Participants, methods, and results

The six schools in Oxford with the largest number of refugee children agreed to participate. At these schools, all 115 children who were refugees or seeking asylum were identified, and we wrote to the parents or guardians of each. We asked permission, in English and the parent's native language, for the child's class teacher to complete a questionnaire on their child's psychosocial adjustment. One parent refused and the teachers of 13 children did not return questionnaires (without providing reasons) leaving 101 refugee children in the study. We individually matched refugee children for age and sex with the next ethnic minority and the next white child in the alphabetical class list. If a parent refused (n = 8) the next appropriate child on the register was recruited.

The sample comprised 303 children. In each group were 61 boys and 40 girls; 32 children were aged 5-9, 35 were aged 10-13, and 34 were aged 14-18 years. Regions of origin for the refugee children included the Balkans (48), Kashmir (16), and Afghanistan (10) and for the ethnic minority children Pakistan (64) and Bangladesh (28).

Teachers assessed the children's emotional and behavioural adjustment using the “strengths and difficulties questionnaire” (SDQ), which is well validated and widely used.3 The questionnaire has 25 items, which generate five subscales—emotional symptoms, peer problems, hyperactivity, conduct disorder, and prosocial behaviour—and a separate impact score. We did two sets of analyses comparing the refugee children to each of the control groups. Firstly, psychiatric cases were compared with non-cases using a definition of “caseness” by combining raised symptoms (SDQ≥14) and impact scores (≥2).4 Comparisons showed significant differences with 27% (95% confidence interval 19% to 36%) of refugee children, 9% (5% to 16%) of children from ethnic minorities, and 15% (9-23%) of white children meeting case criteria. Significantly more refugee children were cases than children from ethnic minorities (P < 0.01) and there was a strong trend for more refugee than white children to be cases (P = 0.059).

Secondly, we compared the groups on total SDQ scores and on each subscale (table). Refugee children scored significantly higher than ethnic minority children on both total (P < 0.01) and emotional (P < 0.001) scores; and refugee children scored higher than white children on total (P < 0.01), emotional (P < 0.01), peer (P < 0.01), and hyperactivity (P < 0.05) scores.