(C) PLOS One [1]. This unaltered content originally appeared in journals.plosone.org.
Licensed under Creative Commons Attribution (CC BY) license.
url:https://journals.plos.org/plosone/s/licenses-and-copyright

------------



Educational and health outcomes of schoolchildren in local authority care in Scotland: A retrospective record linkage study

['Michael Fleming', 'Institute Of Health', 'Wellbeing', 'University Of Glasgow', 'Glasgow', 'United Kingdom', 'James S. Mclay', 'Department Of Child Health', 'University Of Aberdeen', 'Aberdeen']

Date: 2021-12

We calculated population attributable fractions to crudely understand the contribution of looked after care to our observed outcomes. We attributed 7.9% of self-harm admissions and 6.8% of children achieving the lowest level of academic attainment to being in care. These percentages were higher than for other outcomes investigated. This is not immediately obvious why and should be investigated further. We previously reported that children treated for ADHD, depression, and epilepsy have poorer educational and health outcomes including injury and self-harm [ 24 – 27 ]. Higher observed prevalence of these conditions among looked after children is one plausible explanation for poorer outcomes observed in this study. However, our observed associations remained after adjusting for presence of these conditions. Alternative explanations include maltreatment and early life adversity [ 36 , 37 ] and components and characteristics of the care system itself [ 38 ].

We reported greater neurodevelopmental and mental health conditions among looked after children, yet no increased risk of physical conditions such as diabetes, asthma, or skin disorders. Previous studies have reported fewer skin disorders and atopic conditions among children in care [ 8 , 9 ]. Increased epilepsy, ADHD, and depression concurs with previous studies [ 7 , 10 , 11 ]. Neurodevelopmental conditions commonly overlap, share genetic aetiology, and are highly heritable [ 35 ], and we demonstrated higher risk of neurodevelopmental multimorbidity. We reported that looked after children were more likely to be hospitalised, for any cause, injury and self-harm, and die prematurely. Previous studies have observed increased self-harm, suicidal behaviour [ 7 , 12 – 14 ], and mortality [ 16 ].

Compared with peers, looked after children had more school absences and exclusions, greater SEN, poorer examination grades, and increased unemployment postschool. Previous studies report similar patterns around absenteeism, exclusion, and attainment [ 17 – 18 ]. Our findings regarding unemployment and SEN are novel but not surprising given reports of poorer physical and mental health [ 7 – 12 ]. Looked after children in our cohort had increased SEN attributed to all types. Previous studies have reported increased behavioural issues, psychosocial and psychiatric problems, and risky behaviours [ 7 , 8 , 12 – 14 ]. Associations with absenteeism, exclusion, and SEN were stronger among less deprived children, explained by higher baseline prevalence of these outcomes among children not in care living in deprived areas.

Ours was a large, nonselective study including children across the whole of Scotland. Missing data within each variable did not exceed 1.0% and, for most, were considerably lower. Given the low percentage of missing data and the large sample size, we do not believe this affected the results and so did not perform sensitivity analyses or data imputation. Because the sampling frame was all local authority maintained mainstream and special schools, inclusion was not restricted to children from specific backgrounds or geographical areas. The large study size provided power to test for statistical interactions and undertake subgroup analyses, and we analysed a range of outcomes spanning education, health, employment, and death. A recent systematic review not only recommended a need for more population-wide research into educational outcomes of looked after children but also described limitations regarding the relatively small number of existing UK peer-reviewed studies including small samples, poor or no adjustment for confounders, and an imbalanced focus on attainment to the detriment of other equally important educational and indeed health outcomes [ 17 ]. We addressed this gap by adjusting for potential confounders: sociodemographic, maternity, and comorbid conditions and investigated whether associations were independent of physical and mental health. Being in care can impact both health and educational outcomes, which, independently, share a bidirectional relationship. This interconnection was our rationale for investigating and adjusting for both and is a clear strength versus studies investigating one or the other.

Limitations

We used existing, administrative databases established for other purposes. However, they undergo regular quality assurance checks. Linkage of education and health records relied on probabilistic matching; however, this has been validated as 99% accurate for singletons [23]. Our study did not cover private schools, but they account for less than 5% of schoolchildren. Additionally, 0.9% of children could not be linked to the CHI database, and a further 14.5% of those who did link to the CHI database could not be linked to Scottish maternity records. The percentage of children who were looked after was 1.9% among those who could not be linked to Scottish maternity records and was also 1.9% among those who could be linked suggesting bias was unlikely. Wider demographics were roughly similar among those who could versus could not be linked with respect to sex (50.9% boys versus 51.3% boys), age (mean 10.90 [SD = 3.97] versus 10.80 [SD = 4.09]), and deprivation (percentage in least deprived quintile 20.2% versus 21.6%; percentage in most deprived quintile 19.9% versus 18.9%); however, there were more children of white ethnicity among those who linked (94.7%) than did not (80.7%). Many of the nonlinkers were likely born outside Scotland. According to the 2011 Scottish Census, 11% of Scottish residents aged 5 to 19 years were born outside Scotland, consistent with 14.5% of Scottish children we could not link to Scottish maternity records.

A cohort approach ensured that initiation of looked after status predated the majority of the education and health outcomes. The exception to this is that we could not be certain about the specific temporal relationships between entering care and initiation of prescribed medications for the chronic conditions presented, meaning reverse causation is possible whereby conditions such as ADHD could contribute to difficulties that result in becoming looked after. Additionally, as reported previously, while ascertainment of children with type 1 diabetes based on receipt of necessary insulin was previously validated as 95% accurate [28], ascertaining children with asthma [29], epilepsy [26], ADHD [25], depression [27], and skin disorders [30] via prescriptions has limitations because some medications are used for alternative reasons and not all children are medicated. We had no access to primary care records and so could not confirm the underlying clinical indications for receipt of these medications. This issue is particularly apparent for antidepressants, which can be prescribed for other conditions including anxiety and obsessive-compulsive disorder. However, previous studies have reported depression to be the main reason for prescribing selective serotonin reuptake inhibitors [39–42] and 62.4% and 62.2% of children prescribed fluoxetine and citalopram to be depressed [40]. In our own previous study investigating educational and health outcomes of children prescribed antidepressants using similar methodology, our original associations remained after limiting ascertainment to children receiving solely fluoxetine or citalopram, reducing the chances of associations being affected by misclassification. Fluoxetine and citalopram are the recommended treatment and most common second-line treatment, respectively, in children under 16, and fluoxetine is the only drug licensed in the UK for treating depression in this age group.

We could not adjust for early life maltreatment and preschool social care factors, parental influence, and parental physical/mental health conditions, addiction problems, and/or criminality. Health compromises following maltreatment and early life trauma may not emerge for many decades; therefore, reporting short-term health outcomes in childhood and adolescence provides a restricted perspective resulting in likely underreporting of poor outcomes. This may have accounted for lack of association with some of the conditions investigated, and longer-term follow-up should be the focus of future work. While follow-up for interactions with the justice system was not the focus of our study, analysing these outcomes as well as longer-term employment outcomes would also add value in future work. A further limitation may arise through underreporting of neurodevelopmental disorders among looked after children because problems are attributed to early life abuse and neglect [43]. With respect to self-harm, our study could only ascertain self-harm episodes that resulted in hospitalisation, therefore excluding cases that were untreated or treated within accident and emergency or outpatient clinics.

We did not have any information on looked after status before 2009 or before starting school. Therefore, looked after children who came out of care before starting school and looked after children who came out of care before 2009 would not be ascertained as ever having been in care if they did not re-enter the system during our study period. Our exposed group may therefore have been reduced, and this would have resulted in us potentially underestimating the adverse effects of being in care on educational and health outcomes. However, our overall conclusions would remain unchanged. Access to more historic, and, in particular, preschool, social care data would help to solve this issue in future studies.

In our main analyses, looked after children comprised a heterogeneous group who had ever been looked after during the study period regardless of type of intervention, duration of care, or reasons for care. We conducted subgroup analyses to additionally investigate educational outcomes pertaining to children looked after at home versus not at home. However, future work should also consider factors that we were unable to investigate due to lack of reliable data, but which are likely to influence outcomes, namely total time spent in care and age of first contact with social care. The large sample size contributed to small p-values (<0.001). Some of the findings, therefore, while highly significant, represented fairly modest differences in real terms, which should be acknowledged during interpretation. Finally, generalisability to other studies may be limited for 3 main reasons. First, to other countries, where reasons for entry to care may differ. Second, to previous studies, which failed to adjust for important confounders adequately or at all. Third, due to heterogeneity in recording of educational and health outcomes between studies.

Despite these limitations, our findings highlight that those children who become looked after during their school years are at high educational risk compared to peers. It has been previously demonstrated that lower levels of educational achievement and fewer years of schooling are associated with poorer health and greater premature mortality [44,45], so it is essential that this vulnerable group of children receive additional educational focus. New initiatives such as “virtual schools” [46] have been designed to achieve this, and our findings reinforce the need for such additional measures if these children are to reach their full potential. However, the higher risk of neurodevelopmental multimorbidity, depression, self-harm, and premature mortality in this group suggests that closer involvement between child and adolescent mental health services and schools will also be crucial if this group is to receive appropriate mental health support and fully access the school curriculum. Our findings suggest that being placed in care away from home might ameliorate at least some poor outcomes in this group, but future research should focus on whether, and in what way, poorer outcomes are related to characteristics of the care system or precare factors, including neurodevelopmental vulnerabilities, maltreatment, adverse childhood events, and parental neurodevelopmental or psychiatric disorder. This will involve trying to access some of the data sources that we were unable to analyse during this study, such as preschool social care data, primary care data, and data around early life maltreatment and parental factors and influences. Longer-term follow-up, including linkage to wider sources such as justice and welfare data, would help uncover longer-term outcomes that we could not detect during this research. Access to richer data around duration of placement, age at first contact, and reasons for entry to care should be a priority to help uncover more around the specific drivers that influence poor outcomes.

[END]

[1] Url: https://journals.plos.org/plosmedicine/article?id=10.1371/journal.pmed.1003832

(C) Plos One. "Accelerating the publication of peer-reviewed science."
Licensed under Creative Commons Attribution (CC BY 4.0)
URL: https://creativecommons.org/licenses/by/4.0/


via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/

You are viewing proxied material from magical.fish. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.