eTable 1. Description of study outcomes and questionnaires.

Outcome variable	Definition	Comments
Study outcomes
Relative proportion of school lessons attended in the five days preceding the survey	Categorized into No (<10%), partial (≥ 10% and <90%) regular (≥ 90%) school attendance	Primary study outcome The categorization aimed to take into account the clinical relevance of school avoidance with intermittent school attendance. Due to incomplete reports from teachers because of non-contactibility at T 2 und T 3, the analyses were based entirely on information obtained from parents. The possibly lower validity of parents’ reports contrasts with a high correlation between parents’ and teachers’ reports at T 1 (r = 0.877; p<0.001; n = 64).
Psychopathology, externally rated: Child behaviour checklist 4–18 (CBCL) (20)	Age-appropriate T scores for the overall severity grade of the CBCL	Test–retest reliability between 0.70 and 0.90 (21) The results are reported by using the standardized T scores with a mean of M = 50 and a standard deviation of SD = 10
Psychopathology, self rated (from the age of 12) Symptom checklist-90-R (SCL-90-R) (22, 23)	Age-appropriate T scores for the SCL-90-R subscales Anxiety Depression	Use of the scales “Anxiety“ and “depression“ at T3 owing to the time restriction of the telephone interview; the selected scales are regarded as important markers of the severity of the psychopathology (24).
General self efficacy: GSE (25)	Calculation of sum scores	High sum scores represent pronounced self efficacy expectancy. The mean sum score is reported as M = 29.46 (SD = 5.33) (26).
Family climate: Questionnaire to capture children’s health-related quality of life (KINDL-R) (27)	Use of subscale Family climate After transformation of scale values to values from 0 to 100	Mean value of the scale “family“: M = 81.9 (95% confidence interval 81.5 to 82.3) (28) High values reflect a high quality of life.

Note: Because of space constraints we did not report additional outcomes, such as the proportion of regular school weeks over the period of the preceding 25 weeks. The exploratory analysis of the 25-week criterion does not contradict the results relating to the primary outcome