Social service utilisation in relation to class setting - a longitudinal study among children with mild intellectual disability in Sweden

ABSTRACT

The aim of this study was to investigate if the type of class setting is related to the utilisation of disability-related services and child welfare services outside school over time among children with mild intellectual disability (ID). A quantitative study with a longitudinal and comparative design was carried out including data from archival records concerning service utilisation among 405 children. Children in special classes were more likely than children integrated into regular classes to utilise disability-related services. Integrated children who changed school setting from regular classes to special classes were more likely to begin to utilise such services and to increase the number of services utilised, compared to children who remained integrated. Professionals in social services and schools may need to improve their collaboration around families of children with mild ID when assessing needs and providing services. Special attention may need to be devoted to children integrated into regular classes.

KEYWORDS:

• Social service utilisation
• children
• mild intellectual disability
• class setting

Introduction

Children with an intellectual disability (ID) face deficits in intellectual and adaptive functioning (American Psychiatric Association 2013). Compared to their peers, they have a higher prevalence of several disorders (Allerton, Welch, and Emerson 2011), are more likely to face violence (Jones et al. 2012) and maltreatment (Lightfoot, Hill, and LaLiberte 2011), and to live in families facing financial discrepancies (Emerson et al. 2010). Among families of children with mild ID living in Sweden, financial distress (Huus et al. 2017; Olsson et al. 2015), a vulnerability to other social problems (Olsson et al. 2015) and a high occurrence of co-existing disorders among the children has been observed (Huus et al. 2017; Lindblad, Gillberg, and Fernell 2011). Altogether, this means that these children, together with their families, generally require various kinds of services to have their needs met (Douma, Dekker, and Koot 2006; Weiss and Lunsky 2010). In Sweden, disability-related services outside school can be provided by municipal social authorities; i.e. social services, with the aim of promoting independency and full participation in the life of the community for persons with extensive and enduring impairments, such as ID. Examples of such disability-related services are respite care, companion services and short stay away from the home (SFS 1993:387). Social services also provide child welfare services to protect and support children who are at risk for a negative development because of social problems (SFS 2001:453), such as being neglected, abused or exploited (SFS 1990:52). Examples of child welfare services are family therapy, contact family and out-of-home placements. The provision of services by these acts is preceded by an individual assessment of the child’s needs and aims to ensure children a healthy development. International studies indicate that children´s service utilisation is influenced by their type of school setting. For instance, Majnemer et al. (2014) and Parkes et al. (2004) observed that children with cerebral palsy (CP) in regular schools in Canada and Ireland, respectively, were less likely to utilise rehabilitation services than children in special schools. Majnemer et al. (2014) claim that the difference may be due to that children with CP in special schools have more severe deficits and thus have a higher need for rehabilitation services, whilst Parkes et al. (2004) argue that the difference depends upon that physiotherapy services are structured around special schools in Ireland and are thus more easily accessible for children who attend such schools. The latter harmonies with a Canadian study, reporting that parents of children with disabilities find it easier to get access to out-of-school services from health care authorities and social services if their child attends a special school, compared to a regular school (Tétreault et al. 2014). This study investigates how social service provision outside school to children with mild ID in Sweden is related to their school setting, focusing on whether they attend regular or special classes.

In Sweden, compulsory school is free of charge for all children (SFS 2010:800). Children who because of ID fail to achieve educational goals for mainstream compulsory school are entitled to be enrolled in compulsory school for pupils with ID and follow the related curriculum (Swedish National Agency for Education 2016). Children with mild ID are educated according to a reading-based curriculum, in contrast to children with more severe ID, who are educated in accordance with a life-skills based curriculum (Swedish National Agency for Education 2011). This study concerns the former. A child who follows the curriculum for the compulsory school for pupils with ID can be educated in a special class consisting exclusively of pupils with ID, usually located in regular schools (Göransson, Nilholm, and Karlsson 2011), or receive the education in a regular class. Such pupils are named ‘integrated pupils’ in the Swedish Education Act (SFS 2010:800) and require that the teacher educates and evaluates the pupils in the class based on two different curriculums. Integration is a multifaceted concept without a universal definition (Swedish Schools Inspectorate 2016); in the present study, integration means a physical placement. Of the 5889 pupils (in 282 municipalities) in Sweden who followed a reading-based curriculum in compulsory school for pupils with ID the school year 2014/15, 1198 were integrated into regular classes (Swedish National Agency for Education 2016); thus approximately 20% at the national level. However, local numbers vary between 0% and 100%, because of local organisational dissimilarities. Some municipalities integrate all children, while others only offer special classes. In about 35% of Sweden´s municipalities, no pupil is integrated. The differences are due to that the Swedish Education Act allows the municipalities to organise the education for pupils with ID in different ways. A child´s guardians may have requests about class placement, but do not have a statutory right to make such a decision (Swedish Schools Inspectorate 2016).

A decision about whether a child with ID who follows the curriculum for the compulsory school for pupils with ID shall be integrated or not should be based on the pupil´s needs and prerequisites (Swedish Schools Inspectorate 2016). However, national evaluations reveal that other reasons are of great importance for such a decision. Reports from school administrators disclose that integration of pupils with ID is mainly based on other reasons than a child´ s educational needs, such as the structure and the size of the municipality (integration is most common in rural and sparsely populated municipalities to enable schooling close to home and to avoid long trips to schools that offer special classes), municipal traditions and school administrators´ interpretations of the concept of integration. Integration is also used as an approach to get guardians to accept that their child will follow the curriculum for the compulsory school for pupils with ID (Swedish National Agency for Education 2002). School administrators in Swedish municipalities report that they pay attention to a pupil´s pedagogical and social needs and requests, but three quarters of the municipalities state that the most important factor when making a decision about integration of a child with ID is requests from guardians (Swedish Schools Inspectorate 2016).

Regardless of class setting, a child with mild ID can be expected to need disability-related services, since a diagnosis of ID requires that the limitations in adaptive functioning must be limiting to the extent that ongoing support is required for the child to manage daily life (American Psychiatric Association 2013). For children with ID service utilisation can be influenced by factors such as the child´s gender (Martin et al. 2013), age, the level of ID, co-existing problems (Chang et al. 2014), ethnicity (Dura-Vila and Hodes 2009), parent´s knowledge of available services, their perceptions of their child´s problems and needs (Douma, Dekker, and Koot 2006; Weiss and Lunsky 2010) and types of services utilised (Olsson 2016). Knowledge about how to gain access to services (Andersen 1995, 2008) and how services are organised may also influence the utilisation of services (Andersen, Davidson, and Baumeister 2014). If social service provision outside school to children with mild ID in Sweden is related to their class setting, has only been investigated in smaller studies. Among 84 children with mild ID living in Sweden, those who were integrated into regular classes were found to be considerably less likely to utilise disability-related services, compared to those children who were placed in special classes (Olsson et al. 2015, 2017). The fact that integration of pupils with ID in Sweden is largely based on factors other than children´s functional ability (Swedish National Agency for Education 2002; Swedish Schools Inspectorate 2016) indicates that adaptive functioning is not automatically higher among children integrated into regular classes, compared to children in special classes. That integrated children are less likely to utilise services aiming to promote independency and participation in the life of the community is therefore an observation that needs to be studied further.

The aim of this study was to investigate if the type of class setting is related to utilisation of disability-related services and child welfare services outside school over time among children with mild ID, based on following questions:

• Are children in special classes more likely to utilise; a) disability-related services and b) child welfare services, compared to children integrated into regular classes?

• Do children in special classes utilise a higher number of; a) disability-related service types and b) child welfare service types, compared to children integrated into regular classes?

• Do children who are integrated into regular classes become more likely to start utilising; a) disability-related services and b) child welfare services, when they change school setting from being integrated into regular classes to attending special classes, compared to children who stay integrated?

• Do children who change class setting from being integrated into regular classes to attending special classes, increase their number of utilised; a) disability-related service types and b) child welfare service types?

Methods

Study design

This is a quantitative study with a comparative, longitudinal design.

Sample

Children with mild ID who are enrolled in compulsory school for pupils with ID are educated according to a reading-based curriculum, in contrast to children with more severe ID, who are educated in accordance with a life-skills based curriculum (Swedish National Agency for Education 2011). The latter were excluded in this study. The childcare and education departments (primary school authorities) in all 13 municipalities in a county in the south of Sweden were asked to provide class lists of all pupils enrolled in compulsory school for pupils with ID who followed a reading-based curriculum in public schools in the school years 2009–2013. The county included around 4% of all pupils enrolled in compulsory school for pupils with ID who followed a reading-based curriculum the school year 2014/15. Twelve departments agreed to provide such lists. The social service departments providing disability-related services and the social service departments providing child welfare services in the twelve municipalities were asked to provide information about service utilisation. All twelve departments providing disability-related services and eleven departments providing child welfare services agreed to provide such information. The sample thus consists of all children (n = 405) enrolled in compulsory school for pupils with ID who followed a reading-based curriculum in public schools during the school years 2009–2013 in these twelve municipalities (Table 1). Children’s ages were 6–18 years. The number of inhabitants in the municipalities was 100,000 to >150,000 for one municipality, 20,000 to >40,000 inhabitants for four municipalities and 1 to >20,000 for seven municipalities. Sample characteristics are shown in Table 1.

Table 1. Sample characteristics (n = 405).

	2009	2010	2011	2012	2013
In special classes	n	n	n	n	n
Males	90	82	89	88	86
Females	66	57	44	44	48
Total	156	139	133	132	134
Integrated into regular classes
Males	54	41	38	35	40
Females	36	33	30	34	36
Total	90	74	68	69	76
Altogether	246	213	201	201	210
Number of children who finished school^a		55	46	49	46
Number of children who became enrolled in school^a		22	34	49	55
Mean age for special classes	12.8	13.0	12.8	12.7	12.8
Mean age for regular classes	12.5	12.4	12.7	12.9	12.4

^aCompulsory school for pupils with ID.

Because children were included only when enrolled in compulsory school for children with ID, each year some children were added, and some disappeared (Table 1). Of the total sample of 405 children, 31% were included for 1 year, 22% for 2 years, 17% for 3 years, 12% for 4 years and 18% for 5 years. During the years 2009–2013, a total of 28 children changed class setting from an integrated setting to special classes. Half of the children changed class setting when they began secondary school. The mean age at changing class setting was 13.3 years.

Data collection

Class lists provided by the childcare and education departments were submitted to professionals in the social service departments who collected data from existing records about the children’s service utilisation during 2009–2013. Services were operationalised as services that after an individual needs assessment were granted by a formal decision and valid to use during the actual year/s. Data concerned type of service and under which Act the service was granted.

Data analysis

Since age (Chang et al. 2014) (given linear representation) and gender (Martin et al. 2013) have previously been found to influence service utilisation among children with ID, they were included in the model. The municipality in which the children lived (dummy coding) was controlled for where possible, because differences between Swedish municipalities resulting from variations in needs assessments have been reported for both the provision of disability-related services (Lewin, Westin, and Lewin 2008) and of child welfare services (Forkby, Höjer, and Liljegren 2015). The children´s IQ scores were not included in the model since IQ scores are not related to everyday functioning among children with mild ID, and therefore inappropriate predictors of service needs (Arvidsson and Granlund 2018).

Several participants reappeared during several years, and repeated measurements data from the same individual might be correlated. A robust variance estimation for cluster-correlated data was executed for all analyses to minimise the risk of underestimating the standard errors. It specifies that the standard errors allow for intragroup correlation, relaxing the usual requirement that the observations are independent. That is, the observations are independent across groups (clusters), but not necessarily within groups (Williams 2000). The clusters in our case are the observations from the same person. In data with repeated observations on individuals, using cluster-correlated standard errors affects standard errors and variance–covariance matrix of the estimators but not the estimated coefficients (Williams 2000).

The software package STATA 12.0 was used to carry out all data analysis.

To analyse whether children in special classes are more likely to utilise services compared to children integrated into regular classes, binary logistic regressions were used. The analysis relied on repeated cross-sectional analyses. In the analysis of disability-related services, twelve municipalities and 405 children were included. Some children were included in the analyses more than one time. On average, each child contributed to 2.6 observations, yielding a total of 1071 observations (694 from children in special classes and 377 from children integrated into regular classes). In the analysis of child welfare services, eleven municipalities and 368 children were included in the analysis. The municipality in which the children lived was controlled for (See Table 3).

Table 2. Service utilisation 2009–2013 in relation to type of class setting (percent in brackets).

	2009	2010	2011	2012	2013
	n	n	n	n	n
Disability-related services	246 (100)	213 (100)	201 (100)	201 (100)	210 (100)
Service utilisers in special classes	70 (45)	72 (52)	71 (53)	67 (51)	75 (56)
Integrated service utilisers	21 (23)	20 (27)	16 (24)	18 (26)	19 (25)
Child welfare services	224 (91)	193 (91)	181 (90)	180 (90)	187 (89)
Service utilisers in special classes	14 (10)	14 (11)	12 (10)	15 (13)	15 (13)
Integrated service utilisers	7 (9)	10 (15)	11 (18)	12 (18)	9 (13)
Missing values^a	22 (9)	20 (9)	20 (10)	21 (10)	23 (11)

^a One municipality did not participate regarding child welfare services.

Table 3. Class setting in relation to utilisation of services.

	Disability-related services (yes/no)			Child welfare services (yes/no)
	No. of pupils 405		No. of observations 1071	No. of pupils 368		No. of observations 965
Independent variables	p	OR	95% CI	p	OR	95% CI
Special classes (ref. integrated)	0.000	7.12	3.80–13.32	0.833	0.93	0.45–1.89
Males (ref. females)	0.121	1.40	0.91–2.15	0.116	0.60	0.32–1.13
Age (linear)	0.065	1.08	0.99–1.18	0.282	0.94	0.83–1.05
The whole model	χ^2 (14) = 61.66, p< 0.0001			χ^2 (13) = 11.84, p= 0.540

Based on binary logistic regressions. Adjusted for municipality giving dummy coding.

To analyse whether children in special classes utilise a higher number of service types compared to children integrated into regular classes, ordinal logistic regressions were used. The odds ratios (ORs) from such a regression analysis correspond to the average ORs obtained in repetitive binary logistic regressions (O’Connell 2006). The analysis was based on repeated cross-sectional analyses. Only children who utilised services were included. In the analysis for disability-related service types, all twelve municipalities and 189 children were included. In the analyses of child welfare service types, eleven municipalities and 52 children were included in the analysis. Municipality was controlled for regarding disability-related service types, but not for child welfare services, because there were few municipalities in which children utilised more than one child welfare service type.

To analyse whether children who are integrated into regular classes become more likely to start utilising services when they change class setting from being integrated into regular classes to attending special classes, compared to children who stay integrated, a binary logistic regression was used. Data were analysed longitudinally. In the analysis of disability-related services, 177 children from ten municipalities were included. Two municipalities were excluded; in these children could not change class setting because one municipality offered only integration into regular classes and the other only special classes. Municipality was not controlled for, because there were few municipalities in which children changed class setting. Only children who initially did not utilise services were included in the analysis. Analysis for child welfare services could not be carried out because among those 21 integrated children who did not utilise any child welfare services and who changed class setting to special classes, no one began to utilise child welfare services.

To analyse whether children who change class setting from being integrated into regular classes to attending special classes increase their number of utilised service types, a binary logistic regression was used. Data were analysed longitudinally. In the analysis of disability-related services, 236 children from ten municipalities were included. Two municipalities were excluded for the same reason cited above. Municipality was not controlled for, because there were few municipalities in which children changed class setting. Analysis for child welfare services could not be carried out because among those 23 integrated children who changed class setting to attend special classes, no one increased the number of service types utilised.

Results

An overview of service utilisers is presented in Table 2.

Are children in special classes more likely to utilise services compared to children integrated into regular classes?

The model including children’s class setting, age, gender, and municipality was statistically significant for disability-related services. Children’s class setting was statistically significant (OR = 7.12). Children with mild ID in special classes had more than seven times greater odds of utilising disability-related services during 2009–2013 than children who were integrated into regular classes. For child welfare services, the model was not statistically significant. Children’s class setting did not predict the likelihood of utilising child welfare services. There was a non-significant trend (p = 0.065) for an increasing likelihood to utilise disability-related services with increasing age, but no association to child welfare services (Table 3).

Do children in special classes utilise a higher number of service types compared to children integrated into regular classes?

The model including children’s class setting, age, gender, and municipality was statistically significant for the number of disability-related service types. Children’s class setting was not significantly associated with the number of disability-related service types utilised. The only factor that contributed significantly to the model was children’s age (OR = 1.34). As children grew older, they utilised a higher number of service types. For child welfare service types, the model including children’s class setting, age and gender was not statistically significant. Class setting was not related to the number of utilised service types (Table 4).

Table 4. Class setting in relation to the number of utilised service types.

	Disability-related service types^2			Child welfare service types^2
	No. of pupils 189		No. of observations 449	No. of pupils 52		No. of observations 119
Independent variables	p	OR	95% CI	p	OR	95% CI
Special classes (ref. integrated)	0.738	0.85	0.32–2.25	0.259	0.51	0.16–1.63
Males (ref. females)	0.540	0.82	0.45–1.52	0.576	0.70	0.21–2.40
Age (linear)	0.000	1.34	1.16–1.54	0.825	1.03	0.81–1.30
The whole model	χ^2 (14) = 38.38, p< 0.001			χ^2 (3) = 2.32, p= 0.509

Based on ordinal logistic regressions. For disability-related services, we adjusted for municipality giving dummy coding.

Do children who are integrated into regular classes become more likely to start utilising services when they change class setting from being integrated into regular classes to attending special classes, compared to children who stay integrated?

The model including children’s class setting, age, and gender was statistically significant for disability-related services. Children with mild ID who were integrated into regular classes but changed to attending special classes had more than six times greater odds of beginning to utilise services than children who remained integrated. Furthermore, children in special classes who stayed in special classes had more than five times greater odds to start utilising services than children who remained integrated into regular classes. There was no significant association to age (Table 5). For child welfare services results cannot be presented (see Methods section).

Table 5. Change in class setting in relation to the likelihood to start utilising disability-related services (yes/no)^a.

	No. of pupils 177		No. of observations 344
Independent variables	p	OR	95% CI
Integrated children who changed to special classes^a	0.006	6.13	1.68–22.30
Children in special classes who stayed in special classes^a	0.000	5.81	2.46–13.71
Males (ref. females)	0.982	1.01	0.55–1.86
Age (linear)	0.443	1.05	0.93–1.18
The whole model	χ^2 (4) = 17.88, p< 0.01

^aChildren integrated into regular classes who stayed integrated were the reference category.

Do children who change class setting from being integrated into regular classes to attending special classes, increase the number of utilised service types, compared to children who stay integrated?

The model including children’s class setting, age, and gender was statistically significant for disability-related services. Children who were integrated into regular classes but changed to attending special classes had more than five times greater odds of increasing the number of utilised service types, in comparison to children who remained integrated. Furthermore, children in special classes who stayed in special classes had almost three times greater odds of an increased number of utilised service types, compared to children who remained integrated into regular classes. There was no significant association to age (Table 6). For child welfare services results cannot be presented (see Methods section)

Table 6. Change in class setting in relation to increase in the number of utilised disability-related service types (yes/no).

	No. of pupils 236		No. of observations 548
Independent variables	p	OR	95% CI
Integrated children who changed to special classes^a	0.002	5.35	1.90–15.08
Children in special classes who stayed in special classes^a	0.003	2.82	1.41–5.66
Males (ref. females)	0.652	1.10	0.72–1.70
Age (linear)	0.825	1.01	0.92–1.11
The whole model	χ^2 (4) = 12.30, p< 0.05

^aChildren integrated into regular classes who stayed integrated were the reference category.

Discussion

The aim of this study was to investigate the relation between the type of class setting and utilisation of disability-related services and child welfare services outside school over time among children with mild ID living in Sweden. The results prove that children in special classes had more than seven times greater odds of utilising disability-related services during 2009–2013, compared to children who were integrated into regular classes. Furthermore, integrated children who changed school setting to special classes had – compared to children who remained integrated – more than six times greater odds to start utilising disability-related services. How should this be interpreted? Since the aim of this study concerns relational patterns of service utilisation and the analyses were cross-sectional, it is not possible to confirm a causal relationship between class setting and service utilisation. However, that service utilisation is related to class setting is an imperative finding, knowing that ongoing support is required for a child with ID to manage daily life and that disability-related services can improve adaptive functioning (American Psychiatric Association 2013) and decrease parental stress (Hastings and Beck 2004; Robertson et al. 2011).

The differences in service utilisation between non-integrated and integrated children with mild ID found in the present may be due to that the latter have a better adaptive functioning and a lower need for services than the former. This is in line with Majnemer et al. (2014), claiming that children with CP attending special schools in Canada have more severe motor and intellectual deficits and a higher need for rehabilitation services, than children in regular schools. However, the present study consists of children with mild ID, suggesting that service needs differences due to level of ID are less substantial. Nevertheless, everyday functioning also differs between children with mild ID (Arvidsson and Granlund 2018). For instance, co-existing problems are likely to increase service needs (Chang et al. 2014; Douma, Dekker, and Koot 2006). There is a lack of knowledge of possible differences in functional abilities between integrated and non-integrated children in Sweden. Further research is needed. Nevertheless, that integration of children with ID is largely based on other factors than children´s adaptive functioning (Swedish National Agency for Education 2002; Swedish Schools Inspectorate 2016), implies that the origins of differences in service utilisation are not exclusively service needs. Knowledge about available services and how to gain access to them are factors that influence service utilisation (Andersen 1995, 2008). Children are frequently directed to services by adults, such as parents and teachers, who serve as gateway providers. Their assessment of a child´s needs and their knowledge about the service system is crucial for the child to receive necessary services (Stiffman, Pescosolido, and Cabassa 2004). Such knowledge might differ between parents of integrated children and non-integrated children with mild ID. For instance, parents of children in special classes have more opportunities to meet other parents of children with mild ID, with whom they may exchange disability-related experiences. They also have possibilities to meet social service professionals who provide information about available disability-related services at parental meetings in school. This is not possible for parents of children with mild ID who are integrated into regular classes, because of ethical matters. Thus, parents of children in regular classes may have less access to information about available services, which may result in underutilisation of services. In fact, insufficient knowledge about where to find services has been identified as a barrier for seeking services among parents of children with ID (Douma, Dekker, and Koot 2006; Weiss and Lunsky 2010), while knowledge obtained by support networks has been recognised as a facilitator in accessing needed formal services (Olsson and Roll-Pettersson 2012). Whether children with developmental delays in Sweden receive services is strongly dependent on actions taken by influential adults surrounding the child (Wilder 2015). Furthermore, a literature review reveals that parents with higher education level and socioeconomic status (SES) report more positive attitudes towards education of children with disabilities in ordinary classes (de Boer, Pijl, and Minnaert 2010). Such parents may have more knowledge about children´s health and how to support their child in disability-related concerns, resulting in a lower need for disability-related services. According to Trivette, Dunst, and Hamby (2010), parents who experience higher levels of self-efficacy report a lower need for additional services. Parents in families of children with mild ID in which the mother participate in paid work or have higher levels of education have reported lower levels of needs for certain kinds of disability-related services (e.g. counselling, contact with other parents of children with disabilities); however, not for community services (Huus et al. 2017). According to other studies, children of parents with lower levels of education utilise services to a lesser extent than children of parents with higher levels of education (Thomas et al. 2007; Weller, Minkovitz, and Anderson 2003). Parents with lower levels of education and SES may have less knowledge of available services and thus report a lower need for support (Douma, Dekker, and Koot 2006). In Sweden, children of parents with lower levels of education are overrepresented among pupils who receive special education support (Giota and Lundberg 2007), while children with an immigrant background are overrepresented in compulsory school for pupils with ID in large municipalities (Berhanu 2008). However, whether there are differences in levels of education, SES or ethnicity between parents of children with mild ID in Sweden who attend special classes versus children who are integrated into regular classes is not known. Further knowledge is required about such possible differences and in what way they might influence children´s service utilisation.

Like parents, teachers may facilitate or impede service utilisation. Their degree of knowledge about the service system and available services and perceptions of a child´s impairment may play a greater role for decisions about services than a child´s actual need for care (Stiffman, Pescosolido, and Cabassa 2004). Teachers who educate children with ID in special classes in Sweden must have a teaching certificate in special education specialising in ID, in contrast to teachers of regular classes (SFS 2011:326). Teachers of regular classes have reported that they lack enough knowledge about ID and its consequences (Swedish Schools Inspectorate 2016). Teachers of integrated children may therefore be less likely to identify special needs among children with mild ID and to direct them to disability-related services outside school, compared to teachers of special classes. This is in line with a study concerning children with physical impairments living in Sweden, revealing that teachers of regular classes have difficulty identifying these children’s needs for support and prioritise the collective needs of the class (Hemmingsson, Gustavsson, and Townsend 2007). Teachers of regular classes generally have larger classes than teachers of special classes, which might decrease their opportunities to identify children´s needs. According to Stiffman, Pescosolido, and Cabassa (2004) a gateway provider with a great caseload burden may be less likely to direct children to services (Stiffman, Pescosolido, and Cabassa 2004). For child welfare services, no relations were found between class setting and service utilisation. In Sweden, all teachers are obligated to immediately report to social services if they suspect that a child is being maltreated and may need child welfare services (SFS 2001:453), regardless of impairment or class setting. Therefore, there is no reason to believe that teachers of special classes are more familiar with reporting to social services that a child is facing social problems, compared to teachers of regular classes. Therefore, service utilisation is not expected to vary depending on such matters.

In sum, the organisation of a welfare system will inevitably influence utilisation of services (Andersen, Davidson, and Baumeister 2014). Services for children with disabilities may be more accessible for children in special school settings (Parkes et al. 2004; Tétreault et al. 2014). According to the present study, disability-related services outside the school system, aiming to promote independency and participation in the life of the community, seem to be more accessible for children with mild ID who attend special classes than for those who attend regular classes. Although the effectiveness of formal support in fulfiling needs among families of children with developmental disabilities living in Sweden is unclear (Wilder 2015), the results suggest that children with impairments are surrounded by a complex organisational interaction, in which parents are forced to navigate in relation to various logics of the social services and the school. The question is to what extent the organisation of education for pupils with mild ID is compatible with the rationality that govern social service needs assessment? Do participation interventions by the social services somewhat presuppose a child in a more exclusive school environment? To gain a deeper understanding of these premises more research is required regarding children's functional abilities and possible co-existing problems, parent, and family characteristics in relation to children´s class setting and utilisation of services. Further knowledge is also needed about contact channels between school authorities and social services.

A limitation of the study is that children´s adaptive functioning was not measured and that the study does not include socioeconomic variables of the families. Such information is not possible to obtain in a study based on archival data about service utilisation. A limitation of the use of archival records is that the researchers lack opportunities to control the original data collection and documentation procedures. However, Swedish legislation requires that all services that are granted by a formal decision after an individual needs assessment are documented on an individual basis.

Implications for practice

Professionals in social services and schools may need to improve their collaboration around families of children with mild ID when assessing needs and providing services, with special attention to integrated children. Social service professionals need to improve engagement in outreach activities so that families of children with mild ID, independent of the child’s class setting, are informed about available services and how to gain access to them.

Acknowledgments

We gratefully acknowledge the Foundation Sävstaholm and the Foundation Sunnderdahl’s Disability Fund as financiers of this study. We would also like to thank the social service professionals and school professionals who assisted us in carrying out this study. No conflicts of interest exist for this study.

Disclosure statement

No potential conflict of interest was reported by the authors.