 100 MM Count 8 7 15 % within model 53 47 0 0 0 100 Total Count 11 24 9 2 1 47 % within model 23 51 19 4 2 100 Note: the percentage is based on N = 47. In Table 2, it is interesting that eight out of seventeen social workers (47.1 per cent) using the SoS answered either ‘disagree’ or ‘neither agree nor disagree’ to the question and four out of fifteen (26.7 per cent) of the social workers using the ICS answered ‘strongly disagree’ or ‘neither agree nor disagree’. This finding indicates that, even though most of the social workers find that the model is helpful to ensure a holistic approach, twelve out of thirty-two (37.5 per cent) of the social workers who use the SoS or the ICS do not agree on whether the model contributes to a holistic approach. In addition, one social worker using the ICS and five social workers using the MM have not answered this question, which could indicate that the social workers find it difficult to assess whether the model contributes to a holistic approach. These different answers might illustrate that, in the worst case for some social workers, the model actually has the opposite effect of the intentions to ensure holism and a better structure of the assessment. This finding is especially interesting regarding the use of the ICS, as it is based on Bronfenbrenner’s ecological model and one could suggest that this would ensure a holistic approach. One interpretation of this finding is that there is a discrepancy between the theoretical background for the ICS and the practical design of the ICS, for instance in terms of the IT system. Earlier studies of use of the ICS in England have, for instance, clarified that the forms to be completed by social workers made it difficult to understand and sense the central story line of the family. As Shaw and Clayden write: This partialising tendency was one of the most common comments made from all sites and at all stages of the [ICS] evaluation. A feared consequence was that there would be a loss of the central story line, obscuring the family context, and resulting in a loss of narrative (Shaw and Clayden, 2009, p. 19). In the section on risk and protective factors above, I briefly presented a finding regarding the ICS which was shown in Figure 1. Figure 1 shows that it is mainly factors regarding the child which are weighted in the assessment when looking at protective factors. This leaves out almost all protective factors about the family, the family’s network and the social surroundings in assessing the risk of the child when using the ICS, which illustrates that the model does not ensure a holistic approach when looking at protective factors. In terms of the SoS, it is stated in the Danish guidelines that the model is based on an understanding of the assessment as a process which is characterised by being holistic, as the approach combines the perspectives of the family and the professionals (Turnell et al., 2013). However, understanding a holistic approach as a combination of the two sets of perspectives is different from Bronfenbrenner’s (1979) ecological approach. It is also very different from the holistic approach that historically was the backbone of social work in Denmark, as described by Hillgaard and Keiser in 1979. Hillgaard and Keiser argue for seeing the client in connection to his or her problems in a view of society linked to the past, the present and the future (Hillgaard and Keiser, 1979). If we look at how the holistic approach is described in the latest reform about children at risk, it states that the risk assessment has to identify problems as well as resources in regard to the child, the family and the family’s network (Servicestyrelsen, 2011). Based on these two Danish understandings of a holistic approach, one could argue that the SoS approach, according to Turnell and colleagues (2013), lacks a historical view of the problems and the resources, which is clear in Hillgaard and Keiser’s approach. The SoS model also lacks the requirements stated by the latest reform about assessing in regard to the child, the family and the network. This could be one of the reasons why none of the social workers using the SoS answer ‘strongly agree’ and eight social workers answer ‘disagree’ or ‘neither agree nor disagree’ with whether the SoS model ensures a holistic approach to the risk assessment. Table 2 shows that all social workers using the MM have answered either ‘agree’ or ‘strongly agree’. This means that none of the social workers using this model disagrees that the model ensures a holistic approach. This pattern of very satisfied social workers could indicate that these workers might be more involved in how the form they have to complete is designed. It could also indicate that the form is not as standardised and fixed as the forms used in the ICS and SoS models, which have been developed by external consultants who might not know about daily practice in a particular municipality. Concluding discussion In this paper, I have addressed the use of three types of risk-assessment models in Denmark. My findings clarify that it makes a difference which risk-assessment model a municipality chooses to base its risk assessments on. My study indicates that, even though the numbers of risk and protective factors have been uneven, social workers using the SoS ticked off twice as many protective factors than risk factors. This indicates that the municipalities using the SoS are much more protective-oriented in the risk assessment compared to municipalities using the ICS or the MM. Hence, the intention of the reform, that the risk assessment has to identify problems as well as resources concerning the child, within the family and the family’s network, are not fulfilled (Servicestyrelsen, 2011) when using these two models. When looking at the protective factors only, the municipalities who choose to use the ICS or the SoS do have a more child-focused approach compared with using the traditional MM. This is not the case when looking at the risk factors. Here there seems to be more consistency in which factors the social workers weight, independently of the assessment model. However, one could argue that, for social workers using the SoS, this finding is not as prominent, as these social workers mainly base their risk assessment on protective factors and not on risk factors, and therefore they would keep the focus on the child. For social workers using the ICS and the MM, the risk assessments mainly focus on school problems, problems in regard to the children’s leisure time, extrovert or introvert behaviour and insufficient care from the parents or severe disharmony in the home. These factors are distinct for the two models. Based on this finding, one could argue that the ICS and the MM models are equivalent. The questionnaire consisted of more factors concerning the child and the family, which could potentially lead to more focus on these factors. However, this study showed that the social workers refrained from focusing on factors linked to the society. Moreover, 37.5 per cent of the social workers using the ICS and the SoS did not agree on whether the model contributes to a holistic approach. Together, this indicates that these models tend to make it difficult to assess cases holistically. These findings contradict the intentions of implementing new models for risk assessments according to the reform and the Danish literature on the ICS (Servicestyrelsen, 2011; VIA University College et al., 2012). Previous studies have already shown that the complexity of the risk-assessment forms have made it difficult to see the whole story about the family and have instead broken the assessments into boxes (Shaw et al., 2009b). If we look instead at the way the literature on the SoS interprets holism, it is in contrast with the Danish legislation. It is stated in the SoS that a holistic approach is when the approach combines the perspectives of the families with the professionals (Turnell et al., 2013). One could argue that this is also happening in the application of the SoS, according to my findings. However, the social workers who use the SoS do not find the model helpful in regard to ensuring the holistic approach in the way they themselves understand holism and, in that case, the model is insufficient. These different findings lead to a discussion on whether new ways of doing risk assessments have actually led to model-specific risk assessments or whether the assessments still cover the same risk factors that social workers have always found relevant. It also leads to a discussion on whether the attempts of the Danish government to place the child in the centre of the assessment have actually come to fruition. This doubt also concerns the government’s focus on incorporating protective factors in the assessment, which is one of the key elements in the Child’s reform. One could argue that the new risk-assessment models affect other essential factors which are not included in this study. This could concern factors such as use of theoretical knowledge of children and families at risk or ensuring participation of children and families in the risk-assessment process, which both the ICS and the SoS describe as factors to be in focus. Both factors are absolutely relevant and necessary in the risk assessment of vulnerable children and families, and both factors have been noted as missing from the assessments and, for the same reason, they were mentioned in the Child’s reform as important issues (Den Digitale Taskforce, 2004; Servicestyrelsen, 2011). Further research should focus on these factors. The findings of this study also indicate that it takes a longer time to change or incorporate new ways of working with vulnerable children and families than anticipated. Previous research has found that the models are only half-heartedly applied by social workers, as the models conflict with their usual ways of doing risk assessments in the past (Munro, 2008). This is another argument for further research within the area of risk-assessment models. Acknowledgements I would like to thank the six municipalities and all the participating social workers for cooperating with me and for answering my survey. I would also like to thank my supervisors Karin Kildedal (Aalborg University), Marianne Skytte (Aalborg University) and Mette Lausten (SFI—the National Centre for Social Research) for guiding me in my research process. In the writing process, I visited the University of York, where Andrew Hill supervised my writing and Ian Shaw kindly lent me his office—I really appreciated this. I would also like to acknowledge my colleagues from the Department of Sociology and Social Work at Aalborg University, Copenhagen, for participating in discussions about my work. The research project is a part of my Ph.D., which is financed by Aalborg University. References Andersen S. H., Jensen B. ( 2016) Anbringelser—Historik, omfang, årsager og virkemidler [Placements—History, Extent, Reasons and Effects] , Copenhagen, Gyldendal og Rockwool Fondens Forskningsenhed. Baird C., Wagner D., Healy T., Johnson K. ( 1999) ‘Risk assessment in child protective services: Consensus and actuarial model reliability’, Child Welfare: Journal of Policy, Practice, and Program , 78( 6), pp. 723– 48. Bell M., Shaw I., Sinclair I., Sloper P., Rafferty J. ( 2007) The Integrated Children’s System: An Evaluation of the Practice, Process and Consequences of the ICS in Councils with Social Services Responsibilities , University of York, York, UK, Department of Social Policy and Social Work. Bronfenbrenner U. ( 1979) The Ecology of Human Development: Experiments by Nature and Design , Cambridge, MA, Harvard University Press. Camasso M. J., Jagannathan R. ( 2000) ‘Modeling the reliability and predictive validity of risk assessment in child protective services’, Children and Youth Services Review , 22( 11), pp. 873– 96. Google Scholar CrossRef Search ADS Cleaver H., Walker S., Scott J., Cleaver D., Rose W., Ward H., Pithouse A. ( 2008) The Integrated Children’s System: Enhancing Social Work and Inter-Agency Practice , London, Jessica Kingsley. Den Digitale Taskforce ( 2004) Afrapportering: Indsatsområdet ‘udsatte børn og unge’ [Reporting: Focus Area ‘Vulnerable Children and Youth’] , Copenhagen, Digital Forvaltning. Egelund T. ( 1997) Beskyttelse af barndommen: Socialforvaltningers risikovurdering og indgreb [The Protection of Childhood: Risk Assessment and Intervention of Child Protection Services] , Copenhagen, Hans Reitzels Forlag. Egelund T., Hestbæk A. ( 2003) Anbringelse af børn og unge uden for hjemmet: en forskningsoversigt [Placing Children Out of Home: A Review] , Copenhagen, Socialforskningsinstituttet. Egelund T., Sundell K. ( 2001) Til barnets bedste: Undersøgelser af børn og familier—En forskningsoversigt [For the Child’s Best: Examination of Children and Families—A Review] , Copenhagen, Hans Reitzels Forlag. English D. J., Pecora P. J. ( 1994) ‘Risk assessment as a practice method in child protective services’, Child Welfare: Journal of Policy, Practice, and Program , 73( 5), pp. 451– 73. Folketinget ( 2010–11) Vejledning om særlig støtte til børn og unge og deres familier nr. 3 [Guidance about Social Support to Children and Young People and Their Families No. 3], pkt. 277, Copenhagen. Hillgaard L., Keiser L. ( 1979) Social (be)handling: teori og metode i socialt arbejde [Social Treatment: Theory and Methods in Social Work] , Copenhagen, Munksgaard Socialpædagogiske bibliotek. Holmes L., McDermid S., Jones A., Ward H. ( 2009) How Social Workers Spend Their Time: An Analysis of the Key Issues that Impact on Practice Pre- and Post Implementation of the Integrated Children’s System , Loughborough, Centre for Children and Family Research, Loughborough University. Hood C., Rothstein H., Baldwin R. ( 2001) The Government of Risk: Understanding Risk Regulation Regimes , Oxford, Oxford University Press. Høybye-Mortensen M. ( 2011) ‘ Velfærdsstatens dørvogtere: Procesregulering af visitationer på socialområdet’ [‘The doorkeepers of the welfare state: Regulation of the process of visitations on the social area’] , Department of Political Science, University of Copenhagen. Kildedal K., Nordstoga S. ( 2011) ‘Rammer for § 50-undersøgelsen i Danmark og barnevernsundersøkelsen i Norge’ [‘Frame for the risk assessment in Denmark and the risk assessment in Norway’], in Kildedal K., Uggerhøj L., Nordstoga S., Sagatun S. (eds), Å bli undersøkt—norske og danske foreldres erfaringer med barnevernsundersøkelsen [To Be Examined—Norwegian and Danish Parents’ Experiences with Risk Assessments] , Oslo, Universitetsforlaget. Levin C. ( 1998) Uppfostringsanstalten: Om tvång i föräldrars ställe [Reformatories: Coercion In Loco Parentis] , Lund, Arkiv förlag. Munro E. ( 2005) ‘Improving practice: Child protection as a systems problem’, Children and Youth Services Review , 27, pp. 375– 91. Google Scholar CrossRef Search ADS Munro E. ( 2008) Effective Child Protection , London, SAGE. Munro E. ( 2010) ‘Learning to reduce risk in child protection’, British Journal of Social Work , 40( 4), pp. 1135– 51. Google Scholar CrossRef Search ADS Munro E., Taylor J. S., Bradbury-Jones C. ( 2013) ‘Understanding the causal pathways to child maltreatment: Implications for health and social care policy and practice’, Child Abuse Review , 23( 1), pp. 61– 74. Google Scholar CrossRef Search ADS Rak C. F., Patterson L. E. ( 1996) ‘Promoting resilience in at-risk children’, Journal of Counseling & Development , 74, pp. 368– 73. Google Scholar CrossRef Search ADS Rossi P. H., Schuerman J., Budde S. ( 1999) ‘Understanding decisions about child maltreatment’, Evaluation Review , 23( 6), pp. 579– 98. Google Scholar CrossRef Search ADS PubMed Rutter M. ( 1990) ‘Psychosocial resilience and protective mechanisms’, in Rolf J., Masten A. S., Cicchetti D., Nuechterlein K. H., Weintraub S. (eds), Risk and Protective Factors in the Development of Psychopathology , New York, USA and Melbourne, Australia, Cambridge University Press. Sagatun S. ( 2011) ‘Dimensjoner ved omsorg som gis oppmerksomhet i barnevernsundersøkelsen’ [‘Dimensions regarding care which are given attention in the risk assessment’], in Kildedal K., Uggerhøj L., Nordstoga S., Sagatun S. (eds), Å bli undersøkt—norske og danske foreldres erfaringer med barnevernsundersøkelsen [To Be Examined—Norwegian and Danish Parents’ Experiences with Risk Assessments] , Oslo, Universitetsforlaget. Servicestyrelsen ( 2011) Håndbog om Barnets Reform [Handbook about the Child’s Reform] , Odense, Servicestyrelsen. Shaw I., Clayden J. ( 2009) ‘Technology, evidence and professional practice: Reflections on the Integrated Children’s System’, Journal of Children’s Services , 4, pp. 15– 27. Google Scholar CrossRef Search ADS Shaw I., Bell M., Sinclair I., Sloper P., Mitchell W., Dyson P., Clayden J., Rafferty J. ( 2009a) ‘An exemplary scheme? An evaluation of the Integrated Children’s System’, British Journal of Social Work , 39( 4), pp. 613– 26. Google Scholar CrossRef Search ADS Shaw I., Morris K., Edwards A. ( 2009b) ‘Technology, social services and organizational innovation or how great expectations in London and Cardiff are dashed in Lowestoft and Cymtyrch’, Journal of Social Work Practice , 23, pp. 383– 400. Google Scholar CrossRef Search ADS Turnell A. ( 1999) Signs of Safety: A Solution and Safety Oriented Approach to Child Protection , New York and London, Norton. Turnell A., Vesterhauge-Petersen M., Vesterhauge-Petersen H. ( 2013) Signs of Safety: Arbejdet med udsatte børn og deres familier [Signs of Safety: Work with Children and Families at Risk] , Copenhagen, Hans Reitzels Forlag. VIA University College, Professionshøjskolen Metropol, University College Lillebælt and University College Syddanmark ( 2012) Barnets velfærd i centrum—ICS Håndbog [The Welfare of the Child at the Centre—ICS Handbook] , Odense, Socialstyrelsen. Vinnerljung B. ( 1996) Fosterbarn som vuxna [Foster Children as Adults] , Lund, Arkiv Förlag. Werner E. E., Smith R. S. ( 2001) Journeys from Childhood to Midlife: Risk, Resilience and Recovery , New York, Cornell University Press. © 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