The impact of school absence on mental health in children and young people: Analysis of an English national birth cohort

Russell M Viner; Anna Pearce; Steven Hope

doi:10.1371/journal.pone.0336137

. 2026 Jan 14;21(1):e0336137. doi: 10.1371/journal.pone.0336137

The impact of school absence on mental health in children and young people: Analysis of an English national birth cohort

Russell M Viner ^1,^*, Anna Pearce ², Steven Hope ^1,³

Editor: Leonard Moulin⁴

PMCID: PMC12803469 PMID: 41533666

Abstract

Background

Given concerns the role of school closures in increasing mental health problems after the Covid-19 pandemic, we used pre-pandemic data to undertake a causal epidemiological analysis of the associations of school absence with later mental health problems.

Methods

Longitudinal data from the Millennium Cohort Study were collected pre-pandemic at ages 7 (in 2008), 11 (2012) and 14 years (2015), securely linked with English routine educational data on school absence in the 2 years preceding each cohort wave. We constructed marginal structural models for mental health problems (as outcome) and quartiles of absence (exposure), taking account of baseline and time-varying confounding.

Results

Those in the highest quartile of absence had odds ratios (OR) for experiencing later mental health problems of 2.216 (1.629, 3.014) at age 7 (n = 6383), and in lagged models OR of 1.508 (1.072, 2.122) at age 11 (n = 6102) and 1.903 (1.234, 2.934) at 14 years (n = 5616). Persistent absence (>10% of school year) was associated with OR for later mental health problems of 2.00 (1.56, 2.57) at 7 years, and in lagged models OR of 2.26 (1.62, 3.14) at 11 years and 2.00 (1.27, 3.16) at 14 years.

Conclusions

School absence above the second quartile doubled the odds of later mental health problems in both primary and secondary school children in pre-pandemic data. Our findings support there being a strong and potentially causal association between absence from school and later mental health problems, and suggest that absence from school is harmful for CYP’s mental health.

Introduction

High levels of both absence from school [1] and poor child and adolescent mental health [2] in many wealthy countries in the years since the COVID-19 pandemic have focused attention on relationships between school non-attendance and student mental health and wellbeing. There is a large literature showing that poor school attendance is associated with mental health problems, with mechanisms postulated to include social isolation from peer groups, isolation from supportive adults including teachers and pastoral staff, learning loss and removal from the safety-net of essential child support services that schools enable [3–8]. However systematic reviews [5,7,8] have concluded the quality of studies in this area is poor, with the great majority of studies being cross-sectional. This is problematic because there is likely a reciprocal relationship between school absence and mental health, with mental health problems causing children to miss school and school absence worsening mental health and wellbeing [9].

Any discussion of post-pandemic mental health amongst children and young people (CYP) must take into account the COVID-19 pandemic, which witnessed a rise in child mental health problems in many countries [2,10]. In England, probable mental health problems rose from 1 in 9 immediately pre-pandemic to 1 in 6 in 2020 and 2021 [2]. Post-pandemic, rates appear stable amongst 7–16 year olds although 16–17 year olds experienced a further increase to 1 in 4 in 2022 [11]. Whilst enforced absenteeism due to school closures is distinct from absence while schools are open, this form of absence has been argued to be a key factor in these increases through social isolation and learning loss alongside other factors including pandemic anxiety, bereavement, economic and family strains and post-covid effects [12–14].

Education policy requires unbiased longitudinal estimates of the impact of school absence on later mental health problems in CYP. These issues remain highly pertinent for England and many other countries 4–5 years post-pandemic as school absence rates remain stubbornly high. Further, the COVID-19 experience suggests that in future pandemics there may again be high levels of student absenteeism while schools are open, reflecting high parental anxiety. Pre-pandemic data offers the opportunity to study the impacts of school absence without having to take into account pandemic impacts such as school closures. We therefore used a nationally representative UK birth cohort with linked routine education administrative data to estimate the effects of school absence on mental health using marginal structural models and controlling for a wide set of plausible confounders. We hypothesized that higher absence increased the risk of later mental health problems.

Methods

The MCS is an ongoing population-based study representative of all children born in the UK in 2000–2001. The MCS oversampled children from economically disadvantaged areas and areas with a high proportion of ethnic minorities. Participant recruitment and selection is described elsewhere [15]. Routine educational data on school absence and educational attainments from the National Pupil Database (NPD) [16] were linked to MCS participants in England only, with data only available for those attending public schools.

We examined the association of mental health measured in surveys at ages 7, 11 and 14 years [17–19] with school attendance in the 2 years preceding each survey. We used a directed acyclic graph (DAG) to define the temporal sequence related to school absence (as exposure) and mental health problems (as outcome) and the hypothesized confounding structure of other important variables (Fig 1). Our choice of hypothesized confounding variables was guided by the literature noted above.

Figure shows a directed acyclic graph (DAG) of hypothesized pathways between school absence (Ab) and child mental health (MH) at ages 7, 11 and 14 years. Ab: school absence in 2 years prior to MH measurement (i.e., 6&7 years, 9&10 years, 12&13 years). MH: high scorer on SDQ at 7, 11 and 14 years. C: baseline confounders: measured before children started school: sex, ethnicity and deprivation. Note that C baseline confounders were included in each of the models, including for Ab9&10 and Ab12&13 however only a single arrow is shown in Fig 1 for simplicity and as C confounders were only included once in lagged models. T: time varying confounders: maternal mental health, long-term condition, previous educational attainments, engagement with school, special educational needs and bullying, area-level deprivation, prior SDQ.

Outcomes

Mental health problems at each of age 7, 11 and 14 years were assessed by parent-completed Strengths and Difficulties Questionnaire (SDQ), extensively validated as a measure of mental health in UK populations [20]. Scores above published thresholds (SDQ ‘high scorer’, defined as a score>=17) at each age were used as a proxy for probable mental health disorder [21].

Exposures

The NPD [16] provided data for each child on the number of half-day school sessions for which a student was authorized to be absent and sessions absent without authorization and the possible number of sessions for each child. The Department for Education (DfE), which collected the absence data, defines “authorised absence” as absence with permission from a teacher or other authorised representative of the schools, e.g., for illness [22]. Given these analyses aimed to understand the impact of total school absence on mental health, we combined both authorised and unauthorised absences.

To ensure exposures were prior to the outcome, we defined our years of interest for school absence as being the 2 years prior to each MCS survey at which SDQ data were available; for the age 7 survey (in 2008) we used absence data for 2007 and 2008 as no absence data were available for 2006); absence data were used for 2009 and 2010 for the age 11 survey (2012) and 2012 and 2013 and for the age 14 survey (2014−15). We averaged sessions absent for each child over the relevant two years and then calculated the proportion of possible sessions that were absent.

The association of absence with mental health problems is unlikely to be linear, and we chose to categorize absence in quartiles as our primary analysis, using the lowest (fewest sessions absent) quartile as the reference group. In supplementary analyses we a) examined authorized absence as a predictor of later mental health; and b) examined associations using total absence as a continuous logged variable and as a binary variable defining persistent absence as > 10% of possible sessions, consistent with the definition of ‘persistent absence’ used by the DfE [22].

Confounding structure

We included the following potential confounding variables (Fig 1):

Baseline confounding: Potential confounding factors measured before children started school included sex, ethnicity and residential area-level deprivation [3] (index of multiple deprivation (IMD)).

Time-varying confounding: The following variables, measured in the sweep preceding exposure measurement were adjusted for: maternal mental health (Kessler Psychological Distress scale); maternal report of presence of a long-term condition in the child [23], school experience [24] including engagement with school, bullying victimization and special or additional educational needs (SEND); and previous educational attainments obtained from linked NPD data (national data collected at age 5 years (Key Stage 1) and at age 10–11 years (Key Stage 2)). Where a required confounding variable was not available at a wave, the relevant variable from the preceding wave was substituted to ensure temporal precedence was maintained.

Analysis

We constructed marginal structural models (MSM) weighted using inverse probability weights (IPW) calculated from a minimally-sufficient confounding structure identified from the DAG. We first we estimated odds ratios (OR), with 95% confidence intervals (CIs) at 7, 11 and 14 years, in separate models, with the exposure being absence in the previous 2 years and the IPW taking account of baseline confounding and time-varying confounders relevant to each age. The IPWs also included mental health problems (SDQ high scorer) measured at the previous wave.

We then ran ‘cross-lagged’ MSMs at 11 and 14 years which took account of previous associations between absence, SDQ and time-varying confounding (e.g., for 11 year taking into account relationships and variables at 7 years; for 14 years taking into variables at 7 and 11 years), following Hope 2018 [25]. For these cross-lagged MSM we used a combined IPW calculated by multiplication of IPW for the current analysis with that for the previous analyses. We confined our analyses to complete cases as missingness was relatively low for all included variables. We did not use population weighting methods provided for MCS analyses as these are not applicable for analyses across multiple waves as we have undertaken. All analyses were undertaken in Stata 16 (StataCorp, College Station, TX) within the secure environment of the UK Data Service SecureLab.

Ethics: No ethical approval necessary for these secondary analyses of anonymized data. Data owner permissions were obtained through the UK Data Service.

Results

Characteristics of the sample for exposure (school absence) and outcome (mental health) are shown in Table 1 with characteristics of other confounding variables used to construct the IPW shown in Table 2.

Table 1. Characteristics of children in terms of school absence in preceding 2 years and mental health disorder at 7, 11 and 14 years.

			7 years (2008)	11 years (2012)	14 years (2015)
A. ABSENCE			Absence 2007−08	Absence 2010−11	Absence 2012−13
Total number of sessions		Mean	608.33 (607.59, 609.06) n = 7659	607.92 (607.19, 608.64) n = 6689	684.69 (683.78, 685.59) n = 5838
Total number of sessions		Median	616	614	694
Absence (authorized)	Annual sessions absent	Mean	29.22 (28.68, 29.76) n = 7659	24.56 (24.05, 25.07) 6689	24.61 (24.02, 25.20) n = 5838
	Annual sessions absent	Median	24	20	19
	% absent of total sessions	Mean	4.83 (4.74, 4.92) n = 7659	4.05 (3.97, 4..13) n = 6689	3.61 (3.51, 3.69) n = 5838
	% absent of total sessions	Median		3.25	2.73
Absence (unauthorized)	Annual sessions absent	Mean	2.92 (2.71, 3.11)) n = 7659	3.10 (2.90, 3.30) n = 6689	3.83 (3.56, 4.10) n = 5838
	Annual sessions absent	Median	0	0	0
	% absent of total sessions	Mean	0.49 (0.46, 0.53)) n = 7659	0.52 (0.49, 0.56) n = 6689	0.57 (0.53, 0.61) n = 5838
	% absent of total sessions	Median	0	0	0
Total Absence (authorized and unauthorized)	Annual sessions absent	Mean	32.14 (31.52, 32.75)) n = 7659	27.09 (26.66, 28.24) n = 6689	28.44 (27.74, 29.15) n = 5838
	Annual sessions absent	Median	25	22	21
	% absent of total sessions	Mean	5.31 (5.22, 5.42)) n = 7659	4.57 (4.47, 4.67) n = 6689	4.18 (4.07, 4.28) n = 5838
	% absent of total sessions	Median	4.18	3.57	–
B. MENTAL HEALTH			(n = 7659)	(n = 6689)	(n = 5838)
Probable mental health disorder	SDQ high scorer	% (95% CI)	8.04 (7.44, 8.67)	9.10 (8.43, 9.82)	10.23 (9.46, 11.03)

Open in a new tab

Notes: Table shows the characteristics of children in these analyses, i.e., A. the distribution of school absence (authorized, unauthorized and total absence) at each age, characterized by the mean and median; and B. the proportions and 95% confidence intervals (95% CI) with probable mental health problems (defined as being high scorer on the Strengths and Difficulties Questionnaire (SDQ)). Data disclosure rules operating in the SecureLab meant that inter-quartile ranges (IQR) and minimum and maximum scores could not be disclosed due to small sample sizes. Where median is not reported, the sample size for the median value also lay below disclosure rules.

Table 2. Cohort member characteristics for confounding variables.

		7 years (2008)	11 years (2012)	14 years (2015)
		% (n)	% (n)	% (n)
Sex	Male	50.65 (3870)	50.07 (3349)	50.27 (3210)
Sex	Female	49.35 (3771)	49.93 (3340)	49.73 (3175)
Ethnicity	White	80.81 (6189)	81.24 (5434)	79.14 (4620)
	Mixed ethnicity	1.24 (95)	1.17 (78)	1.18 (69)
	Indian	3.45 (264)	3.23 (216)	3.80 (222)
	Pakistani or Bangladeshi	8.16 (625)	8.43 (564)	9.66 (564)
	Black	4.47 (342)	3.93 (263)	4.01 (234)
	Other	1.87 (143)	1.99 (133)	2.21 (129)
Index of multiple deprivation (IMD) decile		% (n)	% (n)	% (n)
	1: Most deprived decile	15.0 (1,100)	14.29 (956)	13.2 (843)
	2	12.11 (888)	11.75 (786)	11.14 (711)
	3	11.19 (821)	10.85 (726)	10.73 (685)
	4	9.48 (695)	9.06 (606)	9.21 (588)
	5	9.86 (723)	9.73 (651)	10.05 (642)
	6	8.43 (618)	9.15 (612)	9.43 (602)
	7	8.41 (617)	8.87 (593)	9.19 (587)
	8	8.22 (603)	8.4 (562)	8.54 (545)
	9	8.41 (617)	8.64 (578)	9.08 (580)
	10: Least deprived decile	8.89 (652)	9.25 (619)	9.43 (602)
Enjoys school	Always	72.04 (5230)	64.15 (4291)	64.51 (3766)
	Usually	24.16 (1754)	29.54 (1976)	29.27 (1709)
	Sometimes	3.33 (242)	5.34 (357)	5.28 (308)
	Not at all	0.47 (34)	0.84 (56)	0.77 (45)
Bullied at school	Never	65.24 (4997)	65.67 (4393)	65.60 (3824)
	Once or twice	27.30 (2091)	26.94 (1802)	27.05 (1579)
	Several times	5.54 (424)	5.47 (366)	5.53 (323)
	Many times	1.67 (128)	1.58 (106)	1.54 (90)
Long-term condition	No	80.18 (5859)	81.25 (5427)	85.96 (4569)
Long-term condition	Yes	19.82 (1448)	18.75 (1252)	14.04 (746)
Child has special needs	No	–	91.90 (6147)	92.03 (5373)
Child has special needs	Yes	–	7.95 (532)	7.74 (452)
KS1 reading & writing score level	Level 3 or above	12.48 (886)
	Level 2b or above	47.96 (3404)
	Level 2 or above	20.78 (1475)
	Level 1 or below	18.78 (1333)
KS1 achieved Maths level	Level 3 or above	23.47 (1666)
	Level 2b or above	53.47 (3795)
	Level 2 or above	14.48 (1028)
	Level 1 or below	8.58 (609)
		Continuous	continuous	continuous
Key Stage 2 point score	Mean (sd)	–	28.79 (3.37) n = 6654	28.80 (4.46) n = 5825
Key Stage 2 point score	Median	–	27 n = 2923	27 n = 2568
Maternal mental health: Kessler score	Mean (sd) & n	3.29 (3.81) n = 6938	3.25 (3.89) n = 6389	3.28 (3.92) n = 5515
Maternal mental health: Kessler score	Median	2	2	2

Open in a new tab

Notes: Table shows the characteristics of the sample in terms of the confounding variables that were used in constructing the inverse probability weighting (IPW). For categorical variables, proportion (%) and sample size (n) are shown. For continuous variables, the mean, median and sample size (n) are shown. Ethnicity was collapsed for analyses into the categories shown.

The sample for these analyses consists of those children with data on both school absence and mental health disorder at each age: n = 7659 at 7 years, n = 6689 at 11 years, n = 5838 at 14 years. Proportions of children with probable mental health problem (SDQ high scorer) were 8.0% at age 7 years, 9.1% at 11 years and 10.2% at 14 years. The mean proportion of time that children were absent from school was 5.3% in the two years before age 7, 3.6% in the 2 years before age 11 and 4.2% in the 2 years before age 14. The majority of absence was authorized at all ages, with unauthorized absence accounting for only 0.49% to 0.57% of possible school sessions across ages. The sample for these analyses was broadly representative of the child population of England, being 50.7% male and 80.8% white at age 7, with similar proportions at 11 and 14 years. Consistent with the MCS recruitment strategy, 15% of children were in the most deprived decile at age 7, with 8–12% in all other deciles.

Table 3 shows odds ratios (OR) and 95% confidence intervals (CI) for risk of probable psychological disorder associated with quartiles of total school absence in the previous 2 years, adjusted for confounding factors and previous measurements of the outcome in unlagged and lagged models (that took account of previous relationships between absence and mental health). Final model samples were 6383 at 7 years (83.3% of the analytic sample at 7 years), 6102 at 11 years (91.2% of the analytic sample) and 5616 at 14 years (96.2% of analytic sample).

Table 3. Odds ratios for probable psychological disorder at 7, 11 and 14 years associated with quartiles of preceding school absence in unlagged and lagged mediation models.

	7 years (re absence aged 6–7 year) Model n = 6383			11 years (re absence aged 10-11y) Model n = 6102			14 years (re absence aged 12-13y) Model n = 5616
Unlagged Models	% absence in quartiles	OR (95% CI)	p-value	% absence in quartiles	OR (95% CI)	p-value	% absence in quartiles	OR (95% CI)	p-value
Absence quartiles: Ref: Lowest quartile	<2.2%	1		<1.9%	1		<1.5%	1
2nd quartile	2.3-4.2%	1.163 (0.825, 1.638)	0.388	2.0-3.6%	1.201 (0.866, 1.666)	0.272	1.5-3%	1.272 (0.937, 1.725)	0.122
3rd quartile	4.3-7.1%	1.421 (1.031, 1.958)	0.032	3.7-6.2%	1.433 (1.041, 1.971)	0.027	3.1-5.7%	1.935 (1.459, 2.566)	<0.0001
Highest quartile	7.2% plus	2.216 (1.629, 3.014)	<0.0001	6.2% plus	1.919 (1.368, 2.693)	<0.0001	5.8% plus	2.690 (2.045, 3.539)	<0.0001
*Lagged models for 11 and 14 years*					Model n = 5517			Model n = 4597
Absence quartiles: Ref: Lowest quartile				<1.9%	1		<1.5%	1
2nd quartile				2.0-3.6%	1.185 (0.857, 1.638)	0.304	1.5-3%	1.038 (0.696, 1.549)	0.856
3rd quartile				3.7-6.2%	1.249 (0.898, 1.736)	0.186	3.1-5.7%	1.759 (1.172, 2.641)	0.006
Highest quartile				6.2% plus	1.508 (1.072, 2.122)	0.018	5.8% plus	1.903 (1.234, 2.934)	0.004

Open in a new tab

Notes: Table shows odds ratios (OR) and 95% confidence intervals (CI) for risk of probable psychological disorder associated with total school absence in the previous 2 years, adjusted for confounding factors as shown in Table 2. Note that absence quartiles were derived separately for each of the 2 year periods prior to age 7, 11 and 14 years. Models are shown unlagged and lagged. The unlagged models only use inverse probability weights (IPW) derived for that analysis (separate IPW were derived for models for 7, 11 and 14 years). Lagged models were then run for 11 and 14 years taking account of previous absence and previous associations between mental health and absence, i.e., the model for 11 years used the product of the IPW for 7 and 11 years, and the model for 14 years used the product of the IPW for 7, 11 and 14 years.

In unlagged models, there was a dose response relationship between absence and mental health problems across all ages, with those in the highest quartiles of absence at around twice as likely (or more) to experience mental health problems as those in the lowest quartile. For example, odds ratios (OR) for experiencing mental health problems (highest quartile of absence compared with the lowest) was 2.216 (1.629, 3.014) at age 7, 1.919 (1.368, 2.693) at age 11 and 2.690 (2.045, 3.539) at 14 years.

In lagged models, the dose response relationships were still seen, but these were weaker than in the individual models. For example, the increased risk of mental health problems in the highest quartile of absences was (OR 1.5 (1.071, 2.122) at 11 years and 1.903 (1.234, 2.934) at 14 years. In the cross-lagged models, there no longer appeared to be an elevated risk in the second absence quartile.

Supplementary analyses Table 4 shows models for absence as a continuous (log transformed) variable and for persistent absence (absence >10%). In the continuous analyses, the risk of later mental health problems increased with duration of school absence at 7 years but not at 11 or 14 years in either unlagged or lagged analyses. Persistent absence was found in 11.7% in 2007−08, 8.2% in 2009−10 and 6.8% in 2012−13. Being persistently absent was associated with OR for later mental health problems of 2.00 (1.56, 2.57) at 7 years and in lagged models OR were 2.26 (1.62, 3.14) at 11 years and 2.00 (1.27, 3.16) at 14 years. A sensitivity analysis examining authorized absence (see Table 5) showed similar findings.

Table 4. Supplementary analyses: odds ratios for probable psychological disorder at 7, 11 and 14 years associated with preceding school absence as a continuous or binary variable in unlagged and lagged mediation models.

	7 years (re absence aged 6–7 year) Model n = 6383		11 years (re absence aged 10-11y) Model n = 6102		14 years (re absence aged 12-13y) Model n = 5616
Unlagged Models	OR (95% CI)	p-value	OR (95% CI)	p-value	OR (95% CI)	p-value
School absence (continuous; anti-logged)	1.373 (1.136, 1.658)	0.001	1.030 (0.935, 1.133)	0.552	1.153 (0.969, 1.372)	0.109
Persistent absence (Absence >10%)	2.005 (1.564, 2.568)	<0.0001	2.98 (2.166, 4.093)	<0.0001	2.370 (1.811, 3.100)	<0.0001
*Lagged models for 11 and 14 years*			Model n = 5517		Model n = 4597
School absence (continuous; anti-logged)			1.055 (0.957, 1.163)	0.281	1.074 (0.935, 1.233)	0.312
Persistent absence (>10%)			2.255 (1.620, 3.138)	<0.0001	1.999 (1.267, 3.155)	0.003

Open in a new tab

Table shows odds ratios (OR) and 95% confidence intervals (CI) for risk of probable psychological disorder associated with total school absence in the previous 2 years, adjusted for confounding factors. Models are shown for total absence first as a continuous variable and second as a dichotomous variable (highly absent, defined as absence >10%), and third in quartiles of absence. Models are shown unlagged and lagged. The unlagged models only use inverse probability weights (IPW) derived for that analysis (separate IPW were derived for models for 7, 11 and 14 years). Lagged models were then run for 11 and 14 years taking account of previous absence and previous associations between mental health and absence, i.e., the model for 11 years used both IPW for 7 and 11 years, and the model for 14 years used IPW for 7, 11 and 14 years.

Table 5. Odds ratios for probable psychological disorder at 7, 11 and 14 years associated with preceding authorized school absence in unlagged and lagged mediation models.

	7 years					11 years					14 years
Unlagged models		OR	p-value	95% CI upper	95%CI lower		OR	p-value	95% CI upper	95%CI lower		OR	p-value	95% CI	CI
	N = 6383					N = 6102					N = 5616
Absence quartiles: Ref: Lowest quartile	<2.1%					<1.7%					<1.4%
2nd quartile	2.1-3.8%	1.158	0.397	0.825	1.625	1.7-3.2%	1.292	0.122	0.933	1.789	1.4-2.7%	1.171	0.296	0.870	1.576
3rd quartile	3.9-6.4%	1.318	0.091	0.957	1.814	3.3%−5.5%	1.412	0.034	1.027	1.941	2.8-4.9%	1.732	0.000	1.313	2.284
Highest quartile	6.5%+	2.124	<0.0001	1.571	2.872	5.6%+	1.965	<0.0001	1.411	2.736	5.0%+	2.385	0.000	1.828	3.112
Lagged models, i.e., 11 and 14y accounting for previous absence and previous associations between mental health and absence
						N = 5517					N = 4597
Absence quartiles: Ref: Lowest quartile						<1.7%					<1.4%
2nd quartile						1.7-3.2%	1.312	0.099	0.950	1.811	1.4-2.7%	1.073	0.725	0.724	1.592
3rd quartile						3.3%−5.5%	1.256	0.167	0.909	1.737	2.8-4.9%	1.423	0.072	0.969	2.092
Highest quartile						5.6%+	1.545	0.010	1.112	2.147	5.0%+	1.551	0.038	1.025	2.349

Open in a new tab

Notes: Table shows odds ratios (OR) and 95% confidence intervals (CI) for risk of probable psychological disorder associated with authorized school absence in the previous 2 years, adjusted for confounding factors as shown in Table 3. Models are shown for 2 variables for total absence; first as a continuous variable and second in quartiles of unauthorized absence. Note that quartiles were derived separately for each of 7, 11 and 14 years. Models are shown unlagged and lagged. The unlagged models only use inverse probability weights (IPW) derived for that analysis (separate IPW were derived for models for 7, 11 and 14 years). Lagged models were then run for 11 and 14 years taking account of previous absence and previous associations between mental health and absence, i.e., the model for 11 years used both IPW for 7 and 11 years, and the model for 14 years used IPW for 7, 11 and 14 years. All N are unweighted. Sample sizes are shown by model.

Discussion

We found that levels of absence above the second quartile of the distribution of sessions missed were associated with greater odds of mental health problems 1–2 years later. This is a relatively low threshold for school absence, and suggests that the upper half of the population distribution of school sessions missed were at risk of mental health problems related to absence. Estimates for lagged models for 11 and 14 years were smaller than in unlagged models, suggesting that failing to account for confounding histories likely results in upwards-biased estimates. We found that the increased risk for mental health problems was apparent from only approximately 4% total absence amongst 7 and 11 year olds, with an approximately 40% increase in odds of mental health problems, although in lagged models for age 11 risk was only significantly increased from around 6% absence. For 14 year olds, increased risk was apparent from approximately 3% total absence, but with larger effect sizes, with an 80% increase in odds of mental health problems from approximately 3% absence and a 90% increase from approximately 6% absence. The risk of mental health problems among those experiencing persistent absence compared to those who did not, was approximately double at all ages.

A number of studies have shown links between school absence and mental health [3–8], with persistent or prolonged absenteeism most strongly associated with these problems [26]. This is true for emotional or internalizing disorders, behavioural or externalizing problems [27] and for those with neurodevelopmental problems [3]. Both absences that are authorized by school and those that are not (unauthorized absence) have been associated with mental health problems [23], with relationships stronger for secondary compared to primary school children [23]. Data from the English National Child and Adolescent Mental Health survey in early 2022 found that those with a mental health disorder were nearly twice as likely to have persistent absence as those without problems [11]. However the majority of studies are cross-sectional.

Most research posits mental health problems as a result of school absence, with poor peer relationships [28,29] and isolation [13] as mechanisms. However, there is some longitudinal evidence that relationships may also run from poor mental health to later absence [5], truancy and school drop-out [30]. Pathways linking absence with mental health problems are likely to be complex. First, there may be reciprocity between the two, so that absence may lead to poor mental health due to isolation [13], but poor mental health may increase the chances of children missing or avoiding school. Second, there may be confounding by educational attainments and success, as those with poorer mental health are more likely to have poor attainments [31], but poor attainments are associated with absence [4]. Third, wider social determinants may confound relationships between absence and mental health, and they both may be influenced by a range of shared risk factors. Both absence and poor mental health may reflect common underlying social determinants such as deprivation and family factors [29,32], although few studies have examined these factors [5]. Fourth, key confounders of the relationship between absence and mental health such as bullying are likely to change over time during childhood and adolescent, however studies have thus far failed to account for time-varying confounding. Studies have also largely failed to recognise that previous absence is one of the strongest predictors of school absence [33].

What this study adds

Ours is the first study in this area to use causal inference methods in nationally representative longitudinal data, taking account of temporality, time-varying confounding and earlier associations between absence and mental health. Standard statistical methods to adjust for confounding that is influenced by previous exposure can lead to biased estimates [34]. First, conditioning on a confounder that is affected by previous exposure removes part of the effect of interest. Second, if there are unmeasured factors associated with both school absence and mental health, conditioning or adjusting for a confounding variable (as is done in standard regression) potentially induces further confounding through unmeasured pathways from exposure to outcome via these unmeasured variables, known as ‘collider bias’ [34]. We used weighted marginal structural models (MSM) [34,35] to account for the complex hypothesised pathways between exposure (school absence), outcome (mental health), baseline confounding and time-varying confounding. MSMs are an effective method of generating unbiased potentially causal estimates in the presence of time-varying confounding in social epidemiology [36].

We identified a potentially causal association between school absence and later mental health problems in primary and secondary school children, and identified low thresholds (e.g., 3–4% total absence) above which risk of mental health problems increased.

Limitations of this study

Our data are subject to a number of limitations. While we were able to adjust for a range of confounders, these will not have been measured perfectly and it is possible that there are other important confounders that we have not adjusted for. Our analyses were based upon our DAG, which included relationships between mental health and absence at previous waves and thus accounted for potential bi-directional relationships between mental health and absence. However we cannot exclude unobserved residual confounding explaining some of our findings.

Not all MCS participants in England had linked educational data, likely reflecting loss during the linkage process and lack of consent for linkage. It is likely these data are missing at random as non-linkage is unlikely to be related to any of the variables under study here. There was minor excess attrition in the sample for cohort members from the most deprived deciles and from some ethnic groups; whilst this is unlikely to impact the substantive relationship between school absence and mental health problems, the association of deprivation with both absence and mental health problems suggest our findings may under-estimate its size.

For analytic parsimony we only examined overall mental health problems using the SDQ total difficulties scale, rather than examine associations of absence with different types of mental health problems, such as self-reported anxiety or depression.

We used pre-pandemic data to examine these associations, as this inherently excluded a range of pandemic-related confounding factors and allowed us to more cleanly study school absence as an exposure. However, we recognise that prepandemic absence from school is an imperfect analogy for enforced absence due to school closures during the COVID-19 pandemic and therefore the associations observed in this study might have changed during and since the pandemic.

Policy implications and conclusions

Our findings support there being a strong and potentially causal association between absence from school and later mental health problems in primary and secondary school CYP. Absence for more than 3–4% of total school-time significantly increased the odds of later mental health problems, with greater impact in adolescence. This threshold is notably lower than the 10% threshold used to define persistent absence in England. Whilst higher levels of post-pandemic absence across the school population mean that quartiles of school absence will be different to those identified here, our findings suggests that mental health harms related to absence are likely at lower levels of school absence.

Our use of prepandemic data within a causal mediation framework allows us to be confident that absence from school is harmful for CYP’s mental health separate to any impacts from other pandemic stressors.

Estimates suggest that pupils in England lost 61 days of schooling on average between March 2020 and April 2021, approximately 32% of a school year [37], far higher than the thresholds used in this study. These findings will be important for policymakers, given the persisting levels of absence post-pandemic seen in countries such as England and in relation to decisions regarding school closure in any future pandemics. There is a need for further study of the association between absence and mental health in the post-pandemic period. Our data suggest that persisting high levels of school absence post-pandemic are contributing to rising levels of mental health problems amongst young people, and support conceptualisation of school absence as a public health problem. Our findings therefore support action on reducing school absence as a potentially useful part of efforts to improve children and young people’s mental health.

Data Availability

The data for the Millennium Cohort Study are held in a public depository, the UK Data Archive. All data used in this paper are publically available on special application through the UK Data Service Secure Lab (https://ukdataservice.ac.uk/help/secure-lab/). University College London, UCL Institute of Education, Centre for Longitudinal Studies, Department for Education. (2024). Millennium Cohort Study: Linked Education Administrative Datasets (National Pupil Database), England: Secure Access. [data collection]. 3rd Edition. UK Data Service. SN: 8481, DOI: http://doi.org/10.5255/UKDA-SN-8481-3 University of London. Institute of Education. Centre for Longitudinal Studies. (2017). Millennium Cohort Study: Third Survey, 2006. [data collection]. 7th Edition. UK Data Service. SN: 5795, http://doi.org/10.5255/UKDA-SN-5795-4 University of London. Institute of Education. Centre for Longitudinal Studies. (2017). Millennium Cohort Study: Fourth Survey, 2008. [data collection]. 7th Edition. UK Data Service. SN: 6411, http://doi.org/10.5255/UKDA-SN-6411-7 University of London. Institute of Education. Centre for Longitudinal Studies. (2017). Millennium Cohort Study: Fifth Survey, 2012. [data collection]. 4th Edition. UK Data Service. SN: 7464, http://doi.org/10.5255/UKDA-SN-7464-4 University of London. Institute of Education. Centre for Longitudinal Studies. (2018). Millennium Cohort Study: Sixth Survey, 2015. [data collection]. 3rd Edition. UK Data Service. SN: 8156, http://doi.org/10.5255/UKDA-SN-8156-3.

Funding Statement

The author(s) received no specific funding for this work.

References

1.Pupil absence in schools in England: autumn term 2021/22. England: Department for Education; 2022. [Google Scholar]
2.Newlove-Delgado T, Williams T, Robertson K, McManus S, Sadler K, Vizard T, et al. Mental Health of Children and Young People in England, 2021: Wave 2 follow up to the 2017 survey. NHS Digital; 2021. [Google Scholar]
3.John A, Friedmann Y, DelPozo-Banos M, Frizzati A, Ford T, Thapar A. Association of school absence and exclusion with recorded neurodevelopmental disorders, mental disorders, or self-harm: a nationwide, retrospective, electronic cohort study of children and young people in Wales, UK. Lancet Psychiatry. 2022;9(1):23–34. doi: 10.1016/S2215-0366(21)00367-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
4.Carter-Pokras OD, Bugbee BA, Gold RS, Lauver PE, Aiken R, Arria AM. Utilizing student health and academic data: a county-level demonstration project. Health Promot Pract. 2021;22(2):193–203. doi: 10.1177/1524839919862796 [DOI] [PMC free article] [PubMed] [Google Scholar]
5.Finning K, Ukoumunne OC, Ford T, Danielsson-Waters E, Shaw L, Romero De Jager I, et al. The association between child and adolescent depression and poor attendance at school: a systematic review and meta-analysis. J Affect Disord. 2019;245:928–38. doi: 10.1016/j.jad.2018.11.055 [DOI] [PubMed] [Google Scholar]
6.Lawrence D, Dawson V, Houghton S, Goodsell B, Sawyer MG. Impact of mental disorders on attendance at school. Aust J Educ. 2019;63(1):5–21. doi: 10.1177/0004944118823576 [DOI] [Google Scholar]
7.Finning K, Ford T, Moore D. Mental health and attendance at school. Cambridge: Cambridge University Press; 2022. [Google Scholar]
8.Epstein S, Roberts E, Sedgwick R, Polling C, Finning K, Ford T, et al. School absenteeism as a risk factor for self-harm and suicidal ideation in children and adolescents: a systematic review and meta-analysis. Eur Child Adolesc Psychiatry. 2020;29(9):1175–94. doi: 10.1007/s00787-019-01327-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
9.Children’s Commissioner for England. New NHS Data shows school absence rates are higher in children with a probable mental health disorder. Office of the Children’s Commissioner for England; 2022. [Google Scholar]
10.Vizard T, Sadler K, Ford T, Newlove-Delgado T, McManus S, Marcheselli F, et al. Mental Health of Children and Young People in England, 2020: Wave 1 follow up to the 2017 survey. NHS Digital, Health and Social Care Information Centre; 2020. [Google Scholar]
11.Mental Health of Children and Young People in England 2022 - wave 3 follow up to the 2017 survey. England: NHS Digital; 2022. [Google Scholar]
12.Fegert JM, Vitiello B, Plener PL, Clemens V. Challenges and burden of the Coronavirus 2019 (COVID-19) pandemic for child and adolescent mental health: a narrative review to highlight clinical and research needs in the acute phase and the long return to normality. Child Adolesc Psychiatry Ment Health. 2020;14:20. doi: 10.1186/s13034-020-00329-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
13.Loades ME, Chatburn E, Higson-Sweeney N, Reynolds S, Shafran R, Brigden A, et al. Rapid systematic review: the impact of social isolation and loneliness on the mental health of children and adolescents in the context of COVID-19. J Am Acad Child Adolesc Psychiatry. 2020;59(11):1218–1239.e3. doi: 10.1016/j.jaac.2020.05.009 [DOI] [PMC free article] [PubMed] [Google Scholar]
14.Ford T, Moore DA. Reflections on the Impact of Covid-19 on Children’s Education and Mental Health. In: Moore DA, Finning K, Ford T, editors. Mental Health and Attendance at School. Cambridge: Cambridge University Press; 2022. pp. 181-91. [Google Scholar]
15.Millennium Cohort Study: Centre for Longitudinal Studies, UCL Institute of Education. 2023. [cited 2023 Mar 24]. Available from: https://cls.ucl.ac.uk/cls-studies/millennium-cohort-study/ [Google Scholar]
16.Millennium Cohort Study: Linked Education Administrative Datasets (National Pupil Database). 2nd Edition. England: Secure Access: UCL Institute of Education Centre for Longitudinal Studies, editor.: UK Data Service; 2021. [Google Scholar]
17.Millennium Cohort Study: Third Survey, 2006. In: Institute of Education: Centre for Longitudinal Studies, editor.: UK Data Service; 2017. [Google Scholar]
18.Millennium Cohort Study: Sixth Survey, 2015. In: Institute of Education Centre for Longitudinal Studies, editor.: UK Data Service; 2018. [Google Scholar]
19.Millennium Cohort Study: Fifth Survey, 2012. 4th ed. In: Institute of Education Centre for Longitudinal Studies, editor.: UK Data Service; 2017. [Google Scholar]
20.Goodman R. The strengths and difficulties questionnaire: a research note. J Child Psychol Psychiatry. 1997;38(5):581–6. doi: 10.1111/j.1469-7610.1997.tb01545.x [DOI] [PubMed] [Google Scholar]
21.Goodman R, Ford T, Simmons H, Gatward R, Meltzer H. Using the Strengths and Difficulties Questionnaire (SDQ) to screen for child psychiatric disorders in a community sample. Br J Psychiatry. 2000;177:534–9. doi: 10.1192/bjp.177.6.534 [DOI] [PubMed] [Google Scholar]
22.A guide to absence statistics. London: Department for Education; 2019. [Google Scholar]
23.Finning K, Ford T, Moore DA, Ukoumunne OC. Emotional disorder and absence from school: findings from the 2004 British Child and Adolescent Mental Health Survey. Eur Child Adolesc Psychiatry. 2020;29(2):187–98. doi: 10.1007/s00787-019-01342-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
24.Van Eck K, Johnson SR, Bettencourt A, Johnson SL. How school climate relates to chronic absence: a multi-level latent profile analysis. J Sch Psychol. 2017;61:89–102. doi: 10.1016/j.jsp.2016.10.001 [DOI] [PubMed] [Google Scholar]
25.Hope S, Pearce A, Chittleborough C, Deighton J, Maika A, Micali N, et al. Temporal effects of maternal psychological distress on child mental health problems at ages 3, 5, 7 and 11: analysis from the UK Millennium Cohort Study. Psychol Med. 2019;49(4):664–74. doi: 10.1017/S0033291718001368 [DOI] [PMC free article] [PubMed] [Google Scholar]
26.Lereya ST, Patel M, Dos Santos JPGA, Deighton J. Mental health difficulties, attainment and attendance: a cross-sectional study. Eur Child Adolesc Psychiatry. 2019;28(8):1147–52. doi: 10.1007/s00787-018-01273-6 [DOI] [PubMed] [Google Scholar]
27.Gonzálvez C, Martín M, Vicent M, Sanmartín R. School refusal behavior and aggression in Spanish Adolescents. Front Psychol. 2021;12:669438. doi: 10.3389/fpsyg.2021.669438 [DOI] [PMC free article] [PubMed] [Google Scholar]
28.Nakamura-Thomas H, Sano N, Maciver D. Determinants of school attendance in elementary school students in Japan: a structural equation model. Child Adolesc Psychiatry Ment Health. 2021;15(1):38. doi: 10.1186/s13034-021-00391-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
29.Heyne D, Kearney CA, Finning K. Mental health and attendance at school: Setting the scene. In: Moore DA, Finning K, Ford T, editors. Mental health and attendance at school. Cambridge: Cambridge University Press; 2022. pp. 1–21. [Google Scholar]
30.Lindhardt L, Lindholdt L, Lund T, Mortensen OS. Self-reported mental health in adolescents attending school and its association with later school dropout: A prospective 2.5-year follow-up study. Scand J Public Health. 2022;50(8):1164–71. doi: 10.1177/14034948221089112 [DOI] [PubMed] [Google Scholar]
31.Smith NR, Marshall L, Albakri M, Smuk M, Hagell A, Stansfeld S. Adolescent mental health difficulties and educational attainment: findings from the UK household longitudinal study. BMJ Open. 2021;11(7):e046792. doi: 10.1136/bmjopen-2020-046792 [DOI] [PMC free article] [PubMed] [Google Scholar]
32.Ingul JM, Klöckner CA, Silverman WK, Nordahl HM. Adolescent school absenteeism: modelling social and individual risk factors. Child Adolesc Ment Health. 2012;17(2):93–100. doi: 10.1111/j.1475-3588.2011.00615.x [DOI] [PubMed] [Google Scholar]
33.Children’s Commissioner for England. Education history and attendance. England: Office of the Children’s Commissioner for England; 2022. [Google Scholar]
34.Hernan MA, Robins JM. Causal inference: What if? Boca Raton: Chapman & Hall/CRC; 2022. [Google Scholar]
35.Cole SR, Hernán MA. Constructing inverse probability weights for marginal structural models. Am J Epidemiol. 2008;168(6):656–64. doi: 10.1093/aje/kwn164 [DOI] [PMC free article] [PubMed] [Google Scholar]
36.Gilsanz P, Young JG, Glymour MM, Tchetgen Tchetgen EJ, Eng CW, Koenen KC, et al. Marginal Structural Models for Life-Course Theories and Social Epidemiology: Definitions, Sources of Bias, and Simulated Illustrations. Am J Epidemiol. 2022;191(2):349–59. doi: 10.1093/aje/kwab253 [DOI] [PMC free article] [PubMed] [Google Scholar]
37.Major LE, Eyles A, Machin S. Learning loss since lockdown: variation across home nations. London: Centre for Economic Performance, London School of Economics; 2021. [Google Scholar]

PLoS One. doi: 10.1371/journal.pone.0336137.r001

Decision Letter 0

Leonard Moulin

18 Jan 2025

-->PONE-D-24-53350-->-->The impact of school absence on mental health in children and young people: analysis of an English national birth cohort-->-->PLOS ONE

Dear Dr. Viner,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.-->-->-->

Please ensure that your revised manuscript addresses all reviewer comments comprehensively; however, I would like to draw your particular attention to the following three points:

-->-->-->-->-->-->

Disentangle school closures and absenteeism: The manuscript conflates these concepts, which have distinct implications, and we recommend clarifying their differences.
Endogeneity concerns in the causal relationship: Despite controlling for maternal mental health, the bidirectional relationship between absenteeism and mental health raises concerns that should be acknowledged.
Mechanisms linking absenteeism to mental health: The manuscript could benefit from a brief discussion on how absenteeism disrupts key benefits provided by schools, such as structure, social interaction, and essential services.

-->-->-->Please submit your revised manuscript by Mar 04 2025 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org . When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:-->

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.
A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.
An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols . Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols .

We look forward to receiving your revised manuscript.

Kind regards,

Leonard Moulin

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. We note that you have indicated that there are restrictions to data sharing for this study. PLOS only allows data to be available upon request if there are legal or ethical restrictions on sharing data publicly. For more information on unacceptable data access restrictions, please see http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions.

Before we proceed with your manuscript, please address the following prompts:

a) If there are ethical or legal restrictions on sharing a de-identified data set, please explain them in detail (e.g., data contain potentially identifying or sensitive patient information, data are owned by a third-party organization, etc.) and who has imposed them (e.g., a Research Ethics Committee or Institutional Review Board, etc.). Please also provide contact information for a data access committee, ethics committee, or other institutional body to which data requests may be sent.

b) If there are no restrictions, please upload the minimal anonymized data set necessary to replicate your study findings to a stable, public repository and provide us with the relevant URLs, DOIs, or accession numbers. For a list of recommended repositories, please see

https://journals.plos.org/plosone/s/recommended-repositories. You also have the option of uploading the data as Supporting Information files, but we would recommend depositing data directly to a data repository if possible.

We will update your Data Availability statement on your behalf to reflect the information you provide.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

-->Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented. -->

Reviewer #1: Yes

Reviewer #2: Yes

**********

-->2. Has the statistical analysis been performed appropriately and rigorously? -->

Reviewer #1: Yes

Reviewer #2: Yes

**********

-->3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.-->

Reviewer #1: Yes

Reviewer #2: No

**********

-->4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.-->

Reviewer #1: Yes

Reviewer #2: Yes

**********

-->5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)-->

Reviewer #1: The authors equate school closures with school absences. However, these are distinct phenomena with different implications for students, educators, and communities. School closures, such as caused by the Covid-19 pandemic, result in the inability of all students to attend their educational institution. In contrast, school absences generally refer to the nonattendance of individual students or smaller groups due to personal reasons, such as illness, family responsibilities, or truancy.

By conflating these two terms, the authors risk oversimplifying the challenges and misaligning proposed solutions. For example, strategies effective in addressing individual absenteeism may be inadequate for managing the systemic disruptions of school closures. Similarly, the long-term academic, social, and emotional consequences differ, requiring nuanced understanding and action. It is essential to disentangle these concepts to create precise and effective interventions that cater to the unique demands of each situation.

The authors rely on pre-pandemic data to make inferences about the post-pandemic world, but this approach has significant limitations. The Covid-19 pandemic has drastically altered the educational landscape, including shifts in teaching methods, technological reliance, student engagement, and systemic inequalities. Pre-pandemic data, while valuable for understanding historical trends, may fail to capture the profound and lasting changes brought about by the pandemic. Therefore, I would avoid the framing around the pandemic and focus on the relation ship between school absences and mental health generally. In addition, I would add this as limitation given that the relationship between school absences and mental health may have changed post-pandemic.

The authors fail to discuss why school absences should have a detrimental impact on adolescents' health and why the relationship between school attendance and health is bidirectional. This omission leaves a critical gap in their argument, as it does not address the underlying mechanisms that connect school attendance with health outcomes. The authors do not have to write about this extensively but should add a short discussion to the introduction, e.g., schools provide structured environments that promote routine, access to physical activity, social interaction, and critical services such as mental health counselling and nutritious meals. Absences disrupt these benefits, potentially leading to negative health outcomes such as increased social isolation, decreased physical activity, and higher stress levels due to academic setbacks or disconnection from peer networks.

The authors provide little information on handling missing data, a critical issue in longitudinal studies like the Millennium Cohort Study (MCS). Missing data, if unaddressed, can bias results and weaken validity. The authors should explicitly discuss the extent and patterns of missingness (e.g., MCAR, MAR, or MNAR) and adopt appropriate methods to address it.

Multiple imputation is recommended for handling missing data. This method generates several plausible datasets by filling in missing values based on observed patterns, then combines the results to produce robust estimates while accounting for imputation uncertainty. It is particularly suitable for complex datasets like the MCS.

The authors also use later waves of the MCS, which are subject to participant attrition. Systematic dropout can lead to nonrepresentative samples, introducing bias. To mitigate this, the authors should apply attrition weights to adjust for the likelihood of participants remaining in the study, ensuring the analysis more accurately reflects the original population.

The authors should elaborate on their use of the marginal structural model (MSM) and its advantages compared to standard regression techniques. My understanding is that MSMs are particularly suited for causal inference in longitudinal studies with time-varying exposures and confounders. They address issues where standard regression might fail, such as time-dependent confounding—when prior exposure influences both the outcome and subsequent confounders. This does not become clear. It may also be helpful to include equations to understand the estimation procedure.

While the authors’ use of a marginal structural model (MSM) is commendable for addressing time-varying confounding and bidirectionality, the analysis relies on several strong assumptions that warrant caution in interpreting causal claims. MSMs assume no unobserved confounding, meaning that all relevant confounders must be measured and included in the model. In practice, this is a strong assumption, as unmeasured variables could bias the results.

Additionally, the validity of the results depends on the correct specification of the parametric models used to calculate the weights. Misspecification can introduce substantial bias and undermine the causal interpretation of the findings.

Given these limitations, it is advisable to tone down causal language and frame the results as robust associations rather than definitive causal effects.

Reviewer #2: Thanks for this paper, it is good to see the linking to absence data to provide this kind of analysis. As mentioned in the article it serves as a rare more longitudinal and causal exploration of school attendance and mental health.

There are some minor and quick corrections to consider indicated below.

The only bigger picture thing to consider, which might only mean acknowledgement in introduction and discussion are the figures around school absence for the last few years which have increased dramatically. The introduction seems to focus on mental health difficulty increased and the potential effect of “absence” from school during lockdown. It therefore seems to miss the policy and practice emphasis on school attendance too and the potential implications of the findings if a risk factor for mental health has worsened since this data was collected.

Review the first sentence of Methods in abstract, it seems to be missing some words to make full sense.

The conclusions of the abstract don’t fit with the findings. Above the second quartile versus highest quartile and persistent absence in results summary. The final sentence of the abstract feels as though it is reaching given the timing of data here versus pandemic stressors.

Ought to be Department for Education throughout not “of”.

Page 4 missing end of sentence “However”

Given that the dataset is not available as used, could more be said about access to the NPD and the linking necessary, so this would be clear to other researchers interested in similar.

Could more be said to explain the decision to categorise into quartiles?

Could more be said to justify the selection of confounding variables.

Why are total session averages much higher in 2012-2013? (that could be explained in table 1 notes.

Table 1: Is 26.66 the correct total for 11 year olds?

The sample is described as broadly representative in terms of gender and ethnicity, is that also the case for amount of school absence?

Might a discussion point consider the alternative causal path from mental health difficulty to absence, the bi-directional relationship has been suggested in some research.

In the policy implications there seems to be a focus on days of schooling lost during lockdown. While this is relevant for the argument made in the paper, there ought to be acknowledgment that absence rates are higher in schools now compared to before pandemic.

I was surprised given the messages around 4% and 6% absence being a seemingly important threshold in this study’s findings, particularly as the profile of quartiles of absence is likely to be quite different now compared to pre-pandemic absence records, that an implication was not raised in terms of persistence absence and its cut-off. The findings seem to suggest that the arbitrary threshold of 10% of sessions missed might miss earlier identification of children who may be at risk of mental health difficulties if we were to use absence as an indicator of risk? Schools after all place emphasis on persistent absence as a key threshold to monitor.

**********

-->6. PLOS authors have the option to publish the peer review history of their article (what does this mean? ). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy .-->

Reviewer #1: No

Reviewer #2: No

**********

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/ . PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org . Please note that Supporting Information files do not need this step.

PLoS One. 2026 Jan 14;21(1):e0336137. doi: 10.1371/journal.pone.0336137.r002

Author response to Decision Letter 1

1 Jun 2025

Response to reviewers

Dear Editor

Thank you for asking us to revise this paper. We respond point by point to the Editorial and reviewer comments below

Russell Viner on behalf of all authors

Editorial comments:

1. Disentangle school closures and absenteeism: The manuscript conflates these concepts, which have distinct implications, and we recommend clarifying their differences.

Response: We agree these are distinct and have distinct implications, although of course in both children are not in school. We have rewritten and restructured the introduction and discussion to emphasize that we are interested in school absence (i.e. absenteeism), but also to recognise that in the current post-pandemic generation these impacts can be difficult to separate from the forced absence due to school closures during the pandemic. Therefore we argue that examining pre-pandemic data can be helpful to ensure these impacts are distinct. We have removed any suggestion that these data might be helpful in understanding impacts of school closures, although we note that in future pandemics there may be very high rates of absenteeism while schools are open, due to parental anxiety, and that understanding such impacts upon mental health will be important for policy.

2. Endogeneity concerns in the causal relationship: Despite controlling for maternal mental health, the bidirectional relationship between absenteeism and mental health raises concerns that should be acknowledged.

Response: We undertook non-lagged and ‘lagged’ analyses specifically to account for prior associations between mental health and absence. Because of the temporal aspects of these variables, this did not require a directionality to the relationship i.e. it adjusted for the relationship at earlier periods regardless of directionality. We recognise that our DAG implied a directionality to this – although this is required of directed acyclic graphs. However our models through our DAG did allow us to account for previous bi-directional relationships between mental health and absence at earlier times. However we recognise in our revised limitations sections that residual confounding may remain.

3. Mechanisms linking absenteeism to mental health: The manuscript could benefit from a brief discussion on how absenteeism disrupts key benefits provided by schools, such as structure, social interaction, and essential services.

Response: We have included a brief discussion of this as suggested in the first paragraph of the revised introduction.

Review comments.

Reviewer #1

1. Re equating school absence with school closure

Response: see response to Editorial comment 1 above

2. Handling of missing data; use of attrition weights.

Response: Our analyses relate to a subset of the MCS for which linked educational data were available. This is a subset, albeit a large subset, of the overall dataset. It is therefore not a more standard MCS analysis. It is difficult to imagine on what basis one might multiply impute this linked dataset, as we note in our current limitations section that “Not all MCS participants in England had linked educational data, likely reflecting loss during the linkage process and lack of consent for linkage. It is likely these data are missing at random as non-linkage is unlikely to be related to any of the variables under study here.”

Similarly, attrition weights are not available for this linked dataset and we believe it would be incorrect to use weights for non-linked datasets. We note in our current limitations section that “There was minor excess attrition in the sample for cohort members from the most deprived deciles and from some ethnic groups; whilst this is unlikely to impact the substantive relationship between school absence and mental health problems, the association of deprivation with both absence and mental health problems suggest our findings may under-estimate its size.”

3. elaborate on their use of the marginal structural model (MSM) and its advantages compared to standard regression techniques

Response: We cite a number of papers in our methods section that outline the advances of MSM methods compared to standard regression techniques, and our Methods and What this study adds sections provide considerable detail on the advantages of these methods. We believe that additional detail is beyond the scope of this paper.

4. unobserved confounding

Response: we have strengthened our limitations section (as above) to better recognise the potential for unobserved confounding.

5. frame the results as robust associations rather than definitive causal effects

Response: We agree and have modified our language to conclude we have shown ‘strong and potentially causal’ associations using causal inference methods.

Reviewer 2.

1. issues about school absence and lockdown.

Response: see response to Editorial comment above

2. various typos and missing words

Response: Thank you, these have been corrected.

3. access to NPD

Response: We have added a note that “These data are available on special application through the UK Data Service Secure Lab (https://ukdataservice.ac.uk/help/secure-lab/).”

4. quartiles?

Response: We explain in our current methods section that we chose a categorical analysis as the association was very unlikely to be linear. This proved to be the case e.g. the medians were 0 as in Table 1. We chose quartiles as we wished to identify which sections of the child population were at risk, whether this was only those more extreme (i.e. highest quartile).

5. selection of confounding variables?

Response: We have revised our methods section to note “Our choice of hypothesized confounding variables was guided by the literature noted above.”

6. total absence sessions for 11 year olds?

Response: we had inverted the mean and the low confidence interval – thank you, this has been corrected.

7. representativeness in terms of absence?

Response: We do not have this information, yet this is a very large sample with linkage to routine administrative data, allowing the presumption of representativeness.

8. bidirectional causal pathways?

Response: see response to Editorial comments.

9. lockdown versus school absence?

Response: see response to Editorial comments.

10. Thresholds of absence – 4 and 6%

Response: we agree that the level of absence which appears to lead to mental health harms is lower than the UK definition of persistent absence at >10%. We have added the following to our conclusions: “This threshold is notably lower than the 10% threshold used to define persistent absence in England. Whilst higher levels of post-pandemic absence across the school population mean that quartiles of school absence will be different to those identified here, our findings suggests that mental health harms related to absence are likely at lower levels of school absence.”

Attachment

Submitted filename: Response reviewers PlosOne absence MH 040225.docx

pone.0336137.s002.docx^{(20.8KB, docx)}

PLoS One. doi: 10.1371/journal.pone.0336137.r003

Decision Letter 1

Leonard Moulin

16 Jul 2025

-->PONE-D-24-53350R1-->-->The impact of school absence on mental health in children and young people: analysis of an English national birth cohort-->-->PLOS ONE

Dear Dr. Viner,

One of the original reviewers has expressed satisfaction with the revisions made, while a new reviewer was invited and has provided a number of constructive comments and suggestions to be considered in the ongoing editorial process.

Please submit your revised manuscript by Aug 30 2025 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org . When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:-->

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.
A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.
An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

We look forward to receiving your revised manuscript.

Kind regards,

Leonard Moulin

Academic Editor

PLOS ONE

Journal Requirements:

If the reviewer comments include a recommendation to cite specific previously published works, please review and evaluate these publications to determine whether they are relevant and should be cited. There is no requirement to cite these works unless the editor has indicated otherwise.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

-->Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.-->

Reviewer #2: All comments have been addressed

Reviewer #3: (No Response)

**********

-->2. Is the manuscript technically sound, and do the data support the conclusions?

Reviewer #2: Yes

Reviewer #3: Yes

**********

-->3. Has the statistical analysis been performed appropriately and rigorously? -->

Reviewer #2: Yes

Reviewer #3: Yes

**********

-->4. Have the authors made all data underlying the findings in their manuscript fully available?

Reviewer #2: No

Reviewer #3: Yes

**********

-->5. Is the manuscript presented in an intelligible fashion and written in standard English?

Reviewer #2: Yes

Reviewer #3: Yes

**********

-->6. Review Comments to the Author

Reviewer #2: Thanks for addressing the review points and the tracked changes on the manuscript.

I think there could have been a little more emphasis on school attendance as an increased concern now compared to during and prior to the pandemic, with some indication of how figures have worsened, particularly the near doubling of persistent absence. But the point around attendance as policy concern and caution around causation is now clear and this is important given the propensity for assuming poor mental health causes absence on less robust data and analysis than this.

Minor point for proofs, “Persistently absence” in the abstract should be “persistent absence”

Reviewer #3: The study “The impact of school absence on mental health in children and young people: analysis of an English national birth cohort“ uses inverse probability weighting and linked administrative and longitudinal survey data to estimate the effect of school absences on mental health. The study addresses an important topic, makes an important contribution to the literature, and uses high-quality data and a rigorous method. Nonetheless, there are a few points that need to be addressed before publication.

1. I suggest that the authors describe the method they used in the final sentence of the introduction, rather than stating that they “estimated the robust effect of school absence on mental health.” I recommend wording such as: “We use marginal structural models and control for a large set of plausible confounders to estimate the causal effect […]”.

2. The authors should describe the dataset in more detail. For example, it should be noted that the MCS oversampled children from areas with a high proportion of ethnic minorities and from economically disadvantaged areas. Additionally, it should be mentioned that the NPD only contains data on students attending public schools.

3. The authors should explicitly state that they did not use any survey weights, which is usually suggested for MCS data.

4. Please clarify exactly how many cases were excluded due to complete case analysis. Also, indicate how frequently missing values were replaced using data from the previous sweep.

5. Figure 1 does not include an arrow from C to Ab9&10 and Ab12&13. Was this omission intentional? Is C included in all models?

6. Why do the authors rely on parent-reported SDQ scores rather than self-reported SDQ scores? Are the results robust when using self-reported SDQ scores?

7. I do not understand why the authors present the results of the “unlagged model” in the abstract. Their DAG clearly shows that lagged absences and lagged mental health are confounders of the association between current absences and current mental health. The authors also acknowledge this in the discussion: “Estimates for lagged models for 11 and 14 years were smaller than in unlagged models, suggesting that failing to account for confounding histories likely results in upwards-biased estimates.” Therefore, if the authors aim to estimate the causal effect, they should focus exclusively on the results of the lagged model.

8. The authors need to provide a more detailed explanation of how they selected potential confounders. Their aim is to estimate the causal effect of absences on mental health, which can be achieved with marginal structural models only if all relevant confounders are included. Therefore, simply stating, “Our choice of hypothesized confounding variables was guided by the literature noted above,” is not sufficient. I recommend that the authors consider the risk factors for absences summarized in the meta-analysis by Gubbels et al. (2019), as many of these factors may also influence mental health.

9. In addition, I believe important confounders are missing, such as the socio-economic status of the family (which affects absences independently of area-level deprivation; Klein et al., 2020).

10. I do not fully understand the following sentences in the “limitations” section: “Not all MCS participants in England had linked educational data, likely reflecting loss during the linkage process and lack of consent for linkage. It is likely these data are missing at random as non-linkage is unlikely to be related to any of the variables under study here.”

First, it should be noted that the NPD does not contain information on students attending private schools. Second, do you mean “missing completely at random” (MCAR) or “missing at random” (MAR)? MCS-NPD linkage is related to demographics and socio-economic status, but conditional on these factors, it is not associated with internalizing and externalizing problems (Dräger et al. 2024, Appendix S1).

References

Dräger, J., Klein, M., & Sosu, E. M. (2024). Trajectories of school absences across compulsory schooling and their impact on children’s academic achievement: An analysis based on linked longitudinal survey and school administrative data. Plos one, 19(8), e0306716.

Gubbels, J., van der Put, C. E., & Assink, M. (2019). Risk factors for school absenteeism and dropout: A meta-analytic review. Journal of Youth and Adolescence, 48, 1637-1667.

Klein, M., Sosu, E. M., & Dare, S. (2020). Mapping inequalities in school attendance: The relationship between dimensions of socioeconomic status and forms of school absence. Children and Youth Services Review, 118, 105432.

**********

-->7. PLOS authors have the option to publish the peer review history of their article (what does this mean? ). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy .-->

Reviewer #2: No

Reviewer #3: Yes: Jascha Dräger

**********

PLoS One. 2026 Jan 14;21(1):e0336137. doi: 10.1371/journal.pone.0336137.r004

Author response to Decision Letter 2

18 Oct 2025

Please see attached cover letter

Attachment

Submitted filename: Viner PlosOne response to reviewers 300925.docx

pone.0336137.s003.docx^{(20.8KB, docx)}

PLoS One. doi: 10.1371/journal.pone.0336137.r005

Decision Letter 2

Leonard Moulin

21 Oct 2025

The impact of school absence on mental health in children and young people: analysis of an English national birth cohort

PONE-D-24-53350R2

Dear Dr. Viner,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice will be generated when your article is formally accepted. Please note, if your institution has a publishing partnership with PLOS and your article meets the relevant criteria, all or part of your publication costs will be covered. Please make sure your user information is up-to-date by logging into Editorial Manager at Editorial Manager® and clicking the ‘Update My Information' link at the top of the page. For questions related to billing, please contact billing support .

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Leonard Moulin

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

PLoS One. doi: 10.1371/journal.pone.0336137.r006

Acceptance letter

Leonard Moulin

PONE-D-24-53350R2

PLOS ONE

Dear Dr. Viner,

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now being handed over to our production team.

At this stage, our production department will prepare your paper for publication. This includes ensuring the following:

* All references, tables, and figures are properly cited

* All relevant supporting information is included in the manuscript submission,

* There are no issues that prevent the paper from being properly typeset

You will receive further instructions from the production team, including instructions on how to review your proof when it is ready. Please keep in mind that we are working through a large volume of accepted articles, so please give us a few days to review your paper and let you know the next and final steps.

Lastly, if your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

You will receive an invoice from PLOS for your publication fee after your manuscript has reached the completed accept phase. If you receive an email requesting payment before acceptance or for any other service, this may be a phishing scheme. Learn how to identify phishing emails and protect your accounts at https://explore.plos.org/phishing.

If we can help with anything else, please email us at customercare@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Leonard Moulin

Academic Editor

PLOS ONE

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Attachment

Submitted filename: Response reviewers PlosOne absence MH 040225.docx

pone.0336137.s002.docx^{(20.8KB, docx)}

Attachment

Submitted filename: Viner PlosOne response to reviewers 300925.docx

pone.0336137.s003.docx^{(20.8KB, docx)}

Data Availability Statement

[pone.0336137.ref001] 1.Pupil absence in schools in England: autumn term 2021/22. England: Department for Education; 2022. [Google Scholar]

[pone.0336137.ref002] 2.Newlove-Delgado T, Williams T, Robertson K, McManus S, Sadler K, Vizard T, et al. Mental Health of Children and Young People in England, 2021: Wave 2 follow up to the 2017 survey. NHS Digital; 2021. [Google Scholar]

[pone.0336137.ref003] 3.John A, Friedmann Y, DelPozo-Banos M, Frizzati A, Ford T, Thapar A. Association of school absence and exclusion with recorded neurodevelopmental disorders, mental disorders, or self-harm: a nationwide, retrospective, electronic cohort study of children and young people in Wales, UK. Lancet Psychiatry. 2022;9(1):23–34. doi: 10.1016/S2215-0366(21)00367-9 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0336137.ref004] 4.Carter-Pokras OD, Bugbee BA, Gold RS, Lauver PE, Aiken R, Arria AM. Utilizing student health and academic data: a county-level demonstration project. Health Promot Pract. 2021;22(2):193–203. doi: 10.1177/1524839919862796 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0336137.ref005] 5.Finning K, Ukoumunne OC, Ford T, Danielsson-Waters E, Shaw L, Romero De Jager I, et al. The association between child and adolescent depression and poor attendance at school: a systematic review and meta-analysis. J Affect Disord. 2019;245:928–38. doi: 10.1016/j.jad.2018.11.055 [DOI] [PubMed] [Google Scholar]

[pone.0336137.ref006] 6.Lawrence D, Dawson V, Houghton S, Goodsell B, Sawyer MG. Impact of mental disorders on attendance at school. Aust J Educ. 2019;63(1):5–21. doi: 10.1177/0004944118823576 [DOI] [Google Scholar]

[pone.0336137.ref007] 7.Finning K, Ford T, Moore D. Mental health and attendance at school. Cambridge: Cambridge University Press; 2022. [Google Scholar]

[pone.0336137.ref008] 8.Epstein S, Roberts E, Sedgwick R, Polling C, Finning K, Ford T, et al. School absenteeism as a risk factor for self-harm and suicidal ideation in children and adolescents: a systematic review and meta-analysis. Eur Child Adolesc Psychiatry. 2020;29(9):1175–94. doi: 10.1007/s00787-019-01327-3 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0336137.ref009] 9.Children’s Commissioner for England. New NHS Data shows school absence rates are higher in children with a probable mental health disorder. Office of the Children’s Commissioner for England; 2022. [Google Scholar]

[pone.0336137.ref010] 10.Vizard T, Sadler K, Ford T, Newlove-Delgado T, McManus S, Marcheselli F, et al. Mental Health of Children and Young People in England, 2020: Wave 1 follow up to the 2017 survey. NHS Digital, Health and Social Care Information Centre; 2020. [Google Scholar]

[pone.0336137.ref011] 11.Mental Health of Children and Young People in England 2022 - wave 3 follow up to the 2017 survey. England: NHS Digital; 2022. [Google Scholar]

[pone.0336137.ref012] 12.Fegert JM, Vitiello B, Plener PL, Clemens V. Challenges and burden of the Coronavirus 2019 (COVID-19) pandemic for child and adolescent mental health: a narrative review to highlight clinical and research needs in the acute phase and the long return to normality. Child Adolesc Psychiatry Ment Health. 2020;14:20. doi: 10.1186/s13034-020-00329-3 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0336137.ref013] 13.Loades ME, Chatburn E, Higson-Sweeney N, Reynolds S, Shafran R, Brigden A, et al. Rapid systematic review: the impact of social isolation and loneliness on the mental health of children and adolescents in the context of COVID-19. J Am Acad Child Adolesc Psychiatry. 2020;59(11):1218–1239.e3. doi: 10.1016/j.jaac.2020.05.009 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0336137.ref014] 14.Ford T, Moore DA. Reflections on the Impact of Covid-19 on Children’s Education and Mental Health. In: Moore DA, Finning K, Ford T, editors. Mental Health and Attendance at School. Cambridge: Cambridge University Press; 2022. pp. 181-91. [Google Scholar]

[pone.0336137.ref015] 15.Millennium Cohort Study: Centre for Longitudinal Studies, UCL Institute of Education. 2023. [cited 2023 Mar 24]. Available from: https://cls.ucl.ac.uk/cls-studies/millennium-cohort-study/ [Google Scholar]

[pone.0336137.ref016] 16.Millennium Cohort Study: Linked Education Administrative Datasets (National Pupil Database). 2nd Edition. England: Secure Access: UCL Institute of Education Centre for Longitudinal Studies, editor.: UK Data Service; 2021. [Google Scholar]

[pone.0336137.ref017] 17.Millennium Cohort Study: Third Survey, 2006. In: Institute of Education: Centre for Longitudinal Studies, editor.: UK Data Service; 2017. [Google Scholar]

[pone.0336137.ref018] 18.Millennium Cohort Study: Sixth Survey, 2015. In: Institute of Education Centre for Longitudinal Studies, editor.: UK Data Service; 2018. [Google Scholar]

[pone.0336137.ref019] 19.Millennium Cohort Study: Fifth Survey, 2012. 4th ed. In: Institute of Education Centre for Longitudinal Studies, editor.: UK Data Service; 2017. [Google Scholar]

[pone.0336137.ref020] 20.Goodman R. The strengths and difficulties questionnaire: a research note. J Child Psychol Psychiatry. 1997;38(5):581–6. doi: 10.1111/j.1469-7610.1997.tb01545.x [DOI] [PubMed] [Google Scholar]

[pone.0336137.ref021] 21.Goodman R, Ford T, Simmons H, Gatward R, Meltzer H. Using the Strengths and Difficulties Questionnaire (SDQ) to screen for child psychiatric disorders in a community sample. Br J Psychiatry. 2000;177:534–9. doi: 10.1192/bjp.177.6.534 [DOI] [PubMed] [Google Scholar]

[pone.0336137.ref022] 22.A guide to absence statistics. London: Department for Education; 2019. [Google Scholar]

[pone.0336137.ref023] 23.Finning K, Ford T, Moore DA, Ukoumunne OC. Emotional disorder and absence from school: findings from the 2004 British Child and Adolescent Mental Health Survey. Eur Child Adolesc Psychiatry. 2020;29(2):187–98. doi: 10.1007/s00787-019-01342-4 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0336137.ref024] 24.Van Eck K, Johnson SR, Bettencourt A, Johnson SL. How school climate relates to chronic absence: a multi-level latent profile analysis. J Sch Psychol. 2017;61:89–102. doi: 10.1016/j.jsp.2016.10.001 [DOI] [PubMed] [Google Scholar]

[pone.0336137.ref025] 25.Hope S, Pearce A, Chittleborough C, Deighton J, Maika A, Micali N, et al. Temporal effects of maternal psychological distress on child mental health problems at ages 3, 5, 7 and 11: analysis from the UK Millennium Cohort Study. Psychol Med. 2019;49(4):664–74. doi: 10.1017/S0033291718001368 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0336137.ref026] 26.Lereya ST, Patel M, Dos Santos JPGA, Deighton J. Mental health difficulties, attainment and attendance: a cross-sectional study. Eur Child Adolesc Psychiatry. 2019;28(8):1147–52. doi: 10.1007/s00787-018-01273-6 [DOI] [PubMed] [Google Scholar]

[pone.0336137.ref027] 27.Gonzálvez C, Martín M, Vicent M, Sanmartín R. School refusal behavior and aggression in Spanish Adolescents. Front Psychol. 2021;12:669438. doi: 10.3389/fpsyg.2021.669438 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0336137.ref028] 28.Nakamura-Thomas H, Sano N, Maciver D. Determinants of school attendance in elementary school students in Japan: a structural equation model. Child Adolesc Psychiatry Ment Health. 2021;15(1):38. doi: 10.1186/s13034-021-00391-5 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0336137.ref029] 29.Heyne D, Kearney CA, Finning K. Mental health and attendance at school: Setting the scene. In: Moore DA, Finning K, Ford T, editors. Mental health and attendance at school. Cambridge: Cambridge University Press; 2022. pp. 1–21. [Google Scholar]

[pone.0336137.ref030] 30.Lindhardt L, Lindholdt L, Lund T, Mortensen OS. Self-reported mental health in adolescents attending school and its association with later school dropout: A prospective 2.5-year follow-up study. Scand J Public Health. 2022;50(8):1164–71. doi: 10.1177/14034948221089112 [DOI] [PubMed] [Google Scholar]

[pone.0336137.ref031] 31.Smith NR, Marshall L, Albakri M, Smuk M, Hagell A, Stansfeld S. Adolescent mental health difficulties and educational attainment: findings from the UK household longitudinal study. BMJ Open. 2021;11(7):e046792. doi: 10.1136/bmjopen-2020-046792 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0336137.ref032] 32.Ingul JM, Klöckner CA, Silverman WK, Nordahl HM. Adolescent school absenteeism: modelling social and individual risk factors. Child Adolesc Ment Health. 2012;17(2):93–100. doi: 10.1111/j.1475-3588.2011.00615.x [DOI] [PubMed] [Google Scholar]

[pone.0336137.ref033] 33.Children’s Commissioner for England. Education history and attendance. England: Office of the Children’s Commissioner for England; 2022. [Google Scholar]

[pone.0336137.ref034] 34.Hernan MA, Robins JM. Causal inference: What if? Boca Raton: Chapman & Hall/CRC; 2022. [Google Scholar]

[pone.0336137.ref035] 35.Cole SR, Hernán MA. Constructing inverse probability weights for marginal structural models. Am J Epidemiol. 2008;168(6):656–64. doi: 10.1093/aje/kwn164 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0336137.ref036] 36.Gilsanz P, Young JG, Glymour MM, Tchetgen Tchetgen EJ, Eng CW, Koenen KC, et al. Marginal Structural Models for Life-Course Theories and Social Epidemiology: Definitions, Sources of Bias, and Simulated Illustrations. Am J Epidemiol. 2022;191(2):349–59. doi: 10.1093/aje/kwab253 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0336137.ref037] 37.Major LE, Eyles A, Machin S. Learning loss since lockdown: variation across home nations. London: Centre for Economic Performance, London School of Economics; 2021. [Google Scholar]

PERMALINK

The impact of school absence on mental health in children and young people: Analysis of an English national birth cohort

Russell M Viner

Anna Pearce

Steven Hope

Roles

Abstract

Background

Methods

Results

Conclusions

Introduction

Methods

Fig 1. Causal diagram of hypothesized pathways between school absence (Ab) and child mental health (MH) at ages 7, 11 and 14 years.

Outcomes

Exposures

Confounding structure

Analysis

Results

Table 1. Characteristics of children in terms of school absence in preceding 2 years and mental health disorder at 7, 11 and 14 years.

Table 2. Cohort member characteristics for confounding variables.

Table 3. Odds ratios for probable psychological disorder at 7, 11 and 14 years associated with quartiles of preceding school absence in unlagged and lagged mediation models.

Table 4. Supplementary analyses: odds ratios for probable psychological disorder at 7, 11 and 14 years associated with preceding school absence as a continuous or binary variable in unlagged and lagged mediation models.

Table 5. Odds ratios for probable psychological disorder at 7, 11 and 14 years associated with preceding authorized school absence in unlagged and lagged mediation models.

Discussion

What this study adds

Limitations of this study

Policy implications and conclusions

Data Availability

Funding Statement

References

Decision Letter 0

Leonard Moulin

Roles

Author response to Decision Letter 1

Decision Letter 1

Leonard Moulin

Roles

Author response to Decision Letter 2

Decision Letter 2

Leonard Moulin

Roles

Acceptance letter

Leonard Moulin

Roles

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases