School absenteeism as a risk factor for self-harm and suicidal ideation in children and adolescents: a systematic review and meta-analysis

Self-harm and suicidal ideation in children and adolescents are common and are risk factors for completed suicide. Social exclusion, which can take many forms, increases the risk of self-harm and suicidal ideation. One important marker of social exclusion in young people is school absenteeism. Whether school absenteeism is associated with these adverse outcomes, and if so to what extent, remains unclear. To determine the association between school absenteeism and both self-harm (including completed suicide) and suicidal ideation in children and adolescents, we conducted a systematic review of observational studies. We conducted meta-analysis and report a narrative synthesis where this was not possible. Meta-analysis of cross-sectional studies showed that school absenteeism was associated with an increased risk of self-harm [pooled adjusted odds ratio (aOR) 1.37, 95% confidence interval 1.20–1.57, P = 0.01] and of suicidal ideation (pooled aOR 1.20, 95% CI 1.02–1.42, P = 0.03). A small number of studies showed that school absenteeism had a longitudinal association with both adverse outcomes. Heterogeneity in the exposure and outcome variables, study design and reporting was prominent and limited the extent to which it was appropriate to pool results. School absenteeism was associated with both self-harm and suicidal ideation in young people, but this evidence was derived from a small number of cross-sectional studies. Further research into the mechanisms of this association could help to inform self-harm and suicide prevention strategies at clinical, school and population levels. Electronic supplementary material The online version of this article (10.1007/s00787-019-01327-3) contains supplementary material, which is available to authorized users.


Bailey et al 2014
Contempt of Court for Truancy was defined as a truant youth who missed school after appearing before a judge for truancy, or a truant youth who did not participate in court-ordered community service or programme as ordered by the Judge or Justice of the Peace. Severe truancy, for this study, was used interchangeably with contempt of court for truancy.
All injury-related deaths among children <18 years of age for which the incident occurred in Dallas County. Included in this study were all youth who died due to injury (homicide, suicide or unintentional injury) during the study period.

Bjarnason & Thorlindsson 1994
How often do you play truant from school? Never, <monthly, monthly, weekly, daily Dichotomous response to the question "have you ever tried to kill yourself?" Lifetime direct self-injurious behaviour (D-SIB) defined as the intentional self-inflicted damage of the surface of an individual's body by self-cutting, -burning, -hitting, -biting, and skin damage by other methods. (a) Have you ever intentionally cut your wrist, arms, or other area(s) of your body, or stuck sharp objects into your skin such as needles, pins, staples (NOT INCLUDING tattoos, ear piercing, needles used for drugs, or body piercing)? (b) Have you ever intentionally burned yourself with a cigarette, lighter, or match? (c) Have you ever intentionally carved words, pictures, designs, or other markings into your skin, or scratched yourself to the extent that scarring or bleeding occurred? (d) Have you ever intentionally prevented wounds from healing, or bit yourself to the extent that it broke skin? (e) Have you ever intentionally banged your head or punched yourself thereby causing a bruise? (f) Have you ever intentionally hurt yourself in any of the above-mentioned ways so that it led to hospitalization or injury severe enough to require medical treatment? Occasional D-SIB was defined as 1-4 reported lifetime acts of D-SIB; repetitive D-SIB was defined as ≥5 previous events of D-SIB acts during lifetime. The cut-off of ≥5 has been chosen according to the diagnostic criteria of frequency in the new proposed diagnostic entity of NSSI according to DSM-5 (American Psychiatric Association, 2013) Chen et al 2005 Number of days absent from school in the past 30 days "During the past 12 months, did you seriously consider ending your life?" "During the past 12 months, did you make a plan about how you would end your life?"; "During the past 12 months, how many times did you actually attempt suicide?"

Cheng et al 2009
'During the past 30 days, on how many days did you miss classes or school without permission? 0, 1-5, 6 or more' 'During the past 12 months, did you ever seriously consider attempting suicide? During the past 12 months, how many times did you actually attempt suicide?' Dichotomised. "I have so far never been truant a whole day" All students who did not check the item received a "positive" truancy score.
"Have you ever seriously tried to commit suicide?"

Epstein & Spirito 2009
Skipped school because unsafe (yes/no) -but no timing in table. In text described as 'felt unsafe at school or on their way to or from school on one or more of the past 30 days' Past 12 months -"did you ever seriously consider attempting suicide?" "Did you make a plan about how you would attempt suicide?". "How many times did you attempt suicide" (0 vs 1 or more)

Evren et al 2014
Absenteeism (none, 1-14 days, more than 15 days) Truancy before age 13 Self-harm behaviour (SHB) within the past year. Self-harm was defined as "deliberate self-injury to body tissue without the intent to die." The SHB question was "Do you harm yourself intentionally? (never/at least once)" . The question also included the most commonly used methods of SHB in parentheses (cutting, burning, hitting oneself, inserting sharp objects into body orifices and pulling out body hair)  et al., 1961), is "Thoughts of killing myself during the past few weeks", with response alternatives ranging from "Don't have any thoughts", "Have thoughts but would not carry them out", "Would like to kill myself", to "Would kill myself if I had the chance." The third and fourth items are from the Zung Index of Potential Suicide (Zung, 1974), and asks how much of the time the following statements apply: "I've said to someone that I wanted to kill myself" and "I've been thinking of ways to kill myself." A scale ranging from 1 to 3 was created by assigning a score of 1 to respondents who responded negatively to each item, 2 if they indicated experiencing suicidal thoughts or talked about suicide at most a little of the time (SCL-90 or Zung items) or thought of suicide but would not kill themselves (Beck item), and 3 if they reported stronger suicidal feelings, stating at least that they would like to kill themselves (Beck item) or that they have thoughts of killing themself, spoke of killing themself (Zung items), or are bothered by suicidal thoughts (SCL-90 item) at least some of the time (Cronbach's alpha .82).

Larsson and Sund 2008
Frequency of truancy within the last year (None to more than once a month) measured at T1 assessed on a 1-4 scale (good to bad) Overdose/Deliberate Self Harm (DSH) without suicidal intent. Suicide attempt. Lifetime at T1 and last 12 months at T2 one year later -latter used as outcome 'During the past 30 days, how many times did you NOT go to school because you felt you would be unsafe at school or on your way to or from school?' Participants asked whether they had ever hurt themselves to deal with stress or other problems (non suicidal self-injury). Suicide attempts excluded. Controls had never injured themselves.

Pages et al 2004
School absenteeism in last 12 months (never, sometimes, often) Suicide attempts with and without hospitalisation. (Items concerning suicide attempts ("During lifetime, how many times have you attempted suicide: no/once/twice and more") and hospitalization ("If you made a suicide attempt, were you hospitalized for this reason: no/once/twice and more").) Pillai et al 2009 Number of days absent from school in the past 3 months. 0 days, 1-3 days, 4-6 days, 7 days or more Three questions to elicit suicidal behaviour-whether seriously considered ending one's life; made a plan about ending one's life; or attempting suicide. The reference period for all three questions was the 3 months before the interview, and a positive response to at least one of the questions was considered as suicidal behaviour. Used the term suicidal behaviour to include suicidal thoughts, plans and attempts.

Peltzer & Pengpid 2017
Truancy: "During the past 30 days, on how many days did you miss classes or school without permission?" (Recoded 0 = 0 times, 1 = 1 or more times) 'During the past 12 months, did you ever seriously consider attempting suicide?' Randall et al (2014) "During past 30 days, on how many days did you miss classes or school without permission" (1-2 vs 3 or more days) "During the past 12 months, did you ever seriously consider attempting suicide?'' and ''During the past 12 months, did you make a plan about how you would attempt suicide?''. ''During the past 12 months, how many times did you actually attempt suicide?''. (0, 1 or more than one) Sharma et al 2015 "During the past 30 days how many days did you miss class without giving notice to the school?" (0-2 days or 3+ days) Suicidal ideation in past 12 months Suicide attempt in past 12 months "In the past 12 months, have you tried to end your life?"

Taliaferro & Muehlenkamp 2014
On a 5-point scale, during the last 30 days, how many days did not go to school because felt would be unsafe at school or on the way to or from school (dichotomised to 0 times/1 or more times) "Have you ever thought about killing yourself?" and "Have you ever tried to kill yourself?". Response items for both items were: "no", "yes, during the last year," or "yes, more than a year ago." To ensure data had relevance to currently suicidal youth, positive responses to "during the last year" were used to drive group classifications. Students were classified into three groups: suicidal ideation only (thought about killing oneself during the last year and never attempted suicide), suicide attempt (attempted suicide during the last year, with any reported suicidal ideation), and no suicidality (never thought about or attempted to kill oneself ever). Youth who only endorsed "more than a year ago" on both items were excluded from the analysis to avoid confounding the classification of youth into current suicide ideation and attempter groups. Wilson et al 2012 "During the past 30 days, on how many days did you miss classes or school without permission?". Response items were "0", "1-2", "3-5", "6-9" and "10 or more days". Students were considered truant if they missed more than 3 days of school within the reference period using a prior threshold by Wilson et al. [17].
"During the past 12 months, did you ever seriously consider attempting suicide?" "During the past 12 months, did you make a plan about how you would attempt suicide?"

Xin et al 2017
Truancy 'In the past 12 months, (a) Have you ever intentionally cut or scalded yourself? (b) Have you ever intentionally bit, scratched, or hit yourself?' Direct Self Injurious Behaviour was recorded if a participant answered "yes" to either item.

Funnel plots for studies included in meta-analyses
Poor school attendance and self-harm Egger's test for small-study effects: Test of H0: no small-study effects P = 0.568

Poor school attendance and suicidal ideation
Egger's test for small-study effects: Test of H0: no small-study effects P = 0.057 Note: results should be interpreted with caution given the small number of studies in each meta-analysis.

Cross-Sectional Studies
Note: A study can be awarded a maximum of one point for each numbered item with the Selection and Exposure categories. A maximum of two points can be given for Comparability.

Selection (maximum 4 stars*)
1. Representativeness of the sample a. Truly representative of the average in the target population (random sample or whole population) * b. Somewhat representative of the average in the target population (purposive sampling of representative schools or evidence that the sample is representative of the source population) * c. Selected group of users d. No description of the sampling strategy 2. Sample size a. Justified and satisfactory * b. Adequately powered to detect a difference (at least 10 events per variable in multivariate analyses)* c. Not justified 3. Non-respondents a. Comparability between respondents and non-respondents characteristics is established, and the response rate is satisfactory (>60%)* b. The response rate is unsatisfactory, or the comparability between respondents and nonrespondents is unsatisfactory c. No description of the response rate or the characteristics of the responders and nonresponders 4. Ascertainment of the exposure (absence or exclusion) a. School administrative records* b. Reported by school staff* c. Other collateral or self-report d. No description Comparability (maximum 2 stars*) 5. The subjects in different outcome groups are comparable, based on the study design or analysis. Confounding factors are controlled a. Study controls for age and gender (or analysis separated by gender)* b. Study controls for any additional factor * Outcome (maximum 2 stars*) 6. Assessment of the outcome (self-harm or suicidal ideation) a. Structured interview or written self-report* b. Clinical record* c. Reported by school or collateral d. No description 7. Statistical test a. The statistical test used to analyse the data is clearly described and appropriate, and the measurement of the association is presented as either an OR, CI and P value or a beta coefficient, SE and P value* b. The statistical test is not appropriate, not described or incomplete Additional parameter not in NOS:

Clear variables (Maximum 2)
a. Both variables clearly defined including time period for measurement (2) b. One or other variable clearly defined including time period for measurement (1) c. Neither variable clearly defined including time period for measurement (0)

*** Cohort Studies
Note: A study can be awarded a maximum of one point for each numbered item with the Selection and Exposure categories. A maximum of two points can be given for Comparability.

Selection (maximum 4 stars*)
1. Representativeness of the cohort a. Truly representative of the average in the target population (random sample or whole population)* b. Somewhat representative of the average in the target population (purposive sampling of representative schools or evidence that the sample is representative of the source population)* c. Selected group of users d. No description of the derivation of the cohort 2. Selection of the non-exposed cohort a. Drawn from the same community as the exposed cohort * b. Drawn from a different source c. No description of the derivation of the non-exposed cohort 3. Ascertainment of exposure (absence or exclusion) a. School administrative records* b. Reported by school staff* c. Other collateral or self-report d. No description 4. Demonstration that outcome of interest was accounted for or not present at start of study a. Yes * b. No

Comparability (maximum 2 stars*)
5. Comparability of cohorts on the basis of the design or analysis a. Study controls for age and gender (or analysis separated by gender) * b. Study controls for any additional factor * Outcome (maximum 3 stars*) 6. Ascertainment of outcome (self-harm or suicidal ideation) a. Structured interview or written self-report* b. Clinical record* c. Reported by school or collateral d. No description 7. Was follow-up long enough for outcomes to occur? a. Yes (>1 year) * b. No 8. Adequacy of follow-up of cohorts a. Complete follow-up -all subjects accounted for * b. Subjects lost to follow-up unlikely to introduce bias -small number lost (<20%) or attrition described and accounted for in analysis * c. Follow up rate not adequate and no description of those lost d. No statement Note: A study can be awarded a maximum of one point for each numbered item for sample size and statistical test. A maximum of two points can be given for clear variables.

Additional parameters not in NOS:
Not in NOS, taken from cross sectional NOS: 9. Sample size (Maximum 1) a. Justified and satisfactory * b. Adequately powered to detect a difference (at least 10 events per variable in multivariate analyses)* c. Not justified 10. Statistical test (Maximum 1) a. The statistical test used to analyse the data is clearly described and appropriate, and the measurement of the association is presented as either an OR, CI and P value or a beta coefficient, SE and P value* b. The statistical test is not appropriate, not described or incomplete Further additional parameter: 11. Clear variables (Maximum 2) a. Both variables clearly defined including time period for measurement (2) b. One or other variable clearly defined including time period for measurement (1) c. Neither variable clearly defined including time period for measurement (0)

*** Case-Control Studies
Note: A study can be awarded a maximum of one point for each numbered item with the Selection and Exposure categories. A maximum of two points can be given for Comparability. Additional parameters: Not in NOS, taken from cross sectional NOS: 9. Sample size (Maximum 1 star) a. Justified and satisfactory * b. Adequately powered to detect a difference (at least 10 events per variable in multivariate analyses)* c. Not justified 10. Statistical test (Maximum 1 star) a. The statistical test used to analyse the data is clearly described and appropriate, and the measurement of the association is presented as either an OR, CI and P value or a beta coefficient, SE and P value* b. The statistical test is not appropriate, not described or incomplete Further additional parameter:

Selection (maximum 4 stars*)
11. Clear variables (Maximum 2 stars) a. Both variables clearly defined including time period for measurement (2) b. One or other variable clearly defined including time period for measurement (1) c. Neither variable clearly defined including time period for measurement (0)

Results of quality assessment Quality Assessment: Poor attendance and suicidal ideation
Cohort studies Reference Representativeness of exposed cohort (0-1) Selection of non-exposed cohort (0-1) Unselected birth cohort One cohort containing exposed and non-exposed Random sample of schools. 90,000 of 120,000 completed survey -core random sample selected for interviews (N 13000 is the core sample)

Ascertainment
One cohort containing exposed and nonexposed Unselected birth cohort One cohort containing exposed and nonexposed Stratified random sampling of schools from 2 counties, 88% response rate One cohort containing exposed and nonexposed groups

Selfreported
Lifetime at T1 and last 12 months at T2 Attendance not in the model but many covariates weregender and age were considered univariately but age not in the model (probably due to non-significance)attendance not in the model Randomly selected from representative high schools, only minor differences compared to census data One cohort containing exposed and nonexposed groups