Concordance of ICD-10 Y90 and Y91 codes for alcohol intoxication in the Americas

Codes for the Y90 (breathalyzer readings) and Y91 (clinical assessment) were added to the ICD-10 “external causes' chapter as categories to be used as supplementary information considering the causes of morbidity and mortality, with commentary implying the Y91 be used as a backup when the Y90 could not be coded (suggesting interchangeability). With preparatory work for the ICD-11 underway, one question to be answered is whether the two codes are accurately capturing the same dimension. This presentation extends earlier work that compared the Y91 assessment with Y90 breathalyzer using emergency room (ER) admissions data from 12 countries comprising the WHO Collaborative Study on Alcohol and Injuries. That study found the level of agreement between the two measures to be moderately concordant, with higher agreement among current drinkers not consuming substantial amounts of alcohol anytime prior to the ER (BAC<.06) but much lower agreement among those who reported drinking within the six hours prior to injury. Using alternative partitioning methods, we examined optimal ways of recording estimated blood alcohol concentration to test its concordance with the Y91 assessment using the Kendall's Tau-B statistic. As well, ten individual signs of intoxication (with four levels of severity) designed to help clinician's make an accurate Y91 assessment were used to further explain the concordance results between the Y90 and Y91 ratings. Socio-cultural contextual variables were considered in analyses. Findings showed concordance to be similar to the prior study, still with some indication of lack of agreement between the individual signs and the Y91 assessment.