Aims To investigate the extent of variability in the reporting of heroin-related deaths in Victoria, Australia. Additionally, to identify opportunities to improve the accuracy and consistency of heroin-related death reporting by examining variability in the attribution, death certification, classification and coding of heroin-related death cases. Methods Heroin-related deaths in Victoria, Australia during a 2-year period (2012-13) were identified using the National Coronial Information System (NCIS) and used as the 'gold standard' measure in this study. Heroin-related death data from the Australian Institute of Health and Welfare (AIHW) and Australian Bureau of Statistics (ABS) were then compared. Differences in the number of deaths reported as well as the classification and coding assigned to the identified heroin-related death cases were investigated by cross-referencing these data sets and examining the assigned ICD-10 codes. Results A total of 243 heroin-related deaths were identified through the NCIS compared with 165 heroin-related deaths reported by the AIHW and assigned the heroin-specific ICD-10 code of T40.1. Forty per cent of all the missed heroin-related death cases resulted from either the attribution of the death to morphine toxicity or with non-specific drug toxicity certification; 30% occurred where the cases had been attributed to heroin but there were irregularities in death certification. Additional missed heroinrelated death cases occurred as a result of late initial registration of these deaths to the Registry of Births, Deaths and Marriages, and where these cases were then not assessed by the ABS for classification and coding purposes.