OBJECTIVE: To identify host and environment characteristics associated with poison exposure among preschool children in Thailand. SETTING: A matched case-control study in 20 public hospitals in Khon Kaen province. METHODS: Cases comprised 100 consecutive preschool children attending hospital between September 1997 and February 1999 because of reported exposure to a poison. Controls were three children matched by age, sex and area of residence to each case, who had never attended a hospital for this reason. Risk factors were elicited by interviewing the children's caregivers at their homes using a structured questionnaire and analyzed using conditional logistic regression. RESULTS: Number of siblings, whether the child was "medicine eating" and "rushing to explore new objects", the child's activities while the caregiver was working during the daytime, the child's distance from the caregiver while the caregiver was working, whether the child was taken to a non-agricultural worksite, and the number of used containers around the residence, were all statistically significant in univariate analyses. However, when all significant variables were included in a multivariate model, only "medicine eating" and the number of used containers around the residence remained statistically significant. CONCLUSION: Attention to "medicine eating" and used containers through increased supervision could be a protective factor against poison exposure for these children. To reduce risk, caregivers should not refer to medicines as foods and used containers should be systematically collected for disposal or recycling.
OBJECTIVE: To identify host and environment characteristics associated with poison exposure among preschool children in Thailand. SETTING: A matched case-control study in 20 public hospitals in Khon Kaen province. METHODS: Cases comprised 100 consecutive preschool children attending hospital between September 1997 and February 1999 because of reported exposure to a poison. Controls were three children matched by age, sex and area of residence to each case, who had never attended a hospital for this reason. Risk factors were elicited by interviewing the children's caregivers at their homes using a structured questionnaire and analyzed using conditional logistic regression. RESULTS: Number of siblings, whether the child was "medicine eating" and "rushing to explore new objects", the child's activities while the caregiver was working during the daytime, the child's distance from the caregiver while the caregiver was working, whether the child was taken to a non-agricultural worksite, and the number of used containers around the residence, were all statistically significant in univariate analyses. However, when all significant variables were included in a multivariate model, only "medicine eating" and the number of used containers around the residence remained statistically significant. CONCLUSION: Attention to "medicine eating" and used containers through increased supervision could be a protective factor against poison exposure for these children. To reduce risk, caregivers should not refer to medicines as foods and used containers should be systematically collected for disposal or recycling.