Selection bias in follow-up interviews with individuals attending the emergency department for occupational injuries

Objective To examine whether supplementary interview participation was comparable for occupationally injured patients attending two hospital emergency departments and to investigate the magnitude of selection bias in relation to sex, age, severity, job tasks and industry risk level. Methods Workers aged 18-70 years who contacted the two emergency departments for an acute occupational injury in 2013 were eligible and given a short questionnaire. Following written consent, a semi-structured interview concerning health and transient risk factors was conducted by telephone. The two departments were compared for study recruitment by age and sex. Respondents and non-respondents to the interview were compared for age, sex, injury severity, job tasks and industry risk level. Results Of 4002 patients attending the two hospitals, 1693 (42%) participated in the interview. One hospital had a markedly higher response rate to the questionnaire, but the proportions of participation in the interview were similar in the two hospitals. Patients aged <30 years were over-represented among non-respondents whereas sex, injury severity, job task and industry risk level were not significantly different between respondents and non-respondents. Conclusions Despite a relatively low interview participation rate among injured individuals attending the emergency department, selection bias was limited. This indicates that results regarding injury risk patterns may be more widely generalisable when examining the causality of occupational injuries. However, the study also showed that young injured workers were less likely to participate in follow-up interviews, which is an important factor when interpreting age-related risk of injury.