Socioeconomic status and severity-based incidence of poisoning: a nationwide cohort study

Abstract

Background. The association between severity-based incidence of poisoning and socioeconomic status (SES) was investigated. Methods. SES was classified based on the premium level of the National Health Insurance cohort, which are medical aid (MA) population not requiring a premium, five self-employed insured (SEI) population groups, and five employed insured (EI) population groups. The poisoning data sets were obtained from the National Injury Database (2001-2003) of Korea. Severity-based standardized incidence rates (SIRs) were calculated using an excess mortality ratio-adjusted injury severity score (EMR-ISS) similar to the new ISS. EMR-ISS was used to categorize poisoning severity as mild (1 <= EMR-ISS <= 8), moderate (10 <= EMR-ISS <= 24), severe (25 <= EMR-ISS <= 74), or critical (EMR-ISS = 75 or death). The association of household-level SES and severity-based incidence was evaluated using a multiple Poisson regression model with adjustment for the urbanization level and seasonal variations. Results. The total number of poisoning patients was 83,378. The SIR was 1,584 per 1 million person-years. Critical, severe, moderate, and mild poisoning rates were 16.3, 56.7, 18.7, and 8.3%, respectively. SIRs and relative risks of moderate to critical injuries in the SEI group and of moderate (9 <= EMR-ISS < 25) injuries in the EI group were strongly and inversely associated with the SES level when compared with MA. When we used the multiple Poisson regression model using the MA group as a reference, adjusted rate ratios were 0.97 (95% confidence interval, 0.94-1.01) for greater than severe poisoning and 0.75 (95% confidence interval, 0.69-0.82) for critical poisoning in the highest SEI. A strong inverse association was found between adjusted incidence of critical poisoning and SEI and EI levels. Conclusion. A strong association between SES levels and severity-based incidence was found in a large nationwide cohort of poisonings.