Child mental ill health and absence from school, England: 2021 to 2022

• For children aged 5 to 16 years living in two-parent households in England, absences from school significantly contribute to children experiencing mental ill health.

• The probability of presenting at hospital with mental health issues more than doubles (increases from 1.82% to 3.77%), when absences increase from 0% to 20%, and nearly triples (increases to 5.27%) at 30% absence.

• The effect of school absences on child mental ill health is amplified if the child has other vulnerabilities, for example, a chronic physical health condition, Education, Health and Care (EHC) plan or any type of special educational needs and disability (SEND) support, or are eligible for free school meals.

• The effect is two-way, with child mental ill health also increasing a child's absence from school.

• The predicted level of absence is nearly three times as great for children who present at hospital with mental health issues (16% of total sessions in the year) compared with children who do not (6%).

• While having SEND support, an EHC plan, a chronic physical health condition, or being eligible for free school meals, is associated with a higher level of absence from school, the increase for those experiencing mental ill health is lessened for children with certain forms of SEND support or other vulnerabilities.

This article presents our analysis of the relationship between a child's absence from school and the child presenting at hospital with a mental health concern (for the purposes of this article, we refer to this as "experiencing mental ill health" (see Section 7: Glossary for further details). The analysis uses a bespoke linked dataset created for this project (including Census 2021 linked to education and health administrative data; see Section 8: Data sources and quality for details), based on a sample of 1.1 million children. It forms part of a programme of work undertaken by the Office for National Statistics (ONS) working with Loughborough University, which was initially funded by the British Academy Innovation Fellowship scheme.

We use causal econometric methods to analyse the complex relationships between a child’s mental ill health and their school absences, also controlling for important factors such as their physical health, home, family and school characteristics, and wider environment. All the data shown in the article, including in Figures 1 to 9, are results from this econometric modelling.

Our research shows that the more times a child is absent from school, the greater the probability that they will experience mental ill health. For a child with no school absences during that year, the probability of experiencing mental ill health is 1.82% on average. However, this probability more than doubles (to 3.77%) when absences increase to 20% and nearly triples (to 5.27%) at 30% of school missed. Furthermore, the effect on child mental ill health accelerates with the level of absences. For example, the increase in the probability of experiencing mental ill health when school absences increase from 30% to 40% is twice as large (increasing from 5.27% to 7.20%), compared with an increase in absences from 0% to 10% (increasing from 1.82 to 2.65 %). This indicates a nonlinear, accelerating relationship (see Figure 1 for more detail).

The impact of absence from school on the probability of experiencing mental ill health varies across different characteristics and vulnerabilities:

• The impact is greater for girls than for boys; for every additional 1% of sessions missed, the probability of girls experiencing mental ill health increases by 0.10 percentage points on average, compared with 0.08 for boys.

• The effect of absence from school on mental ill health for children who have a chronic physical health condition, or an Education, Health and Care (EHC) plan, is around 3.5 times greater than on those who do not; for every additional 1% of sessions missed, their probability of experiencing mental ill health increases by 0.29 and 0.30 percentage points, respectively, compared with 0.08 for those children who do not.

• For children with identified special educational needs and disability (SEND), the average increase in probability of mental ill health for every 1% increase in absence is nearly double that of their peers without SEND (0.16 compared with 0.08 percentage points).

• The average increase is also larger for children who are eligible for free school meals (FSM) than those who are not (0.12 and 0.09 percentage points, respectively).

The effect of absence from school on mental ill health varies by the age of the children. Analysis shows that for those in key stage 1 and 2 (aged 6 to 7 years and 10 to 11 years in 2021 to 2022), the more times a child is absent from school, the greater the probability of experiencing mental ill health, but the results are more complex for those in key stage 4 (KS4, aged 15 to 16 years). While the impact of increasing absence is statistically significant for boys in KS4, that is not the case for the full sample of KS4 children. The relationship for girls is more complex, which made it difficult to test accurately with the data available.

The relationship between a child's mental ill health and their absence from school is two-way, with the child's mental ill health also affecting their level of attendance. The predicted level of absence from school for children who experience mental ill health is nearly three times higher than for those without (16% and 6%, respectively).

Again, this effect is stronger for children with particular characteristics.

Chronic physical health condition

For children who do not experience mental ill health, the predicted level of absence increases from 5.96% for children without a chronic physical health condition to 9.53% for those with such a condition. For children experiencing mental ill health, the respective increase is from 10.60% to 16.40%.

Education, Health and Care plan

For children who do not experience mental ill health, the predicted level of absence increases from 5.99% for children who do not have an Education, Health and Care (EHC) plan to 8.90% for those who do. For children experiencing mental ill health, the respective increase is from 10.70% to 15.50%.

Special educational needs and disability

For children who do not experience mental ill health, the predicted level of absence increases from 5.83% for those without special educational needs and disability (SEND) support to 8.13% for those with SEND support. For children experiencing mental ill health, the respective increase is from 10.40% to 14.20%.

Free school meals eligibility

For children who do not experience mental ill health, the predicted level of absence increases from 5.79% for those not eligible for free school meals (FSM) to 8.66% for those who are FSM eligible. For children experiencing mental ill health, the respective increase is from 10.30% to 15.00%.

Although having any type of SEND support is associated with a higher average level of absence from school, this is not true for all forms of support. For example, while having SEND support for social, emotional and mental health (SEMH) needs or physical disability are associated with higher levels of absences, the opposite is true for those with specific learning difficulties, moderate learning difficulties, severe learning difficulties, and speech, language and communication need (SLCN) support.

More importantly, unlike the effect of absence from school on child mental ill health, experiencing mental ill health and having certain forms of SEND support do not compound each other, magnifying the level of absence. The impact of mental ill health on school attendance is slightly reduced for children with certain forms of SEND support.

For those with SEMH support specifically, the average increase in absence from school if a child experiences mental ill health is reduced by a magnitude of 0.6 percentage points. These children have a proportionally smaller increase in the predicted level of absence from school if experiencing mental ill health, compared with children without SEMH support.

We see the same relationship for children with SEND support for SLCN, severe learning difficulties, physical disability, and autistic spectrum disorder, as well as those eligible for free school meals. There is a slightly stronger effect for those with an EHC plan, suppressing the average increase by 1 percentage point.

The analysis uses a sample of 1.1 million children from a bespoke linked dataset created for this project (linking Census 2021 data to education and health administrative data; see Section 8: Data sources and quality for details).

We use an instrumental variable (IV) estimator to explore the impact of absence from school on the probability of a child experiencing mental ill health. The outcome variable is the child experiencing mental ill health between 01 April 2022 and 31 March 2023, and the treatment variable is the child's level of absence from school in the September 2021 to August 2022 academic year, as a percentage of the total number of sessions they could attend. The instruments used that satisfy the required conditions of the model are the average level of absence from school for the same key stage in the same school as the child ("relative absence"), and the level of free school meal eligibility in the school.

We use a matching estimator to explore the impact of child mental ill health on school attendance. The outcome variable is the level of absence from school in the September 2021 to August 2022 academic year. The treatment variable is the child experiencing mental ill health between 01 April 2021 and 31 March 2022. As the proportion of children experiencing mental ill health is small compared with the number who do not and given the large volume of data, generalised full matching is used.

For definitions of IV and generalised full matching estimator modelling, see Section 7: Glossary.

In all analyses, we control for variables also likely to influence a child's mental health, such as environmental features of their surrounding area, and the health of their parents. A list of these variables is in the associated dataset, and includes individual, parental and household characteristics. Further work is planned to explore the mediating effects of such characteristics, that is, how they may exacerbate or mitigate the observed relationships.

Descriptive statistics, the analysis results and a list of the factors controlled for in each of the estimators can be found in our accompanying dataset.

Chronic physical health conditions

Individuals were recorded as having a chronic physical health condition if they had attended hospital between 01 April and 31 March in the relevant year and were diagnosed as having one or more of the following:

• cancer/blood disorders

• chronic infections

• respiratory conditions

• metabolic/endocrine/digestive/renal/genitourinary conditions

• musculoskeletal/skin conditions

• neurological conditions

• cardiovascular conditions

The coding of these conditions is based on previous academic research and can be found in the dataset associated with this release. Data from 2021 to 2022 is used for most analyses. Where relevant, data from 2020 to 2023 is used and is indicated accordingly.

Education, Health and Care plan

An Education, Health and Care (EHC) plan is for children and young people aged up to 25 years who need more support than is available through special educational needs support. An EHC plan is a legal document that describes special educational needs, the support needed, and the outcomes to be achieved. The special educational provision described in an EHC plan must be provided by the child or young person's local authority. 

Free school meal eligibility

A child can be eligible for free school meals (FSM) if their parents/guardians receive Income Support, income-based Jobseeker's Allowance, income-related Employment and Support Allowance, support under Part VI of the Immigration and Asylum Act 1999, the guaranteed element of Pension Credit, Child Tax, Working Tax Credit or Universal Credit (some exceptions apply).

Generalised full matching estimator

A matching estimator is a statistical model that can enable causal inference in observation data. It emulates randomised control trials, the gold standard for causal inference, which are often not feasible for ethical, financial and feasibility reasons. A matching estimator can be applied to longitudinal and cross-sectional observational data, and it is especially useful when the treatment variable is binary.

Generalised full matching estimator matching allows the full sample to be retained where the proportion of the sample in one of the groups is comparatively small, and where the nature and volume of data does not support exact matching. This method optimises the matching process, while minimising the computing power required to do so.

Instrumental variable estimator

Instrumental variable (IV) estimator is a causal statistical model that can be applied to cross-sectional data where the input ("treatment") variable is continuous. The IV estimator uses additional variables ("instruments") to separate the proportion of change in the treatment variable that is attributable to unobserved external factors. Instruments have to satisfy two conditions: they must be related to the treatment variable but not be directly related to the outcome variable or the unobserved external factors.

Mental health conditions

Individuals were recorded as having experienced mental ill health if they had attended hospital between 01 April and 31 March, and were recorded as having a diagnosis of one of the following conditions:

• alcohol use disorder

• substance use disorder

• schizophrenia

• schizotypical and delusional disorder

• personality disorders

• other mood disorders

• bipolar disorder

• depression

• anxiety

• dementia

• obsessive-compulsive disorder

• post-traumatic stress disorder

• eating disorders

• conduct disorders

• self-harm

• behavioural/development problems

Children were also recorded as having experienced mental ill health if they met the criteria for a stress-related presentation (SRP). These are attendances at hospital for, or where they exhibited emotional, behavioural or physiological manifestations of stress (the SRP coding is based on academic research).

Adults were additionally recorded as having experienced mental ill health if they were referred to NHS Talking Therapies during the period of interest.

The coding for this can be found in our accompanying dataset.

Special educational needs and disability (SEND)

A child or young person has SEND if "they have a learning difficulty or disability which calls for special educational provision", as outlined in the Department for Education's Code of Practice. They may also have a disability, which is "a physical or mental impairment which has a long-term (a year or more) and substantial adverse effect on their ability to carry out normal day-to-day activities", as defined by the Equality Act 2010. SEND needs can be broadly categorised into four areas:

• communication and interacting

• cognition and learning

• social, emotional and mental health difficulties

• sensory and/or physical needs

Linked dataset

This work involved linking data from the following sources:

• Census 2021 (England only)

• Hospital Episode Statistics (HES) Admitted Patient Care (APC) and outpatient care (OP) records from 1 April 2021 to 31 March 2023

• Emergency Care Dataset (ECDS) records from 1 April 2021 to 31 March 2023

• NHS Talking Therapies records from 1 April 2021 to 31 March 2023

• National Pupil Database records from 1 September 2021 to 31 August 2022

• Office for National Statistics (ONS) death registrations, covering deaths registered from 01 January 2020 to 31 December 2023

• Ordnance Survey data specially curated for ONS

• Index of Multiple Deprivation (2019)

We used the Demographic Index (PDF, 549KB) (DI) to link the data. This allows anonymous data linkage across the different datasets without using personally identifiable data such as the person's name or date of birth.

The full linked dataset includes everyone who completed Census 2021 and who were usually resident in England at the time of completion. Each person was linked to a unique DI reference.

Data on admissions with relevant mental or chronic physical health conditions (see Section 7: Glossary) were extracted from Hospital Episode Statistics admitted patient care and outpatient datasets, and the Emergency Care Dataset, along with referrals to NHS Talking Therapies. These were linked using NHS reference numbers to create a single record for each person indicating any involvement with health services for a relevant condition during that financial year. These data were then linked to the DI, and the NHS number removed, before the data were linked to the census records using the DI reference.

Similarly, school data were extracted from the National Pupil Database, linking across different data tables using the anonymous Pupil Matching Reference (aPMR), to create a single record for each child. These data were then linked to the DI, and the aPMR removed, before the data were linked to the census and health records using the DI reference.

Data on the area of green space where a person lives is from Ordnance Survey data curated specially for the Office for National Statistics (ONS). These data were linked to the census data using the Unique Property Reference Number (UPRN) assigned to each census record by the ONS. Each record was linked via UPRN to a Lower-layer Super Output Area (2011 code) and then the Index of Multiple Deprivation (2019).

Data inclusion criteria

Owing to resource and time limitations, analysis was conducted on a 1.1 million sample of the over 8 million children in the full dataset. Initial analyses have focused on children living in "traditional" households with two parents because of interest in the effects of both mothers' and fathers' health on children. Children were included if:

• they were aged 5 to 16 years on census day (21 March 2021)

• they were living in a household with two parents (or step-parents)

• they had a record of completing a key stage 1, key stage 2 or key stage 4 assessment and had an absence record in the National Pupil Database for academic year September 2021 to August 2022

Quality

Overall, we consider the quality of the data to be good. Use of the demographic index to link data has enabled us to create a very large and complex dataset, supporting exploration of factors in a way that would not otherwise be possible within the project's timeframes. The linkage rates between the different datasets and the demographic index is good, varying between 95.26% (for Census 2021) and 98.99% (for National Pupil Database). Because of the robustness of the statistical techniques and volume of data included, the trends and insights shown in the results are deemed to provide a valuable contribution to the empirical evidence on this topic, as well as providing a basis for further research. As with all exploratory research, however, there are some limitations users should be aware of, outlined in the rest of this section.

While we have tried to identify the most suitable data sources, there are limitations on the datasets available for linkage in the project. The use of hospital data (Hospital Episode Statistics and Emergency Care Dataset) is likely to under-identify those experiencing mental ill health. These capture only the most severe cases. To partially overcome this issue, we undertake robustness checks where we also include stress-related presentations (SRPs). The results appear consistent and stable across the two definitions of mental ill health, and efforts are continuing to access and incorporate the Mental Health Services Dataset to improve capture of mental health issues. Furthermore, administrative health data can also under-identify those experiencing mental ill health because of uncontrolled selection effects (as some people will be more likely to seek hospital treatment).

There also exists bias introduced during linkage, as certain groups of children (for example, those of minority ethnic background and children living in more deprived neighbourhoods) are less likely to successfully link to both health and education records. This is evident in our modelling dataset, where 70% of children are of British White ethnicity, as compared with 67.1% in the census population of children. We also see an under-representation of those in the lower deciles of the Index of Multiple Deprivation, and an over-representation of those in the higher deciles. Caution should, therefore, be exercised when applying these findings to those from ethnic minority groups and more deprived areas.

Acknowledgements

We are grateful to the British Academy for funding this project via the 2022 to 2023 Innovation Fellowships scheme.

We would like to thank colleagues at the Department for Education, the Department for Health and Social Care, the Health Foundation, and Right to Succeed, as well as colleagues within the Office for National Statistics for their valuable feedback and insights on this work. Their contributions have been instrumental in refining our approach, ensuring that the analysis is both methodologically sound and directly relevant to current policy priorities.

The relationship between mental ill health and school absence is complex. Our analysis provides evidence that the association between child mental ill health and absences is bi-directional and likely to be causal. Further research exploring the impact of other factors, such as the child's home and local environment, will help us improve our understanding of the other influences on a child's mental health. It would also be beneficial to investigate the mechanism of the relationship, such as behavioural changes that might mediate the relationship between child mental ill health and school absences. Furthermore, longitudinal data would enable a more in-depth analysis of the temporal nature and sequence of events.