%0 Journal Article %A Gordon S Smith %A John D Langley %T Drowning surveillance: how well do E codes identify submersion fatalities %D 1998 %R 10.1136/ip.4.2.135 %J Injury Prevention %P 135-139 %V 4 %N 2 %X Objectives—The aim was to determine in New Zealand: (1) to what degree the International Classification of Diseases Supplementary Classification (ICD) external cause (E) codes for drowning identify all deaths involving drowning; (2) how the other drowning deaths are distributed across E codes; and (3) whether the proportion of drownings not identified by traditional ICD E codes has changed over time. Methods—Mortality files for the period 1977–92, which were coded in the range E800-E999 (external causes of injury and poisoning), were searched electronically using the keyword “drown”. Results—2718 cases that involved drowning were identified. This represents a 17.7% increase in the number of cases one would identify using ICD drowning E codes alone. The majority (65%) of the 408 drownings not coded as such were coded as motor vehicle traffic crashes. The number of drownings that were not identified by ICD E codes remained relatively constant over time, although the number of deaths E coded as drowning declined significantly in recent years. Conclusion—Standard ICD E codes for drowning do not identify all drowning related deaths, which may make comparisons of injury rates between countries difficult, especially for injuries such as drownings and burns that can be both nature of injury and external cause codes. Multiple cause coding and the inclusion of free text narratives are an important tool to improve the value of a country's vital statistics for injury prevention, and facilitate comparisons with other countries. %U https://injuryprevention.bmj.com/content/injuryprev/4/2/135.full.pdf