Erasmus MC heeft ervoor gezorgd dat je Mijn BSL eenvoudig en snel kunt raadplegen. Je kunt je links eenvoudig registreren. Met deze gegevens kun je thuis, of waar ook ter wereld toegang krijgen tot Mijn BSL.
Om ook buiten de locaties van Erasmus MC, thuis bijvoorbeeld, van Mijn BSL gebruik te kunnen maken, moet je jezelf eenmalig registreren. Dit kan alleen vanaf een computer op een van de locaties van Erasmus MC.
Eenmaal geregistreerd kun je thuis of waar ook ter wereld onbeperkt toegang krijgen tot Mijn BSL.
Login
Als u al geregistreerd bent, hoeft u alleen maar in te loggen om onbeperkt toegang te krijgen tot Mijn BSL.
Adolescents in aftercare services and transitioning from out-of-home care, also known as care leavers, are more prone to social exclusion, risk behaviors, mental health problems, and lower quality of life compared to adolescents without a history in out-of-home care. These factors have a detrimental effect on their coping. This study was a cross-sectional document analysis which examined Finnish care leavers’ coping profiles and identified health related, educational, and behavioral background factors as well as their combinations predicting those profiles by utilizing Bayesian method. The data comprised information on care leavers that were aftercare services’ clients in one large Finnish city in the Fall 2020 (N = 698). Care leavers were divided into three groups based on the severity of factors affecting their coping: a minority belonged to the “moving on” group (6.7%), the majority (73.4%) to the “survivors” group, and a fifth (19.9%) to the – most problematic – “strugglers” group. Overall, 16 independent health related, educational, and behavioral factors were found to be associated with coping profiles. The most strongly associated variables were alcohol and drug use in personal history as well as during aftercare, and impulsivity. In addition, having access to a student’s social benefits had a protective effect on care leavers’ coping. Therefore, coping is a multifaceted phenomenon, which emphasizes the importance of service development and multiprofessional collaboration. In addition, the factors affecting the coping profile are also interconnected, which emphasizes the significance of further research, including intervention studies, to increase the knowledge of the phenomenon.
Opmerkingen
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Out-of-home care (OOHC) is a general term for temporary, medium-, or long-term placement when a child or adolescent is unable to live with their own family. OOHC includes placements in professional and non-professional settings such as foster homes, kinship care, professional foster homes, child welfare institutions, and reform schools. (Mendes & Snow, 2016) According to the latest global estimates, more than 3.2 million children or adolescents are living in institutional settings on an annual basis (Desmond et al., 2020). Placements in other settings than institutions (e.g., foster home and kinship care) are not included in that estimate, an omission which means that the total number of children and adolescents in OOHC can confidently be assumed to be considerably higher. In 2023, 17,299 children were placed OOHC in Finland, of which 3840 were placed for the first time in their lives (Finnish Institute for Health and Welfare [THL], 2024a). A placed child may have several different placements during the year. For example, the child may first have been urgently placed in a reception home, and later in the same year it has been decided to take the child into care, and a place has been found for them, for example in a foster family. In 2023, of all the Finnish children registered according to the last place of placement, 42 percent were in family care, 10.5 percent in professional family care, 49 percent in institutions, and the remaining 4 percent in other care. (Finnish Institute for Health and Welfare [THL], 2024a)
Adolescents in aftercare services or transitioning from OOHC (in this study, they are referred to as care leavers) form a vulnerable group (Taylor et al., 2024). The transition from OOHC to adulthood is known to be a critical time in care leavers’ lives because they often do not have the skills to survive independently, to which aftercare services are designed to respond. The criteria, content, and duration of aftercare services vary between countries – mainly due to legislation – however, the basic purpose of aftercare services is universal: to systematically support adolescents in their transitioning phase from OOHC to adulthood. (Tyler et al., 2017) It is worth noting that in some countries, aftercare services respond to extended OOHC placement, or after OOHC placement ends, care leavers receive services easing their transition (Taylor et al., 2024). According to Finnish legislation (Child Welfare Act, 2007), a child or adolescent is entitled to aftercare services when they have been placed in open care for at least half a year or through placement in OOHC. In Finland, the purpose of aftercare is to support an adolescents’ independence by helping them to achieve sufficient skills for independent everyday life and adulthood. In the case of an adolescent moving on their own, the focus is on strengthening sufficient capabilities to manage independently in everyday life. (Finnish Institute for Health and Welfare [THL], 2024b) Aftercare is a voluntary service that ends at the latest when the adolescent turns 23 or when five years have passed since the end of the clientship in child welfare services (Child Welfare Act, 2007). In 2023, 13,260 children and adolescents in Finland received aftercare services, and 12,090 of them were 18 years or older (Finnish Institute for Health and Welfare [THL], 2024a).
Care leavers’ prevalence of risk behaviors is known to be relatively high, as the majority (88%) are engaging in one or more forms of risk behavior. In addition, different forms of risk behavior among care leavers are also known to be intertwined, for example substance use is associated with problems with money management and smoking. (Petäjä et al., 2023) It is also known that many care leavers often continue with their risk behaviors in adulthood, which adversely affects their education, employment, housing, physical, and psychological health, as well as their social relationships (e.g., Kääriälä & Hiilamo, 2017). According to Eastman et al. (2019), these numerous challenges which care leavers face have been found to accumulate in certain individuals, highlighting their known vulnerability. In addition, these challenges are intergenerational in nature, which often manifests as a continuation of problems (passed from the parent to the offspring) and intergenerational placements in OOHC (Petäjä et al., 2023) due to insufficient ability to cope (Siverns & Morgan, 2019).
As care leavers struggle with many aspects of life, the challenges they experience can be seen as a collective threat to their coping ability (Kääriälä & Hiilamo, 2017; Sariaslan et al., 2022). Coping refers to the behavioral and cognitive efforts that are used to manage stressors and associated emotions which can be seen as threats to one’s well-being (Chesmore et al., 2017). Care leavers themselves have identified various components which they think are indicative for their coping and successful future. These are meaningful lives which include caring for others and having a sense of personal development, having successful personal relationships, being independent, maintaining psychological well-being, and having a structured timetable to underpin life. (Van Audenhove & Vander Laenen, 2017) In addition, care leavers approach their transition to adulthood either as a possibility for a new beginning in life or as a negative change to their life situation. Their point-of-views are affected by their experiences of their capabilities, emotions, and personal identity. (Häggman-Laitila et al., 2018) The aforementioned are affected by beliefs related to the future and a sense of resilience. Resilience refers to the ability to adapt despite a stressful situation and to succeed in coping despite the challenges that may be encountered. (Chmitorz et al., 2018)
This study used data from the electronic client and patient systems and information from aftercare professionals’ experiences of Finnish care leavers in aftercare services. The Bayesian method was used to determine the prevalence of care leavers’ coping and to predict the impact of risk behaviors and life difficulties. As a result, we generated new knowledge about the future prognosis of these adolescents that can be utilized in political decision-making and in the development of child protection and aftercare services.
Background
In addition to risk behaviors, care leavers are known to be more prone to social exclusion (Power & Raphael, 2017), mental health problems (Dubois-Comtois et al., 2021), and a lower quality of life (Rome & Raskin, 2019) compared to non-care leaver adolescents. The most recognized forms of risk behavior among care leavers are substance use and delinquency (Petäjä et al., 2022). The prevalence of substance use among care leavers varies between studies from four percent (Liu et al., 2020) to 52 percent (Petäjä et al., 2023). Substance use is more common among male care leavers (Cusick et al., 2012), as well among those who have been placed in out-of-home care three times or more (Toivonen et al., 2020). In addition, placements in institutional settings or group homes and placements due to behavioral or emotional problems are associated with substance use (Prince et al., 2019). Furthermore, substance use is associated with problems of money management and smoking (Petäjä et al., 2023). From two (Liu et al., 2020) to 46 percent (Cusick et al., 2012) of care leavers engage in delinquency. It is more common among male care leavers (Yang et al., 2017) and among those who are placed multiple time (Cusick et al., 2012) in group homes or institutional settings (Prince et al., 2019). In addition, delinquency is associated with problems of money management, drug use history and aggressive behavior (Petäjä et al., 2023).
Manifestations of risk behaviors have known to expose adolescents to social and economic problems, illnesses, accidents, and premature death (Grahn et al., 2020; Murray et al., 2020; Sariaslan et al., 2022; Xie et al., 2020). Care leavers are known to have higher rated of physical and mental morbidities (Kääriälä et al., 2021; Toivonen et al., 2020), compared to non-care leaver adolescents. The prevalence of depression and anxiety disorders among care leavers is 39 percent, compared to 7 percent among non-care leavers. In addition, neurodevelopmental disorders are more common among them than non-care leavers (26 vs. 7%, respectively) as well as conduct disorders (25 vs. 1%, respectively). Care leavers are also frequently diagnosed with comorbid combinations of these aforementioned diagnostic classes and have the highest risk-ratio (26.1, respectively) for self-harm and suicidality. (Kääriälä et al., 2021) The highlighted physical and mental morbidities increase also their all-cause morbidity prevalence (Sariaslan et al., 2022; Xie et al., 2020). In addition, obtaining upper secondary or higher education is more unlikely among care leavers than non-care leavers. They also spend 52–80 fewer days annually in employment after graduation (Pasanen et al., 2023).
The challenges care leavers experience make them more likely than their peers without OOHC background to experience social exclusion; however, there are variations in their coping strategies. Care leavers’ different coping strategies are associated with the quality of care, their transitioning phase, and the support that they receive after care. (Stein, 2006) Understanding the complex history of care leavers as well as the differences in how they cope, Stein’s (2012) theoretical model categorizes care leavers into three groups. Care leavers in the “moving on group” are likely to have some degree of reasonable stability and continuity in their lives, which includes secure attachment relationships. In addition, they have, to some extent, achieved educational success, and their preparation for leaving care has been gradual and planned. The care leavers in this group see independent living as a positive challenge and as an opportunity to gain more control over their lives. Their experiences in and after OOHC have enhanced their resilience. Unlike the “moving on group”, the “survivors” have experienced more challenges in OOHC, including disruptions and instability of placements. Care leavers in this group are more likely to leave OOHC at a younger age and with lower education levels and experience more problems afterwards. In addition, they are more likely to experience unemployment, homelessness, and problems with their relationships. The “strugglers” group of care leavers is the most disadvantaged of the three. According to Stein (2012), care leavers in the “strugglers” group have experienced multiple placements and disruptions related to relationships and education. In addition, they have earlier problems with emotional and behavioral trajectories and problems at school. After transitioning from OOHC, they were likely to be homeless, unemployed, isolated, and suffer from mental health problems. They also lack personal support before and during aftercare.
Recent studies have aimed to understand the coping and resilience of care leavers better by using profiling (Häggman-Laitila et al., 2019a; Karki et al., 2023; Shpiegel et al., 2022; Yoon et al., 2023). According to Häggman-Laitila et al. (2019a), the majority of care leavers belong to the “survivors” group, and the risk of becoming a member of the “strugglers” group is heightened for children who were taken into custody after the age of 12. In addition, substance use between the ages of 13 and 15 is associated with a higher risk of becoming a member of that most disadvantaged group. Completing secondary education, being placed in a foster family, and participating in employment services have positive effects on care leavers’ risk behaviors. It is worth noting that diagnosed psychiatric illness negatively affects care leavers’ risk behaviors whether they are enrolled in mental health services (Karki et al., 2023). Shpiegel et al. (2022) reports sustained, periodic, and patterns of non-resilience among care leavers, which are associated with several risk, protective, and child welfare factors. According to their study, females are more likely to report resilient outcomes, and having a supportive adult is associated with both sustained and periodic resilience. Diagnosed emotional or behavioral disorder by the age of 17, running away from a placement at age 17, or giving birth or fathering a child by age 19, are associated with a lesser likelihood to exhibit sustained and periodic resilience. While resilience has a multifaceted nature (Shpiegel et al., 2022; Yoon et al., 2023), children placed in OOHC are more likely to have low behavioral and emotional resilience compared to those who have not been in OOHC (Yoon et al., 2023).
Additional support programs, such as single and multiple-focused interventions, targeting care leavers’ complex needs and their coping are mainly focused on housing, education, employment, life and social skills, as well as access to health services (Häggman-Laitila et al., 2020a). According to Häggman-Laitila et al.’s (2020a) systematic review, half of the support programs offers holistic support to care leavers, and only one intervention was focused on reducing risk behavior among care leavers. Evaluations of these interventions have major problems regarding the use of definitions, assessment of instruments, and setting study designs. Due to these issues, the assessment of intervention efficacy remains limited (Chmitorz et al., 2018). According to Häggman-Laitila et al. (2020a), in the single-focused and multiple-focused programs, weak evidence of effectiveness was found, which supports the findings from Chmitorz et al. (2018).
Few previous studies have focused on explaining the factors that affect care leavers’ coping profiles (Häggman-Laitila et al., 2019a; Karki et al., 2023) and resilience (Shpiegel et al., 2022; Yoon et al., 2023); further, research understanding and predicting coping remains lacking. The aim of this study was to categorize care leavers into groups according to their coping based on a retrospective document analysis utilizing Stein’s (2012) theoretical model. In addition, the aim was to predict the factors and their combinations associated with the coping profiles.
The research questions were:
1.
How are care leavers in Finnish aftercare services distributed to the coping profiles?
2.
What kind of health related, educational, and behavioral background factors and their combinations predict care leavers’ coping profiles?
Design and Method
This cross-sectional retrospective document analysis was carried out using Moilanen et al.’s (2022) seven-phase theoretical framework (Fig. 1). In the first phase, the purpose, data and study design were decided based on the existing knowledge gap and the available data. In the second phase, the time limitation regarding cross-sectional design was determined and the sampling method was set. In the third phase, the structured electronic worksheet, which was based on systematic reviews (Häggman-Laitila et al., 2018; Häggman-Laitila et al., 2019b; Häggman-Laitila et al., 2020a) was updated based on empirical studies on the topic (Häggman-Laitila et al., 2019a; Häggman-Laitila et al., 2020b; Karki et al., 2023; Toivonen et al., 2020) in collaboration with multidisciplinary team to respond the purpose of this study. The multidisciplinary team included researchers and professionals from medicine, social science, and nursing science. In the fourth phase, the social workers and a public health nurse received training on how to fill out the structured electronic worksheet several times prior to study to ensure the uniformity of data collection. In the fifth phase, the data were collected from electronic client and patient records by the social workers and the public health nurse who handled daily the adolescents’ affairs. Additionally, the data were analyzed using appropriate methods of quantitative statistics. In sixth phase, the credibility of the study was assessed by evaluating the rigor of the documents, data and analysis. In the seventh phase, the research ethics were evaluated before, during and after the research. Thus, before starting the research, the ethical aspects related to document analysis were considered and an appropriate research permit was applied for the study. During the research, the data were handed over to the researcher anonymized and in the analyzing phase, the vulnerability of the target group was considered by reporting the results at the group level. After conducting the research, the reliability and ethics of the data based on document-analysis were critically viewed to ensure the reliability of the results.
Fig. 1
A seven-phase theoretical framework (Moilanen et al., 2022) for conducting research as a retrospective document analysis
×
Target Group and Data Collection
The target group of the study was care leavers who were active clients in aftercare services in a large Finnish city in the fall of 2020 (N = 698). They were either in aftercare for their first year, or had been for several years, or were about to leave. The data were collected into an updated structured electronic worksheet by the aftercare social workers (N = 19) and one public health nurse as they worked with care leavers on a daily basis and knew them most deeply. The data were collected from electronic customer/patient record systems, namely, ATJ, Effica, and Pegasos between September 2020 and August 2021. In addition, the aftercare social workers and the public health nurse used the information they had obtained from their conversations with the adolescents when answering the questions on the worksheet.
The structured electronic worksheet contained a total of 107 questions, which mapped the lives of the adolescents comprehensively. In more detail, the structured electronic worksheet filled in by social workers contained 101 questions, and the structured electronic worksheet completed by the public health nurse included 32 questions. Of all the questions, 20 overlapped, which meant that the same question was answered by both parties. Overlapping questions were related to care leavers’ health behavior, social problems, and use of services. The purpose of these questions was to get a comprehensive overview of these research themes, as the social workers could only use social care information, and the public health nurse could only use health care information to answer the questions on the worksheet. Both closed and open questions were posed in the worksheet, all of which were divided into 11 themes: background information, social situation, child protection background, care leavers’ social relationships, physical health, lifestyle and everyday life management skills, learning, health behavior, social problems, care leavers’ future prospects, and the services used by the care leavers. All the themes contained from one to 25 clarifying questions to form an overall picture.
With the help of background questions, a picture of the care leavers’ age, gender, nationality and mother tongue was formed. The questions surveying the social situation were related to the care leaver’s marital status, childhood family relationships, education and socioeconomic status. The questions related to the child protection background formed an overall picture of the reasons for the care leaver’s placement (which could be originated from the parents/family situation or the care leavers themselves), the form of placement(s), the number of placement(s) (“ordinary”, urgent and outpatient care placements) and the care leaver’s age at the time of the first placement. The care leaver’s social relationships were described through current interpersonal relationships, parenting, and the resources and goals related to the care leaver’s social relationships. The physical health of care leavers was comprehensively described by care leaver’s related resources and goals but also diagnosed physical and mental illnesses and the use of medications. Lifestyles and everyday management skills mapped their nutrition and exercise, sexual behavior, everyday life skills and money management, as well as support needs related to their own parenting. The questions related to learning created a comprehensive picture of the care leaver’s studying and possible learning problems as well as their resources and goals related to the theme. Questions related to health behavior mapped in detail the care leaver’s use of substances both in the past and in the present, in addition to which related goals and resources were described. Care leavers’ social problems described the threat of violence, crime, self-destructive behavior and challenges in social interaction. Their future prospects and the use of services included questions that were used to map their goals and resources in relation to the themes, and to find out the degree of use of the services. In addition to the use of services, care leavers’ attitude towards the authorities was described.
Data Analysis
Descriptive Analysis
The data were received anonymized, and frequencies and percentages were used to describe the characteristics of it (Field, 2018). We analyzed the data using IBM SPSS Statistics 27 for Windows software (IBM Corporation, 2020). The characteristics of the data were measured by standardized assessments (yes/no questions and Likert scale measurements) and included demographic information of care leavers: gender, age, language, marital and socioeconomic status, living situation, educational status, and details of placement history. These are presented in the paragraph detailing demographic information; note that the variables described therein are not in the prediction model.
The outcome variable “Coping Profile” was constructed based on research by Häggman-Laitila et al.’s (2019a) since they had a similar population, and the structure of the outcome variable suited the purpose and design of our study. The outcome variable “Coping Profile” is a sum variable consisting of three variables and four sum variables (Table 1). The three variables were delinquency; neurological, neuropsychiatric, and psychiatric illnesses; and everyday life skills. The four sum variables were compulsory education; pregnancies and parenthood; substance use; and the level of social skills. We gave points between 1 and 4 to the standardized answer options of each variable (whether it was a sum variable or not), where 1 meant the most unproblematic scenario and 4 implied the most problematic scenario (see Table 1). After the formation of the outcome variable, we categorized the care leavers into three subgroups, according to Stein’s (2012) theoretical categorization. The first group – the “moving on group” – included care leavers who had no problems in the seven measured variables; their sum variable value was 7. The second group – the “survivors” – had some problems but they managed to cope, as their sum variable value was between 8 and 14. The third group, the “strugglers”, had significant problems in the seven measured variables, and their sum variable value was 15 or more.
Table 1
The formation of the outcome variable “Coping Profile”
Variables of the Coping Profile with the scale from most unproblematic (1) to most problematic (4) scenario
Name of the (suma) variable
Included variables
Scores and their explanations according to the content of the variables
Compulsory educationa
Completed education and currently ongoing education
1= completed comprehensive school or completed higher education other than comprehensive school and currently ongoing education is higher than comprehensive school
2= completed comprehensive school (no ongoing higher education)
3= currently ongoing education is comprehensive school
4= neither completed nor ongoing comprehensive school
Delinquency
Delinquency
1= none, or no delinquency has been identified
2= minor (e.g., fines)
3= one or more convictions
4= in a crime spiral
Everyday life skills
Everyday life skills
1= has good control over all areas
2= deficiencies in one of the areas
3= deficiencies in several different areas
4= does not control
Neurological, neuropsychiatric, and psychiatric illnesses
Neurological, neuropsychiatric, and psychiatric illnesses
1= none, or no neurological, neuropsychiatric, or psychiatric illnesses have been identified
2= diagnostic tests/under investigation
3= diagnosed illness
4= several diagnosed illnesses and/or one or more diagnosed illnesses and diagnostic tests/under investigation
Pregnancies and parenthooda
Number of pregnancies, terminations of pregnancies, and children
1= no pregnancies, no abortions, no children
2 = 1 pregnancy and/or 1 abortion and/or 1 child
3 = 2 pregnancies and/or 2 abortions and/or 2 children
Substance use history and degree of substance use currently
1= no use of substances and no history of problematic/treatment-requiring substance use, or use of substances has ended and no history of problematic/treatment-requiring substance use, or no substance use has been identified and no history of problematic/treatment-requiring substance use
2= entertainment use of substances and no history of problematic/treatment-requiring substance use, or no use of substances and a problematic/treatment-requiring substance use history, or use of substances has ended, and a problematic/treatment-requiring substance use history, or no substance use has been identified and a problematic/treatment-requiring substance use history
3= periodic substance use (regardless of personal history of problematic/treatment-requiring substance use), or entertainment use of substances and a problematic/treatment-requiring substance use history
4= daily/regular use of substances (regardless of the history of problematic/treatment-requiring substance use)
Level of social skillsa
Degree of social interaction and loneliness
1= functional social interaction or no social interaction has been identified and no loneliness or no loneliness has been identified
2= occasional problems with social interaction and occasional loneliness
3= constant/regular problems with social interaction and experiences of constant/daily loneliness
4= significant/detrimental problems in social interaction and withdrawn from social relationships
a= sum variable
Bayesian Analysis
Bayesian analysis (BA) offers a powerful framework for making predictions and inferences in complex systems; in fact, it outperforms classical statistical methods in various ways (Pearl & Mackenzie, 2018; Van de Schoot et al., 2021). In contrast to classical statistics, BA excels in handling intricate data sets that exhibit high levels of uncertainty, missing data, outliers, and non-linear relationships. Also, BA can give individual-level predictions. (Arora et al., 2019; Kaplan, 2023; Van de Schoot et al., 2021) BA can be learned from data, purely from expert knowledge without data, or from a combination of the two approaches. Another advantage is that results can be presented in a visual form that is easy to interpret. (Arora et al., 2019; Kaplan, 2023; Van de Schoot et al., 2021)
The structure of a Bayesian network is a directed acyclic graph (DAG), which provides a graphical representation of the probabilistic relationships between different variables (Nilsson et al., 2021). This graphical representation allows for the depiction of complex joint probability distributions among a high number of variables. Each node in the graph represents a variable, and the arcs between nodes signify the existence of statistical dependencies. Arc direction can represent both non-causal (predictive or explanative) or causal modeling. Moreover, the conditional probability table (CPT) attached to each node describes the size of statistical dependency between two nodes. (Conrady & Jouffe, 2015) The analyst can answer “what if” questions by fixing variables to certain values, when the model shows new distributions of variables according to a fixed state or states. For example, the analyst can set the evidence fixing the state of “ADHD” to “yes”. This action prompts the whole model to update the conditional probability distributions based on the fixed value of “ADHD”. Consequently, the analyst can observe the changes in the model with this new evidence. By manipulating the variables in this manner, analysts can gain insights into how changes in one variable affect the probabilities of other variables in the network. (Conrady & Jouffe, 2015)
By using Bayesialab 11.1 software, we established a tree-augmented naïve (TAN) Bayesian modeling to demonstrate associations between independent variables (that were not included in the coping profiles) and the outcome variable (Arora et al., 2019; Kaplan, 2023; Van de Schoot et al., 2021). In the TAN model the size of each node represents the mutual information (MI) of the corresponding variable with the outcome variable. Node colors indicate the total effect on the outcome variable, with green signifying the highest impact, red the lowest, and yellow representing intermediate effects, considering both direct and indirect influences. (Friedman et al., 1997) The lines between nodes depict the relationships between variables, with line thickness corresponding to the strength of Pearson’s correlation. Blue lines represent positive correlations, while red lines denote negative correlation, indicating a potential protective effect. While traditional Naïve Bayes assumes that all predictor variables are independently associated with the outcome, TAN allows dependencies between the features. (Friedman et al., 1997) The data were randomly divided into a learning set (80% of data) and test set (20% of data). Variables’ values before splitting the data into learning and test sets, and before imputing missing values, are presented in Table 2. To mitigate the risk of overfitting, we limited the number of predictor variables to less than 5% of the sample size, which allowed for the inclusion of up to 27 variables (5% of the learning set). However, we observed that after including 16 variables, the predictive value of additional variables diminished without improving prediction results. Therefore, we selected these 16 independent variables for the final model. (Raudys & Jain, 1991) We only used arcs which were present in 90% of networks obtained from the jackknife resampling. Variables having no connection with the target variable, or showing only a minimal or unstable connection, were dropped from further analysis. We performed missing value imputations by predicting missing values using a structural equation model (structural EM) algorithm (Myers et al., 1999). We controlled the imputation of missing data by using missing data analysis included in the Bayesialab package, finding imputed values to have a distribution comparable to existing data.
Table 2
Target variable and the list of predicting variables in bayesian analysis
Variable name
Purpose of the variable
Distribution of the variable values
f (%)
Profile
Target variable: Constructed outcome variable to identify care leavers’ coping profiles
1= “Moving on group” 47 (6.7)
2= “Survivors” 512 (73.4)
3= “Strugglers” 139 (19.9)
ADHD
Measure the number of adolescents diagnosed with attention deficit hyperactivity disorder (ADHD)
0= “No” 329 (47.1)
1= “Yes” 118 (16.9)
−1= “Missing value” 251 (36.0)
Conduct dis
Measure the number of adolescents diagnosed with conduct disorder (CD)
0= “No” 333 (47.7)
1= “Yes” 114 (16.3)
−1= “Missing value” 251 (36.0)
Depression
Measure the number of adolescents diagnosed with depression
0= “No” 222 (31.8)
1= “Yes” 225 (32.2)
−1= “Missing value” 251 (36.0)
Student SB
Measure the number of adolescents who receive social benefits (SBs) for students
0= “No” 460 (65.9)
1= “Yes” 216 (30.9)
−1= “Missing value” 22 (3.2)
Delinquency (PR)
Measure the number of adolescents whose reason for the placement (PR) is due to the adolescents’ delinquency
0= “No” 408 (58.5)
1= “Yes” 94 (13.5)
−1= “Missing value” 196 (28.1)
Daily rhythm
Measure whether the adolescent has problems with the daily rhythm/daily program and, if so, how serious they are
Measure how many adolescents have problems with money management which manifest as being broke
0= “No” 255 (36.5)
1= “Yes” 99 (14.2)
−1= “Missing value” 344 (49.3)
Alcohol use (H)
Measure how many adolescents have a history of alcohol use
0= “No” 61 (8.7)
1= “Yes” 276 (39.5)
−1= “Missing value” 361 (51.7)
Drug use (H)
Measure how many adolescents have a history of drug use
0= “No” 96 (13.8)
1= “Yes” 241 (34.5)
−1= “Missing value” 361 (51.7)
Mixed use (H)
Measure how many adolescents have a history of mixed use of substances
0= “No” 285 (40.8)
1= “Yes” 52 (7.4)
−1= “Missing value” 361 (51.7)
Alcohol use (DA)
Measure how many adolescents are using alcohol during aftercare
0= “No” 46 (6.6)
1= “Yes” 268 (38.4)
−1= “Missing value” 384 (55.0)
Drug use (DA)
Measure how many adolescents are using drugs during aftercare
0= “No” 174 (24.9)
1= “Yes” 140 (20.1)
−1= “Missing value” 384 (55.0)
IV drug use
Measure how many adolescents are using intravenous drugs
1= “No” 396 (56.7)
2= “Yes” 35 (5.0)
−1= “Missing value” 267 (38.3)
Threat violence
Measure how many adolescents are experiencing the threat of violence
1= “There is no threat” 522 (74.8)
2= “There is a threat” 72 (10.3)
−1= “Missing value” 104 (14.9)
Impulsive
Measure how many adolescents have problems with social interaction, which manifest as impulsivity
0= “No” 193 (27.7)
1= “Yes” 107 (15.3)
−1= “Missing value” 398 (57.0)
Emotion EXP
Measure how many adolescents have problems with social interaction, which manifest as difficulties in emotional expression
0= “No” 160 (22.9)
1= “Yes” 140 (20.1)
−1= “Missing value” 398 (57.0)
We performed validation of the predictive model using predictive precision (percentage of correct predictions in the confusion matrix) and an ROC curve (receiver operating characteristic curve), which is a graphical representation of a predictive model’s performance across different decision thresholds (Yumoto et al., 2018). Validation was done with both learning set and test set. We calculated mutual information (MI) between all pairs of variables. MI is a Bayesian method used to quantify the statistical association between the outcome variable and the independent variables. It measures the mutual dependence between independent values of two variables, reflecting how much knowing the value of one variable reduces the uncertainty about the other. (Kaplan, 2023; Van de Schoot et al., 2021) A higher MI indicates a stronger relationship between the variables. The key feature of MI is its independence from the linearity of the variables, providing a comprehensive view of complex relationships, which makes it particularly valuable in analyzing non-linear associations (Yumoto et al., 2018).
To identify important combinations, we used Most Relevant Explanation (MRE) analysis. MRE aims to determine the critical combinations of independent variables that are associated with a particular value of the outcome variable. (Yuan et al., 2011) The MRE approach strives to locate a partial instantiation of the outcome variables that maximizes the generalized Bayes factor (GBF), thus serving as the most suitable explanation for the provided evidence. GBF provides flexibility for model comparison in complicated models where defined priors are not available. (Kaplan, 2023; Van de Schoot et al., 2021) GBF describes the strength of the explanation being maximized, with values between 3 and 10 indicating a substantial explanation, and values between 10 and 30 indicating a strong explanation. We used MRE analysis to search for combinations of variables that are associated with the outcome variables’ value “strugglers” (the most pessimistic scenario) (Yuan et al., 2011).
Ethical Considerations
The research was conducted according to good scientific practices following ethical guidelines. An appropriate research permit was obtained from the target organization. According to Finnish legislation and the Finnish National Board on Research Integrity [TENK], the use of patients’ and customers’ registry data does not require ethical approval, and the presented data can be utilized without the consent of the participants (Finnish National Board on Research Integrity, 2019).
Results
Demographic Information of the Care Leavers
The analyzed data were drawn from a total of 698 care leavers, whose ages varied between 17 and 22 years (Table 3). Nearly equal number of care leavers were males (52%) and females (47%). Native language was primarily Finnish (84%), while 16% spoke other languages (e.g., Swedish) as their mother tongue. The majority of care leavers (88%) were unmarried. The most common socioeconomic status was being a student (39%), followed by being unemployed (33%), and employed or in the military/civilian service (13%). A minority of care leavers (3%) were retired and received pension or disability pension. Most of care leavers (71%) lived in a rental apartment or as a subtenant, 11% lived with parents, friends, or relatives, and six percent resided in housing services/supported housing units. The highest completed education for the majority was comprehensive school (75%), while the highest currently ongoing education was vocational school (38%). Eight percent of care leavers had comprehensive school currently ongoing.
Table 3
Demographic data of care leavers (N = 698)
Variable name
Variables’ value options
The variable values f (%)
Age
≤18
63 (9.0)
19
155 (22.2)
20
157 (22.5)
21
157 (22.5)
22
166 (23.8)
Gender
Female
325 (46.6)
Male
362 (51.9)
Native language
Finnish
584 (83.7)
Other
111 (15.9)
Marital status
Unmarried
612 (87.7)
Cohabiting/Married/Registered partnership
57 (8.2)
Socioeconomic status
Employed or Military/Civil service
90 (12.9)
Unemployed
232 (33.2)
Student
269 (38.5)
Employment activity/course
11 (1.6)
Parental leave
21 (3.0)
Sick leave or Rehabilitation allowance
18 (2.6)
Disability pension or Pension
20 (2.9)
Socioeconomic status unknown
22 (3.2)
Form of residence
Rental apartment or subtenant
494 (70.8)
Homeless
20 (2.9)
Lives with parents/friends/relatives
73 (10.5)
Housing services or subsidized living
44 (6.3)
Other
52 (7.4)
Highest completed education
Comprehensive school
521 (74.6)
High school
35 (5.0)
Vocational school
71 (10.2)
Highest currently ongoing education
Comprehensive school
52 (7.4)
High school
61 (8.7)
Vocational school
264 (37.8)
Preparatory (employment policy) education
13 (1.9)
University of Applied Sciences
15 (2.1)
Reasons for placement, parents/family situation
Substance use
210 (30.1)
Mental health problems
143 (20.5)
Parental problems with parenting
331 (47.4)
Interparental violence
64 (9.2)
Parental fatigue/exhaustion
107 (15.3)
Parental criminality
21 (3.0)
Conflicts between parents
34 (4.9)
Somatic illness of the parent
18 (2.6)
Violence against a child/an adolescent
71 (10.2)
A chaotic life situation
72 (10.3)
Reasons for placement, the child’s/adolescents’ problems
Delinquency
94 (13.5)
Substance use
175 (25.1)
Self-destructive behavior
43 (6.2)
Mental health problems
257 (36.8)
School-related problems
269 (38.5)
Disregarding the rules of family and/or society
142 (20.3)
Running away
62 (8.9)
Somatic illness
14 (2.0)
Other
135 (19.3)
The form of placement/placements
Placement(s) in family settings
117 (16.8)
Placement(s) in institutional setting
436 (62.5)
Placement(s) in family and institutional settings
108 (15.5)
Childs’/adolescents’ age during the first placement
0–3
81 (11.6)
4–6
45 (6.4)
7–10
111 (15.9)
11–14
243 (34.8)
15–17
210 (30.1)
Number of placements/ taken into custody
1
404 (57.9)
2
191 (27.4)
3
57 (8.2)
4≥
24 (3.4)
Number of emergency placements
0
44 (6.3)
1
355 (50.9)
2
174 (24.9)
3
37 (5.3)
4≥
11 (1.6)
Number of open care placements
0
166 (23.8)
1
262 (37.5)
2
53 (7.6)
3≥
23 (3.3)
The most common reasons for the care leavers’ placement, which originated from parents or the family situation, were parental problems with parenting (47%), substance use (30%), and mental health problems (21%). Moreover, problems related to school (39%), mental health problems (37%), and substance use (25%) were the most common reasons which originated from care leavers themselves. Over half of care leavers (63%) were placed in institutional settings, while 17% were placed in family settings. A minority of them (16%) had experienced both placement forms. Care leavers had their first placement between the ages of 0 and 17, while it was the most common (35%) to be placed between the ages 11 and 14. The second most common (31%) was to be placed between the ages 15 and 17. The majority of care leavers (58%) were placed or taken into custody once, while 28% had experienced it twice. Regarding emergency placements, six percent of care leavers did not have any and 51% had experienced it once. Over a fifth of care leavers (24%) did not have any open care placements, while 38% had experienced one.
The Predictive Model of Care Leavers’ Coping Profiles
A minority of care leavers (6.7%, n = 47) belonged to the “moving on group”, while the majority (73.4%, n = 512) belonged to the “survivors” group. One-fifth of the care leavers in this study (19.9%, n = 139) belonged to the most problematic “strugglers” group.
The original data set consisted of 698 individuals, and after dividing the data into training and test sets, we had a data set with n = 559 and a test set with n = 139. In the learning sample, the model’s overall performance was strong: precision = 77.1%, reliability = 82.9%, area under ROC curve (AUC) = 93.6%, and R2 = .47. In the test set, overall precision was 76.3%, overall reliability was 76.7%, AUC was 94.9%, and R2 was 0.46. In TAN analysis we identified 16 independent variables associated with the outcome variable “Coping Profile” (see Fig. 2). Six of the identified variables were health related: alcohol use during aftercare (DA) and alcohol use in history (H); drug use during aftercare (DA) and drug use in history (H); intravenous drug use; and mixed substance use in history (Mixed use (H)). Eight of the identified variables were related to behavior: attention deficit hyperactivity disorder (ADHD); conduct disorder (Conduct dis); daily rhythm and daily program (Daily rhythm); delinquency as a placement reason (PR); depression; difficulties with emotional expression (Emotion EXP); impulsiveness (Impulsive); and threat of violence. Finally, two of the identified variables were educational related: being broke and social benefits (SB) for students. Moreover, alcohol use in personal history, alcohol use during aftercare, drug use in personal history, drug use during aftercare, and impulsivity were most strongly related to the outcome variable, which can be seen in the thickness of the lines relative to the outcome variable. Furthermore, some of the variables were also associated to each other in addition to the outcome variable. For example, care leavers’ threat of violence was associated with their drug use during aftercare. In addition, the different forms of substance use both in the present and in the past were associated with each other. The red line in the figure shows the possible protective effect of that variable, which is shown in the case of social benefits for students.
Fig. 2
Tree-Augmented Naïve Bayes (TAN) model illustrating factors associated with the outcome variable, profile. The size of each node represents the mutual information (MI) of the corresponding variable with the outcome variable. Node colors indicate the total effect on the outcome variable, with green signifying the highest impact, red the lowest, and yellow representing intermediate effects, considering both direct and indirect influences. The lines between nodes depict the relationships between variables, with line thickness corresponding to the strength of Pearson’s correlation. Blue lines represent positive correlations, while red lines denote negative correlation, indicating a potential protective effect
×
Table 4 presents the local analysis of mutual information (MI) between the outcome variable, “Coping Profile,” and the independent variables. The independent variables are listed in descending order of their binary mutual information, which represents each variable’s contribution to the value of the outcome variable. The column labeled Max. Weight of Evidence (WoE); Predicting Variable Value provides the value of the predictor variable most strongly associated with the outcome variable. This column also includes the responding WoE values, with interpretations based on established thresholds: WoE values between 5 and 10 indicate substantial evidence, while values between 10 and 15 signify strong evidence (Jeffreys, 1998; Kass & Raftery, 1995). A WoE value of ∞ indicates an indefinite association, meaning that all cases in the training set share the same value for the variable. Similarly, the Min. Weight of Evidence, Predicting Variable Value column identifies the variable states associated with the lowest WoE values, representing the weakest associations with the outcome. Results are presented separately for each value of the outcome variable “Coping Profile,” ordered by importance.
Table 4
Local analyses of mutual information (MI) between outcome variable “Coping Profile” and independent variables
Node
Relative Binary Mutual Information
Max. weight of evidence, predicting variable value
Min. weight of evidence, predicting variable value
Profile = struggler (19.9%)
Daily rhythm
20.19%
daily
6.657
no problem
−7.4236
Drug use (DA)
18.69%
yes
7.3684
no
−7.3684
Drug use (H)
17.53%
yes
7.5673
no
−7.5673
Impulsive
13.77%
yes
6.5213
no
−6.5213
IV drug use
11.48%
yes
7.2532
no
−7.2532
Threat violence
10.67%
yes
6.1455
no
−6.1455
Mixed use (H)
10.59%
yes
6.5663
no
−6.5663
Alcohol use (H)
9.46%
yes
5.5162
no
−5.5162
Emotion EXP
6.85%
yes
4.6571
no
−4.6571
Conduct dis
6.72%
yes
4.7873
no
−4.7873
Student SB
5.97%
no
6.2631
yes
−6.2631
Broke
5.64%
yes
4.5849
no
−4.5849
Delinquency (PR)
4.76%
yes
4.286
no
−4.286
Alcohol use (DA)
3.35%
yes
3.1861
no
−3.1861
ADHD
2.22%
yes
2.9632
no
−2.9632
Depression
0.28%
yes
0.9691
no
−0.9691
Profile = survivor (74.2%)
Drug use (DA)
10.13%
no
2.7664
yes
−2.7664
Daily rhythm
9.99%
no problem
0.9104
daily
−2.5247
IV drug use
7.85%
no
7.499
yes
−7.499
Impulsive
7.65%
no
2.7385
yes
−2.7385
Drug use (H)
7.18%
no
1.689
yes
−1.689
Mixed use (H)
6.66%
no
4.2193
yes
−4.2193
Threat violence
6.15%
no
2.9416
yes
−2.9416
Conduct dis
3.03%
no
1.4958
yes
−1.4958
Emotion EXP
2.75%
no
1.2482
yes
−1.2482
Delinquency (PR)
2.68%
no
1.5346
yes
−1.5346
Broke
2.58%
no
1.4868
yes
−1.4868
Alcohol use (H)
2.55%
no
0.9196
yes
−0.9196
Student SB
2.05%
yes
0.8012
no
−0.8012
ADHD
0.56%
no
0.5841
yes
−0.5841
Alcohol use (DA)
0.30%
no
0.312
yes
−0.312
Depression
0.15%
yes
0.2228
no
−0.2228
Profile = moving on (5.9%)
Alcohol use (H)
13.94%
no
∞
yes
−∞
Alcohol use (DA)
13.78%
no
∞
yes
−∞
Daily rhythm
12.65%
no problem
10.6382
occasional
−9.7788
Drug use (H)
11.43%
no
∞
yes
−∞
Depression
9.83%
no
∞
yes
−∞
Emotion EXP
6.52%
no
∞
yes
−∞
Drug use (DA)
6.27%
no
∞
yes
−∞
ADHD
5.07%
no
∞
yes
−∞
Conduct dis
5.01%
no
∞
yes
−∞
Impulsive
4.72%
no
∞
yes
−∞
Broke
4.20%
no
∞
yes
−∞
Threat violence
3.36%
no
∞
yes
−∞
Mixed use (H)
2.23%
no
∞
yes
−∞
Student SB
2.23%
yes
3.5168
no
−3.5168
IV drug use
1.45%
no
∞
yes
−∞
Delinquency (PR)
0.91%
no
4.0672
yes
−4.0672
The variable values most associated with “struggler” group were daily problems in daily rhythm (Daily rhythm=Daily), drug use during aftercare (Drug use (DA)=yes), drug use in history (Drug use (H)=yes), being impulsive (Impulsive=yes), and intravenous drug use (IV drug use = yes). The first row shows that the variable “Daily rhythm” is associated with the outcome category “struggler,” contributing 20.19% to the MI. The highest weight of evidence (WoE) is found in the “Daily rhythm” variable with the value “daily” (WoE = 6.657), indicating substantial evidence. Conversely, the “Daily rhythm” value “no problem” shows the lowest WoE (−7.4236), indicating a protective effect with substantial evidence. In the “survivor” group no reported drug use during aftercare (DA) contributed 10.13% of the probability to belong to this group and showed the highest WoE (2.7664). In the “moving on” group no alcohol use in history (H) or during aftercare (DA) contributed 13.94% and 13.78%, respectively, of the probability to belong to this group based on infinitive WoE values.
Table 5 presents the variable combinations showing the strongest association with the outcome value “strugglers”, which was the most problematic scenario. Generalized Bayes Factor (GBF) values >5 indicate strong association. Care leavers’ drug use in history and not having social benefits for students showed the strongest association with the coping profile “strugglers” (GBF = 5.9471), meaning that this combination predicted the strongest their ending up in the most vulnerable group. Furthermore, constant problems with daily rhythm/program combined with drug use in history (GBF = 5.7946) or drug use during aftercare (GBF = 5.7744) where other strong combinations enhanced participants probability to belong “strugglers” group. This enhanced probability was also evident in further combination of drug use during aftercare and not having social benefits for students (GBF = 5.7597). Overall, different forms of substance use were emphasized in the combinations. Impulsive behavior and the threat of violence also combined with substance use. In addition, a lack of students’ social benefits was included in a few combinations.
Table 5
Most relevant explanation (MRE) analysis for variable combinations showing strongest associations with the outcome variable value “Strugglers”
Most relevant explanations
Alcohol use (H)
Daily rhythm
Drug use (DA)
Drug use (H)
Impulsive
IV drug
use
Mixed
use (H)
Student SB
Threat violence
Generalized Bayes factor
yes
no
5.9471
daily
yes
5.7946
daily
yes
5.7744
yes
no
5.7597
yes
5.7113
yes
yes
5.6058
yes
5.4556
yes
no
5.4132
daily
yes
5.3774
yes
yes
5.3283
yes
5.3128
yes
daily
yes
5.2424
daily
yes
5.2263
yes
yes
5.1846
daily
yes
5.1499
In addition to the most problematic scenario, we also analyzed the variable combinations showing the strongest association with the outcome value “moving on”, which was the most optimistic scenario. The results revealed that the most optimistic scenario and the most problematic scenario were almost mirror images of each other (same variables, but opposite values). Therefore, the optimistic scenario is not shown.
Discussion
In this study we examined Finnish care leavers’ distribution to the coping profiles and identified health related, educational and behavioral background factors as well as their combinations which predicted those profiles. Overall, care leavers’ coping has been an understudied research topic as only a few studies have partially addressed it by explaining the factors affecting care leavers’ coping profiles (Häggman-Laitila et al., 2019a; Karki et al., 2023) and resilience (Shpiegel et al., 2022; Yoon et al., 2023). This study is the first to use Bayesian methods and to combine register data and employee experiences to examine the complexity of care leavers’ coping and the interconnectedness of the factors predicting it. Hence, the complexity of care leavers’ coping should be considered when planning further research. In addition to this, identifying the coping profiles of care leavers is important so that aftercare services can be developed to be more individualized and services corresponding to their coping can be offered.
Our results showed that the majority of care leavers belonged to the “survivors” coping group, which supports the findings of earlier research (Häggman-Laitila et al., 2019a; Shpiegel et al., 2022). However, the “strugglers” group (the most problematic scenario) was found to be the second largest coping group, which differs from previous profiling studies (Häggman-Laitila et al., 2019a; Karki et al., 2023; Shpiegel et al., 2022). This divergence in results can possibly be explained by the accumulation of problems for certain individuals, and the relatively high prevalence of risk behaviors among care leavers in aftercare services (Petäjä et al., 2022; Petäjä et al., 2023) as well as the different structure of the profiles depending on the research. The accumulation of problems can also be explained by care leavers’ greater risk of social exclusion when compared to non-care leavers (Power & Raphael, 2017). The reasons behind the higher risk are related to their problems with substance use (Petäjä et al., 2023), delinquency (Yang et al., 2017), educational challenges (Pasanen et al., 2023), physical and mental morbidities (Kääriälä et al., 2021; Toivonen et al., 2020), and self-supporting problems (Kääriälä & Hiilamo, 2017). As the “strugglers” are known to do worse in life than other care leavers in these coping groups (Sariaslan et al., 2022; Stein, 2012), paying special attention to this most problematic and at the same time most vulnerable group is particularly important in aftercare services. These adolescents are often the ones who disappear from the scope of services due to the severity of their problems (substance use, extreme delinquency, and so forth), although reaching them would be particularly important in terms of preventive or problem-reducing work. For them, aftercare services could be a last resort, however, reaching and engaging with them in a meaningful way can be challenging. Thus, training aftercare professionals to identify and engage with such a challenging group as well as utilize methods that promote genuine engagement is crucial.
We identified 16 health related, educational, and behavioral variables associated with the coping profiles which predicted care leavers’ coping and their threat of social exclusion. The most strongly associated variables were related to care leavers’ health and behavior problems, as the forms of using different substances in the present and in the history as well as the manifestation of conduct disorder were highlighted. These findings are supported by previous research (e.g., Häggman-Laitila et al., 2019a; Karki et al., 2023; Shpiegel et al., 2022). In this study, ADHD, depression, and conduct disorder were psychiatric disorders that most strongly predicted care leavers’ coping profiles. According to Shpiegel et al. (2022) care leavers with emotional or behavioral disorders are at higher risk for low resilience in which case the risk of social exclusion and challenges in life is greater. Additionally, over 60% of care leavers have been diagnosed with neurodevelopmental or psychiatric disorder, the most common diagnoses being anxiety disorders, depression, oppositional defiant disorder, or conduct disorder (Kääriälä et al., 2021). As emotional, behavioral, neurodevelopmental and psychiatric disorders are connected to care leavers’ coping, their early diagnosis and proper treatment are essential. However, commitment to care can be challenging for care leavers, due to their complex daily life challenges, in which case an irregular daily rhythm can make it difficult to commit to therapy, for example. In aftercare services, in multiprofessional collaboration, care leavers should be supported to commit in using mental health services and to treat their own illness. Furthermore, behavioral or emotional problems as the reason for placement have been connected to care leavers’ substance use (Prince et al., 2019). Substance use is more common among care leavers than non-care leavers, with the prevalence being up to more than half of care leavers (Petäjä et al., 2023), depending on the studies. Factors that predispose adolescents’ substance use have been extensively studied (e.g., Hagborg et al., 2020), including being academically disengaged, associating with substance-using peers, being delinquent, and having low parental monitoring (Yamada et al., 2016).
It is also worth noting that, according to the results of our study, the different forms of substance use both in the present and in the past were connected to each other, which indicates the continuous nature of these problems. Thus, it is important to strengthen multidisciplinary collaboration, especially with substance use services, in order to develop early identification of substance use risk factors. In addition, attention should be paid to identifying the quality of the adolescent’s relationships, and guiding the adolescent toward supportive relationships, because according to studies such relationships have a positive effect on the adolescents’ coping (Nuñez et al., 2022; Park & Courtney, 2023; Van Audenhove & Vander Laenen, 2017). For this reason, a working approach that focuses on relationships should be developed in aftercare and in accordance with the adolescent’s own wishes. As our analyses emphasized the destructive and long-term effects of substance use, it can be stated with some confidence that many care leavers’ substance use has not been stopped during OOHC and aftercare, which also raises the question of the proper identification of substance use and the effectiveness of the services offered. According to Häggman-Laitila et al. (2020a), while there are holistic support programs for care leavers, they identified only one intervention that aimed to reduce risk behavior. It is also worth noting that evidence of the effectiveness of those programs has remained weak (Chmitorz et al., 2018). In addition, substance use history and substance use during aftercare were connected to intravenous drug use and the threat of violence, which increases the adolescent’s vulnerability and endangers them even more (Brumley et al., 2017). Therefore, early identification, prevention, and reduction of substance use for these adolescents is crucial and requires multi-professional collaboration both in aftercare and at an earlier stage in child protection.
Problems with daily rhythm and daily program were found to be a strong risk factor for care leavers’ coping and social exclusion as it was one of the strongest independent variables that predicted care leavers’ belonging in “strugglers” group. According to Kaasinen et al. (2023) care leavers have reported improvement in their social inclusion when participating in daily activities and having a regular daily rhythm. In addition, care leavers have identified having a structured timetable as one of the components which they feel are indicative for coping and achieving a successful future (Van Audenhove & Vander Laenen, 2017). Therefore, supporting care leavers in a regular daily rhythm and enabling meaningful activities are important to consider when developing aftercare services. In addition, our results emphasized the harmfulness of substance use, especially when it came to drug use, since it strongly predicted the care leaver ending up in the most problematic group. It is worth noting that even if the use of drugs had been stopped before entering aftercare services, the effect was still significant, which reinforces their long-term harmful effects. Discussing substance use history and current substance use should be adopted as a general procedure in aftercare services among healthcare providers and other professionals that serve this population. With an open culture of discussion, it could be possible to identify drug use that also may be hidden, enabling cessation-related support to be provided in multiprofessional collaboration.
When looking at the combinations of health related, educational, and behavioral background variables predicting care leavers’ belonging in the “strugglers” group, the possible protective effect of social benefits for students, in addition to different forms of substance use and conduct disorder, were highlighted. This strengthens the conclusions of previous studies, according to which education has a protective effect on adolescents (Cutler et al., 2015; Petäjä et al., 2023). Therefore, supporting care leavers in matters related to studying and supporting education in the most possible ways that consider their possible challenges related to everyday life and daily rhythm are justified profitable areas of development.
Strengths and Limitations
Our study is essentially retrospective, as it uses data from electronic customer/patient record systems; this inevitably makes the study more susceptible to biases and confounding factors compared to prospective studies. The study misses the most marginalized people, the dead, people in institutions, and people who are unable to participate in aftercare services due, for example heavy substance addiction or severe illnesses. On the other hand, those who have been the most successful in their lives may not have considered aftercare services necessary, so they are also missing from the data. This has a potentially significant effect on the distribution of the outcome variable, but we assume that the effect on the factors of the prediction model is small.
In our study we utilized structured electronic worksheets based on multidisciplinary collaboration, systematic reviews (Häggman-Laitila et al., 2018; Häggman-Laitila et al., 2019b; Häggman-Laitila et al., 2020a), and previous studies (Häggman-Laitila et al., 2018; Häggman-Laitila et al., 2019a; Häggman-Laitila et al., 2019b; Häggman-Laitila et al., 2020a; Toivonen et al., 2020) to compile the data. Records and documentation regarding aftercare services are regulated by law; thus, completing and updating them is mandatory. Moreover, document-based data provides structured content (Moilanen et al., 2022). These are the essential strengths of the current study.
It should be noted that the purpose of the documents is to support the production of aftercare services – they are working documents which are not primarily intended for research purposes. Moreover, some documents contain information derived from conversations between care leavers and aftercare professionals; thus, these documents include a risk of misunderstanding which should be considered while interpreting analyses. In addition, the data were compiled by several aftercare professionals from documents that had been completed by other professionals. Although the professionals participating in the data collection were thoroughly instructed, there is still a risk of misunderstanding. It is worth noting that double-checking of the answers as a method for inter-rater reliability could not be carried out due to lack of resources concerning the data collectors. In addition, part of the study data were based on conversations between the employee and the care leaver, in which case double checking could not have increased the reliability of the study. It is also good to consider that aftercare is based on interaction, which is why only themes that the care leavers were willing to talk about are brought up in the documents. Therefore, for some themes, the real situation may be rather different than as reported (Beal et al., 2018).
Recommendations for Practice and Future Research
As the majority of care leavers belonged to coping profiles that had some or considerable challenges with different aspects of life, it is good to consider their complex needs and behaviors regarding aftercare services and the professionals who provide them. Care leavers’ coping was most strongly predicted by variables related to health and behavior, of which different forms of substance use and conduct disorder emerged the strongest. As they are highly prevalent among care leavers, they pose a real threat to their coping as well as their future prognosis. Consequently, the earliest possible identification, prevention and reduction of substance use and mental health problems with interventions should be a priority, especially for these adolescents. Moreover, by utilizing the research results, a checklist can be prepared for the systematic consideration and recording of mental health and substance use problems in every customer contact. In addition, further research on the effects of these services on the coping of care leavers would bring valuable additional information to support the development of aftercare services and political decision-making.
Problems with the daily rhythm and daily program, as well as drug use along with impulsivity strongly predicted care leavers’ placement to the most problematic scenario in the coping profiles and increased their threat of social exclusion. In aftercare services, more attention should be focused on the content of the care leavers’ everyday life and the circadian rhythm, which is important to maintain. Enabling meaningful and participatory activities are important elements in normalizing their daily rhythm. Therefore, further research should be targeted at the long-term effects of a broken daily rhythm and the challenges faced by care leavers as a result. Moreover, results of our study showed an unbroken cycle of substance use from past to aftercare, in which case targeting interventions focused on substance use already at the child protection stage is justified. Furthermore, training child welfare professionals to identify factors predicting coping and social exclusion is important, because “strugglers”, who were the second largest group in the results, are the most vulnerable but also multi-problematic; especially for them, early intervention can be significant. In addition, the creation of an operating model for aftercare, in which factors affecting coping profiles and social exclusion are identified, reduced, and prevented, should be at the very center of further research.
Conclusion
This study shows that identifying the coping profiles of care leavers enables more individualized aftercare services and development of interventions targeted to profiles. It is worth noting that since most care leavers belong to the “survivors” and “strugglers” groups, the aftercare services have reached, at least up to a certain point, the people who needed them. However, the systematic development and strengthening of aftercare services are justified based on this study to better meet their complex needs. Among the factors predicting coping, the most significant are related to their psychiatric illness and substance use, both factors for which early intervention and treatment are essential. In addition, our results highlight the unbroken cycle of care leavers’ substance use from history to aftercare, which in turn speaks to the lack of effective social and healthcare services and possible challenges related to identification and prevention of it in child protection services. Belonging to the most vulnerable group is predicted by problems with the daily rhythm and daily program, in addition to drug use and impulsivity. Therefore, identifying and supporting these care leavers is particularly important, as their problems often continue into adulthood, which can manifest as an increased use of society’s resources. Based on the results of our research, it can be concluded that care leavers’ coping is a complex phenomenon and that the factors influencing it are elaborately interconnected.
Acknowledgements
We would like to thank the steering group of the aftercare project for their valuable comments and guidance throughout the study. In addition, we would like to thank the research organizations’ aftercare professionals, who have made this research possible with their work.
Authors’ contributions
All of the authors contributed to the study conceptualization and design. The first draft of the manuscript was written by U.K.P. The data presented in the manuscript were analyzed by U.K.P., A.T.M., and O.P.R. The manuscript was critically revised by A.H.L., A.T.M., and O.P.R. The final version of the manuscript was read and approved by all of the authors.
Compliance with ethical standards
Conflict of interest
The authors have no competing interests to declare that are relevant to the content of this article.
Ethical approval
The research was carried out as a retrospective document analysis, in which case, in accordance with the legislation of the study country, it was not necessary to apply for a prior ethical evaluation and no separate permission from the target group was needed to obtain the data.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Met BSL Psychologie Totaal blijf je als professional steeds op de hoogte van de nieuwste ontwikkelingen binnen jouw vak. Met het online abonnement heb je toegang tot een groot aantal boeken, protocollen, vaktijdschriften en e-learnings op het gebied van psychologie en psychiatrie. Zo kun je op je gemak en wanneer het jou het beste uitkomt verdiepen in jouw vakgebied.
