This study focused on two key traumatic exposures and their associations with absenteeism in a sample of Sri Lankan school children aged 12–17 years. The findings from this study are important because these types of exposure have received relatively little attention with respect to their later impact on children’s education despite the importance that this has in determining later social mobility and socioeconomic status. Key findings were that overall previous tsunami-related exposure was significantly associated with absenteeism, overall conflict-related exposure was not significantly associated with absenteeism, although some associations were found with individual conflict-related exposures in secondary analyses. Mental disorders were associated with absenteeism but did not appear to account substantially for the association between tsunami exposure and this outcome.
There is no consensus definition of significant absenteeism to our knowledge, and it varies according to the cut-off points applied by different authors [11, 12, 37, 38]. Although numerical definition of absenteeism varies, the broad concept of absenteeism has been consistent, defined as the total number of absent days from school from the total number of available school days of a calendar year [11, 37, 38]. In our study, absenteeism, defined as non-attendance on at least 20% of school days, as recorded by the school, was present in 26.8% of the sample. This is generally consistent with the 20% absenteeism rate found in other studies in Sri Lanka and elsewhere [11, 37–39].
This study was conducted in mid 2007, approximately two and half years after the 2004 December tsunami. In contrast, the civil conflict had been ongoing for almost three decades at the time of the survey. These timeframes have to be considered in interpreting the results for the two exposures of interest. A previous study on exposures to tsunami and conflict were both shown to be associated with mental ill health among Sri Lankan children . Another study found a range of tsunami-associated psychosocial events such as loss of parents, prolonged displacement, losses of other family or friends to be associated with psychiatric symptomatology including depression and PTSD among Sri Lankan children . However, our findings did not show any strong association between exposure to tsunami and mental disorders. Mental disorder therefore did not appear to be an important factor in our analyses accounting for the tsunami-absenteeism association. However, the impact of tsunami exposure on absenteeism may be due to the more disruptive, sudden impact of the exposure on social structures in the Sri Lankan context, and it is noteworthy that bereavement was the strongest individual tsunami-associated risk factor for absenteeism. Although exposure to conflict may be devastating, families tend to group together even in displacement, which might have limited the strength of association below significance for this overall exposure. The tsunami on the other hand may have disrupted whole social support systems, breaking families, communities and destroying alternative support systems such as schools within a very short span of time. In exploratory analyses, displacement and loss of property due to conflict were associated with absenteeism, which may support this suggestion of social disruption as an important underlying factor. In interpreting the findings from this study, attention must be paid to the fact that absenteeism may be linked more to the loss of family and its continuing effects (possibly exacerbated by more widespread social disruption), rather than to a direct link with the disaster in question.
A study carried out in eastern Sri Lanka found that daily stressors partially mediated the relation between exposure to conflict or disaster and psychological outcomes among adolescents . The association with absenteeism in our sample remained relatively independent of most covariates, with paternal/maternal education level (proxy measure of socio-economic status) appearing to be the only major confounding factor. Furthermore, no significant associations were found between exposure to conflict and mental disorders in this study. Exposure to conflict can bring devastating changes to lives of children, especially if they are subjected to mass displacement and separations, leading to disruptions in their daily lives, most importantly to their education [2, 9]. In the Sri Lankan setting, several districts included in the study bordered the conflict areas and people were subjected to displacement because of conflict-related violence spilling over, prior to and at the time of survey. This may be reflected in absenteeism of children in these populations [25, 27] and the significant associations with conflict-related family injury, displacement and property loss in our sample.
Conflict and natural disaster related events can be highly traumatic to children with both short and long term consequences, affecting their physical and mental performance, and education is a key facet of their lives with a high vulnerability to disturbance, especially as education is perceived as the gateway for escaping poverty and attaining upward social mobility in most resource-poor settings in the world. The Strengths and Difficulties Questionnaire used in the study gave rise to an overall prevalence of 8.5% for any psychiatric disorder among the population of children in the sample. This figure is similar to findings on behavioral disorders (using similar outcome measures) among children in Sri Lanka and other countries . Conduct disorders had the highest prevalence in the studied sample. Associations were observed for SDQ-derived conduct and emotional disorder with absenteeism. These findings add to existing evidence that mental illness such as behavioral disorders in children have a detrimental effect in their education, inducing and increasing absenteeism among them [11, 16, 17]. Regression analysis also showed a significant association between absenteeism and any psychiatric disorder that lies within the predictive range of the SDQ. This includes conduct, emotional and hyperactive disorders in a broader aspect. This evidence again suggests that psychopathology may play a key role in children’s education.
This study has certain limitations. The results may be subject to type II statistical error due to lack of power-for example, the relatively small numbers with specific conflict- or tsunami-related exposures which precluded detailed analyses of individual exposures. Type I errors are unlikely because of the primary focus on two exposures and one outcome; however, secondary associations need to be treated with more caution. The main likely biases are selection bias stemming from under-sampling of absentee students and information bias due to the data collection strategy used in the study. The fact that sampling was not possible in the districts most strongly and recently affected by conflict may have diluted the influence of this exposure, as well as limiting national generalisability. Under-sampling of absentee students, which may have lead to an increase of students without absenteeism being recruited to the study has been potentially minimized by conducting three visits to track selected participants in case they were absent while the overall participation rate was high at 92.5%. Another issue is that the analysis is limited to the secondary age students (12–17 years) and the findings do not reflect mental disorder prevalence of the younger age school children nor their school attendance patterns. Paternal education level was the only possible confounder shown in the regression analysis, although illnesses, stress among family members and daily stressors (not tested in the study) may act as confounders. As this was a cross-sectional study, causality underlying associations cannot be assumed.