Analysis of agreement between coding of underlying cause of death by traffic accidents

Abstract

An evaluation of information researched as to basic underlying cause of death by traffic accidents and a comparison with codings attributed on the basis of death certificates. The official coding of underlying cause of death on Death Certificates was compared with coding based on primary data obtained from five hospitals and accident reports. Kappa statistics with 95% CI were used to assess the agreement between the two coding systems. The research covered 1,719 patients. 57 of whom died. Of these 57 deaths, the official coding for 50 was obtained. Data showed an underreporting of deaths by traffic accidents, since 32% of the 50 deaths were coded as non-specific accidents (E928.9) There were also 38% of deaths coded as non-specific traffic accidents (E 819.9). Using primary data, non-specific traffic accidents dropped to 4%, accidents to pedestrians (E814.7) being responsible for 48% of deaths. The Kappa coefficient (0.124), with 95% confidence interval (-.1533-0.4022) was calculated to assess the inter-rater reliability between the two codings, which was considered poor. It is concluded that coroners, who perform autopsies on casualties of traffic accidents, should dedicabe greater effort to filling out Death Certificates correctly.