Social work in an assembly line? The development of specialisation in child welfare and further internal division of work between 2003 and 2018

ABSTRACT Organisational specialisation characterises present social work in Sweden as internationally. Specialisation is often expected to lead to higher professional expertise. This study provides an overview of the development of specialisation and workforce characteristics in child welfare based on survey data collected in 2003, 2014 and 2018. The aim was to describe the development of specialisation and discuss the conditions for professional expertise in child welfare work. The results indicate that child welfare work is increasingly work task specialised. Social workers reported performing fewer tasks in 2014 and 2018 compared to 2003. This could be understood mainly as a form of statutory specialisation, indicating a shift in the professional role. Also, with the tendency to divide the work into only one or two phases of the process, child welfare work today almost resembles work at an assembly line. Compared with 2003, the statutory child welfare social workers surveyed in 2014 and 2018 were also significantly younger and less experienced. Since decision-making in child protection is a highly complex and challenging task, a question is if the development may be a survival strategy of the organisations rather than an opportunity for developing expertise.


Introduction
In Sweden, like in other parts of the Western world, the evolving landscape of welfare systems has affected social workers' professional roles and job responsibilities.Both empirical evidence (Arches, 1991) and theoretical analysis (Mintzberg, 1979) have long suggested that highly bureaucratic organisations hinder workers from holistically approaching their tasks.The increase in administrative burdens (Brante et al., 2015) and standardisation (Ponnert & Svensson, 2016) has more recently been recognised as negatively impacting the professional autonomy of social workers.
Swedish child welfare work is recognised as family service-oriented, in contrast to child protection systems found in the USA, Canada and the UK (Lundström et al., 2021).However, child welfare in Sweden has undergone major changes in recent decades (National Board of Health and Welfare, 2015).A trend towards expanding the protective aspects has been identified (Lundström et al., 2021), and children's rights perspectives have likewise been strengthened (e.g. the United Nations (UN) Convention on the Rights of the Child has become law).The requirements for documentation have increased to strengthen legal certainty; standardised documents and process directives have been implemented as part of the Swedish version of the Integrated Children's System, Barns behov i centrum (which translates to 'Children's needs in focus') (Skillmark & Oscarsson, 2020).In addition, to perform exercise of authority in Child Welfare, a Swedish degree in social work or a comparable degree is required (National Board of Health and Welfare, 2015).
Furthermore, social work in Sweden, as internationally, is increasingly characterised by organisational specialisation (Bergmark & Lundström, 2007;Grell et al., 2016Grell et al., , 2021;;Lundgren et al., 2009).Much like the reorganisation of local authority personal social service following the Seebohm report in England (Parton, 2009), the idea of an integrated, community-based social service was introduced and implemented in Sweden in the 1970s and early 1980s (Sjöberg & Turunen, 2022).However, as in England, integrated social services were a short-lived phenomenon.Social services in Sweden today are, once again, divided into separate divisions responsible for specific groups, such as service to the elderly, child welfare, social assistance, and adult services.This form of organisational specialisation is often combined with a division between groups that handle tasks involving exercise of public authority (statutory social work) and groups that handle only specialised service or treatment tasks (e.g.family treatment) (Bergmark & Lundström, 2005).In child welfare, further internal specialisation based on the children's age was practised in approximately one-fifth of Sweden's mid-sized municipalities investigated in 2001 (Bergmark & Lundström, 2007).
The definitions and operationalisation of specialisation differ in previous research, making studies hard to compare (Lundgren et al., 2009), but both advantages and disadvantages of specialisation have been described (Grell et al., 2021).LeBlanc et al. (2007) found no clear evidence that specialisation based on work tasks in child welfare serves families better, but indications of a more prominent child perspective (Östberg, 2010) have been found in work task-specialised child welfare organisations.However, previous research also indicates worse conditions for making appropriate assessments in specialised organisations (Bergmark & Lundström, 2005;Perlinski et al., 2012) and reports that clients usually prefer to meet with only one or a few social workers (Blom et al., 2009;Boklund Palm, 1995;LeBlanc et al., 2007), which can be hard to achieve in any specialised organisation.
Perhaps the most prominent argument for specialisation is the possibility for workers to develop expertise in a specific area (Bergmark & Lundström, 2007;Grell et al., 2021;Perlinski et al., 2011).In complex, often conflictual and ethically challenging statutory child welfare work, and particularly tertiary child protection work, it has been described that specific knowledge, skills and a 'values framework' are necessaryrequiring specialisation (Healy & Meagher, 2007, p. 333).In a study in which politicians and managers were interviewed about positive and negative aspects of the organisational model, one argument for specialised organisation was that by handling fewer types of tasks, the social workers would become more skilled in, and less ambivalent about, their work (Perlinski et al., 2011).Other studies indicate such advantages of specialisation in specific categories of social problems or in target groups of clients (Astvik, 2003;Gómez-García et al., 2020;Grell et al., 2016Grell et al., , 2021;;van der Tier et al., 2021) although none of these studies has explicitly examined specialisation in child welfare.
If specialisation is expected to lead to increased professional expertise, what are social workers expected to be experts in?Previous studies based partly on the same material used in this study have raised important questions about the development of specialisation in child welfare (Tham, 2018).Between 2003 and 2014, no increase in specialisation based on age groups was found, but about half of the investigated municipalities and town districts had divided the investigative phase so that some work groups handled intake and others conducted investigations.Although social workers who handled only intake in 2014 were excluded from further analysis for comparative reasons, the content of the work of the included social workers had changed considerably.In 2003 they performed various tasks, including advice and support, treatment, and preventive work.By 2014, most social workers mainly reported conducting investigations, suggesting a specialisation in statutory work.The many changes in child welfare in Sweden described above make it an interesting case to study.Child welfare is the largest section of statutory social services in Sweden and many other European countries (e.g.re Ireland, see Burns, 2011).Since specialisation is assumed to lead to expertise, but child welfare is at the same time described as becoming an entry-level job, the present study provides an overview of the trend of specialisation as well as workforce characteristics of the child welfare workers assumed to be(come) specialists.The study aims to describe the development of specialisation and discuss the conditions for professional expertise in child welfare.Research questions are: (1) To what extent are social workers specialised in only child welfare cases, and how can further specialisation in child welfare be understood?; and (2) What are the characteristics of those social workers expected to be specialists?
The context and the theoretical points of departure Child welfare units in Sweden mainly receive cases through reports of concerns from professionals with mandatory reporting status and through reports of concern and applications from civilians.
Theoretically, the work in child welfare could be seen as a continuum, where decisions must be made at key points of the process (Baumann et al., 2011).
The continuum runs through the stages of service involved in cases processed by child welfare, starting with the first assessment and decision (to start an investigation or not), continuing via the investigative phase, supposed to result in a decision to initiate intervention or not, and ending with the decision to close a case (when the situation in the family is deemed to be good enough for the child).These three stages correspond to the work tasks municipalities in Sweden are legally obligated to perform in child welfare work: (1) assessment and decision about whether an investigation should be initiated (intake); (2) investigation and decision of the need for interventions or other measures; and (3) follow-up of decided interventions.There are legally binding time limits and guidelines for each task.However, the legislator also stresses the importance of preventive and outreach work in child welfare (National Board of Health and Welfare, 2015).Figure 1, inspired by the Decision-Making Ecology approach described by Baumann et al. (2011), shows the typical work task continuum for a social worker in statutory child welfare in Sweden for every stage.As can be seen, an assessment followed by a decision leads the process forward.
Further, according to the Decision-Making Ecology approach, the actual and perceived consequences of a decision, for the decision maker, as well as for the agency and the client (the outcomes of a decision), constitute the experience base underlying new decisions.This is also elaborated in a theory of professional expertise in social work by Fook et al. (2000).Although this theory differentiates between experienced and expert practice, it does relate becoming an expert to experience.About 2 years of practice seems enough to consolidate skills and confidence.Having around 3 years of practice makes a noticeable change in the ability to respond to practice vignettes in a prioritised and specific way (Fook et al., 2000).
The concept of specialisation Mintzberg (1979, p. 69) defines horizontal job specialisation as 'the division of labor according to tasks.' Vertical job specialisation can be understood as the degree of autonomy a worker possesses.

EUROPEAN JOURNAL OF SOCIAL WORK
In this study, we examine horizontal specialisation.As the concept of specialisation differs between previous studies, we will outline our understanding of different types of horizontal specialisations as follows: . There are differences in the juridical aspects of the cases handled by the family and personal social services in Sweden, with specific legislation enabling coercive measures to be taken in child welfare and adult services aimed at persons with substance abuse.However, most often, contact with adult services is voluntary.As in most countries, that is not the case with child welfare, where being subject to an investigation is mostly involuntary for the clients.Therefore, the division into social assistance, adult services, and specialised child welfare departments can be understood as a division of cases with different logic: case-type specialisation. .In child welfare, further internal specialisation is often based on a division between working with families with children (approx.0-12 years old) and working with families with youth (approx.13-21 years old), which can be understood as (within case-type) target group specialisation.(Tham, 2018) Since this can be understood as dividing the exercise of public authority into smaller parts, we also introduce the term phase specialisation for this kind of division of work.

Research design
This study is based on survey data collected on three occasions (2003, 2014, and 2018).Data were collected at the same workplaces in half of the municipalities of Stockholm County in all 3 years.One additional workplace in Stockholm City was added in 2014 and was also included in 2018.
The questionnaire was based on the General Nordic Questionnaire for Psychological and Social Factors at Work (OPS Nordic) (Dallner et al., 2000).Questions more specific to child welfare work were added, such as questions about what kind of cases and work tasks the respondents worked with.This study is primarily based on the responses to these questions.Before data collection in 2003, the survey was tested on active social workers and modified by their feedback.The questionnaire was found to be comprehensive and clear (Tham, 2008).
Data collection for all 3 years followed the same protocol.Workplace visits were conducted during regular staff meetings (by author P.T in 2003 and 2014 and author K.S. in 2018).Respondents were given information about the study and asked to answer the questionnaire privately.Questionnaires and addressed envelopes were provided for the social workers not participating in the meetings.A summary of the total sample and this study's analytical sample is provided in Table 1.

Ethical considerations
This study follows ethical guidelines and national laws (SFS, 2003;Swedish Research Council, 2017).The study includes no sensitive personal data.Permission to conduct the research was received from the heads of the divisions in all included town districts and municipalities for all three measuring occasions.Written information about the study was distributed.The data collection method made it possible to orally inform the respondents and answer questions about the study.The respondents were assured that the study results would be presented so that individuals or working groups would not be identifiable.Submission of the completed questionnaire was deemed to constitute informed consent.

Differences between the measuring occasions
The questionnaire was revised before data collection in 2014 and 2018, and a few questions were added.Here we only report the question with relevance to the present study.In 2003, the population was defined as social workers in Stockholm County handling reports of concerns or applications and conducting investigations of child welfare cases (Tham, 2007).When contacting the organisations before the second data collection, nearly half of the municipalities and town districts had set up specialised groups whose main task was handling reports of concern and deciding whether those should result in investigations (Tham, 2018) but did not conduct the investigations.Therefore, the specific work task termed 'intake' (i.e.handling reports of concern and administering applications) was added to the questionnaire before data collection in 2014, and those groups were surveyed.The population was consequently re-defined as social workers handling intake and/or investigations of child welfare cases.
At the time of data collection in 2018, most municipalities and town districts had special workgroups for handling cases concerning unaccompanied refugee children or assigned those cases to one or a few social workers in the larger workgroup.Moreover, one municipality had started separate work groups primarily for handling follow-ups, which, however, did conduct investigations if new concerns emerged during the ongoing service provision.Those groups were all surveyed.
Another difference between the measuring occasions was that the social workers more often were absent during the data collection visits in 2014 and 2018.Their supervisors often explained this as they had to prioritise emergencies in client work.

Study sample
The total sample consists of 1,351 social workers (2003 n = 321, 2014 n = 435, 2018 n = 595).This study's sample consists of all social workers reporting handling intake and/or conducting investigations of child welfare cases except those also reporting being managers, and social workers specialised in handling cases concerning unaccompanied refugee children.Even though the latter could be seen as another type of specialisation, we regard these groups as a consequence of the extraordinary circumstances that prevailed around 2015 due to the war in Syria, when Sweden received many refugees and many unaccompanied children.Most of the municipalities and town districts were about to dismantle these groups around the time of the third data collection.To enable comparison with other measuring occasions and other studies, these work groups and social workers who only reported handling cases regarding unaccompanied refugee children have been excluded from this study's sample.The study sample consists of 1,017 social workers (2003 n = 293, 2014 n = 337, 2018 n = 387).

Analysis
All analyses were conducted with SPSS Statistics 27.Differences between years for all variables were tested with non-parametric tests (Kruskal-Wallis for means and Chi-square test for frequencies).

Results
Table 3 shows the sample characteristics and different aspects of specialisation at the three measuring points.Statistically significant differences between the measuring points are indicated.At all three measuring points, around 90% of social workers were female and as many had at least a Bachelor of Science in Social Work.In 2018 the social workers were younger and had less work experience than the respondents in 2014 and 2003.Twice as many social workers were under 30 years old in 2018 compared with 2003.Consequently, fewer participating social workers were older; for example 2018, there were significantly lower proportions of social workers in the 41-50 and 51-60-year age categories.The changes in work experience as a statutory social worker were considerable; in 2018, more than 50% of the social workers had been working 3 years or less in social services, compared with 39% in 2003 and 2014.Phases were divided based on the statutory work tasks 'intake', 'investigation', and 'follow-up'.However, respondents could also be working with non-statutory work tasks.f Respondents have not marked any of the non-statutory work tasks.*p < 0.05; **p < 0.01; ***p < 0.001.

Specialisation
The results indicate that the move towards specialisation is persisting.To a large extent, work was case-type specialised at all three measuring points, meaning that almost every social worker who handled child welfare cases worked only with child welfare cases.However, even though the proportion of social workers who reported working in case-type specialisation was high already in 2003 (91%), there was a significantly larger proportion of social workers in 2014 (99%) and 2018 (98%) who worked only with this case type.
The proportion of respondents who worked only with child cases or youth cases, i.e. those reporting target group specialisation, had decreased in the later years even if it still was a common type of specialisation in 2018.From our data, however, the primary trend seems to be that work task specialisation, i.e. being responsible for a smaller part of child welfare work, has increased substantially.
This finding was further examined by dividing the work tasks into statutory and non-statutory work tasks.In 2014 and 2018, 71% and 78% of respondents reported performing only statutory work tasks, compared with 13% in 2003.Furthermore, the social workers in this study reported in 2014 and 2018 that they were working with only one (23% and 35%, respectively) or two (67% and 56%, respectively) out of three phases of the continuum (phase specialisation).This is further implicated because at least one municipality had started separate groups for handling follow-ups.These groups were included since they also conducted investigations actualised within an ongoing service provision.The decreasing trend of social workers handling follow-ups in this study implies that also other municipalities have set up separate groups that perform this legally obliged task.
Phase specialisation usually only included statutory work tasks.Unfortunately, the data do not allow complete comparisons with conditions in 2003.However, in 2014 and 2018, the tendency is clear: At the last measuring point, as many as 70% of the social workers reported working with only one (26%) or two (44%) of the phases and only performing statutory work tasks.

Discussion
This study aimed to describe the development of specialisation and discuss the conditions for professional expertise in child welfare work.
Comparing the 2003 and 2018 responses, the statutory child welfare social workers in the latter sample were younger and more inexperienced.They reported working with more work task specialisation but less target group specialisation.The trend of work task specialisation could be understood mainly as a form of statutory specialisation, which strengthens previous findings of a shift in the role of social workers (Tham, 2018).This means that prevention, outreach, treatment services, and giving advice and supportusually seen as classic social work tasksare not performed unless handled by other work groups specialised in these tasks.This can be understood as a more pronounced development of the social service, from family-centred child welfare to childcentred child protection (see Lundström et al., 2021, for a similar discussion).
In addition, specialisation also seems to be taking place inside the statutory phase of decisionmaking, phase specialisation, i.e. with different social workers handling different statutory phases or decisions.
What impact might this development have on clients and social workers?Firstly, suppose we assume that a report of concern is handled by one social worker who decides that an investigation is needed and hands the case to an investigation specialist.The client has already met at least two social workers in that case.If then the investigation leads to ongoing service or intervention, the client may meet yet another (a third) social worker who is responsible for the follow-up, as well as another (a fourth) who performs the actual service (e.g.family treatment).Since clients have been reported to prefer to have contact with only one social worker (Blom et al., 2009;Boklund Palm, 1995;LeBlanc et al., 2007), this way of organising child welfare may be considered problematic from the client's perspective.
Secondly, social workers responsible for a single phase of the process may lack information about the outcomes of their efforts.Paradoxically, being a 'one-decision specialist' could diminish decisionmaking knowledge and hinder expertise development over time, as decision-making knowledge relies on perceived or actual outcomes from decisions made in the past (Baumann et al., 2011).Social workers might be less effective in assessing situations holistically, and risk becoming detached from the real-life consequences of their decisions when only involved in specific phases, potentially compromising the quality of their assessments.Moreover, they must rely heavily on the information they receive from colleagues (Gevaert et al., 2022(Gevaert et al., , 2023)).This aligns with the discussion of Blom (2004) and Bergmark and Lundström (2005), who suggest that specialised organisations can encounter difficulties in making accurate assessments.The relatively short experience of most social workers (less than 3 years) and an average workgroup duration of 1.9 years accentuates this concern, particularly when 3 years of practice is known to significantly improve social workers' abilities (Fook et al., 2000).
The development of specialisation found in this study might rather be understood as a simplification of work (i.e. with respondents in 2014 and, even more pronounced, in 2018 having fewer work tasks).According to theory, when a job is highly specialised in the horizontal dimension, the worker's perspective narrows.Consequently, control over the work often shifts to a manager who coordinates the tasks through direct supervision or to an analyst who can achieve this through standardisation (Mintzberg, 1979).Thus, jobs often need vertical specialisation because they are horizontally specialised.In other words, having to perform a narrow task reduces the worker's perspective and, consequently, their control over it.Therefore, when tasks and knowledge areas are narrowly defined, social workers have less potential to control the services they provide clients (Arches, 1991).Less professional autonomy has been found to influence social workers' intentions to leave their positions (Kim & Kao, 2014;Kim & Stoner, 2008;Zheng et al., 2021), although some studies have produced conflicting results (Westbrook et al., 2012).Since previous research shows a persistently high turnover rate among social workers (Astvik et al., 2020;Itzick & Kagan, 2017;McFadden et al., 2015;Tham, 2007;Tham & Kåreholt, 2022), further research should examine whether the increasing work task specialisation is associated with reduced professional autonomy for social workers.
As the statutory work tasks are the only ones that cannot be de-prioritised, a question is whether simplifying work may be a survival strategy to prevent high turnover and burnout.However, the limited existing evidence suggests this positive effect only for target group specialisation (Gómez-García et al., 2020).The decrease in target group specialisation and the increase in work task specialisation indicate that the changes were made to meet the organisation's needs (e.g. the need to meet legally binding time limits) rather than the needs of the clients and social workers.

Limitations
The possibility of generalising the results of this study to a larger population is limited, as all respondents worked in the County of Stockholm.The data used in the present study were collected with the primary aim of measuring work conditions and therefore includes limited information about specialisation.The population used when collecting data has also likely excluded social workers who could have contributed important information to this study, such as those working only with follow-ups of decided interventions.
Furthermore, the data are individual-level data, and the results should be understood as presenting what social workers perceive as being included in their work tasks at each measuring occasion.Two important limitations are connected with this: firstly, as mentioned earlier, in 2003, the survey did not distinguish intake and investigation tasks.Before data collection in 2003, the survey was tested with active social workers, and none of them emphasised the need to regard intake as a distinct task.This suggests that intake was not perceived as separate from the investigation, meaning all social workers conducting investigations in 2003 also handled reports of concern.However, rather than assuming this, we took a cautious approach and refrained from making that assumption in our analysis.As a result, respondents who answered the survey in 2003 had one less possible task for the variable 'work task specialisation.'Despite this, the number of reported work tasks significantly decreased in 2014 and 2018.Assuming that some or all of the social workers in 2003 did indeed perform the work task intake would only strengthen these findings.
Secondly, organisational or other changes might affect how social workers define their tasks.For example, more social workers who answered the survey in 2014 and 2018 might support and advise clients while conducting investigations but did not define 'support and advise' as a separate work task when answering the survey.Data collection methods other than surveys may better capture information on what social workers actually do and how the work is organised.

Concluding reflections
The present study confirms previous studies' findings of specialisation being a persistent trend in child welfare (Bergmark & Lundström, 2007;Grell et al., 2021;Lundgren et al., 2009).More interestingly, this study shows that the workforce has, at the same time, become increasingly younger and less experienced.Despite this finding and the results of many other studies, it is still unclear why specialisation continues to increase and whether the two trends have anything to do with each other.Is specialisation a way to manage the complexity of statutory child welfare work by dividing the work into ever smaller parts rather than being an opportunity to develop professional expertise?The tendency to specialise in only one or two phases of the process almost makes the work resemble an assembly line.This may mean that conditions for developing professional expertise are deteriorating since social workers are not around to see the consequences of their decisions.
working conditions have lately received much media attention with articles in approximately 50 newspapers and journals, interviews in radio and TV.He has been a member of the research council at the School of Health and Welfare at Jönköping University and is presently chairman of the Research Council Review Board at the School of Health and Welfare.

Figure 1 .
Figure 1.Continuum in the municipal child welfare services in Sweden, with examples of work tasks.Note: This figure is influenced by the figure 'Flow of Clients through the Decision-Making Continuum' by Baumann et al. (2011, p. 6).

a
Works only with child welfare cases (child and/or youth cases).b Within case-type specialisation: works only with child or youth cases.c Number of work cases marked by the respondent.Note that in 2003, the task 'intake' was not an alternative in the questionnaire.Hence, there were six possible work tasks to choose from in 2003 and seven possible work tasks in 2014 and 2018.d Comparison between 2014 and 2018.e . Work task specialisation is broadly understood as any form of division of work based on work tasks, where performing fewer work tasks is understood as being more specialised. .That statutory social workers perform services or supportive tasks (e.g.advice and support) within the framework of statutory work and parallel with the exercise of authority is an implied part of statutory work in Sweden (National Board of Health and Welfare, 2006).Only performing work tasks involving exercise of public authority (assessments and decisions) is what is understood as statutory specialisation. .Previous studies on parts of the data included in this study have shown specialisation in separate work groups for intake and investigation

Table 1 .
Sample by year.MeasuresCoding and definitions of each variable are provided in Table2.Case-type specialisation is operationalised as respondents handling only child and/or youth cases (i.e.not working with other areas of social work such as social assistance).Target group specialisation is operationalised as respondents handling only child or only youth cases.Work task specialisation is operationalised as the number of specified work tasks a respondent reports conducting.The different work tasks are divided into statutory and non-statutory work tasks.Phase specialisation is operationalised as the number of statutory work tasks (intake, investigation, and follow-up) a respondent reports conducting, even if the same respondent also reports performing one of the nonstatutory work tasks (preventive, outreach, advice and support or treatment).Statutory specialisation means that respondents report working only with one or more of the three statutory work tasks.
How many years have you worked in your present work group?

Table 3 .
Mean differences between years, individual characteristics, and specialisation.Differences were tested with non-parametric tests (Kruskal-Wallis for means and Chi-square test for frequencies).Comparisons were made between 2003 and 2014 and 2018, respectively, except when 2003 data were missing, in which case, comparisons were only made between 2014 and 2018.