School absenteeism among children and its correlates: a predictive model for identifying absentees.

OBJECTIVES To determine the magnitude of absenteeism and its correlates and to develop a model to predict absenteeism in school children. DESIGNS A cross-sectional study. SETTING three government schools in Delhi. PARTICIPANTS 704 students, aged 10 to15 years. METHODS students were registered and interviewed using a pre designed questionnaire. The frequency and causes of school absenteeism were ascertained by school records, leave applications and one months recall. The factors were subjected to univariate analysis and a stepwise multiple logistic regression analysis and a predictive model was developed. RESULTS The average absenteeism of a student over 6 months was 14.3±10.2 days (95% CI 13.5 -15.0). 48% children absented themselves for more than two days per month on an average. The main factors associated with school absenteeism were younger age, male sex, increasing birth order, lower levels of parental education and income, school truancy, school phobia and family reasons. The discriminating ability of the predictive model developed was 92.4% CONCLUSIONS it is possible to identify potential absentees in school children.