Determinants of response in a longitudinal health study following the firework-disaster in Enschede, The Netherlands

Very few longitudinal health studies after disasters published data on the determinants of loss to follow up. However, these determinants provide important information for future disaster studies to improve their response and reduce selection bias. For this purpose we analyzed the data of a longitudinal health survey which was performed among residents and emergency workers, at 3 weeks (n = 3662) and at 18 months (n = 2769) after a major firework disaster in The Netherlands (Enschede, May 13, 2000). The response was lower among immigrants (54%) than among native Dutch (81%). Severe damage to the house due to the disaster (OR: 1.8; 95% CI: 1.1-3.0) and being involved as an emergency workers (OR: 2.1; 95% CI: 1.2-3.4) were associated with higher response among native Dutch, while this was not the case among immigrants. Non-western immigrants with health problems in the first study were more likely to participate in the second study (for example physical symptoms OR: 2.5: 95% CI: 1.4-4.4), while the native Dutch with these symptoms were less likely to participate (OR: 0.7; 95% CI: 0.5-0.9). In conclusion, disaster-related characteristics were associated with higher response in native Dutch. Health problems were associated with higher response among non-western immigrants and with lower response among the native Dutch.