Enhancing identification of suspected unregulated drug toxicity deaths in administrative health records: a case study among people with HIV in British Columbia, Canada.

backgroundAccurate and timely cause of death information is important for public health monitoring. In recent years, deaths with an unknown/ill-defined underlying cause have increased markedly in Canada, particularly in British Columbia (BC), among people living with HIV (PWH) in concert with the 2016 unregulated drug toxicity crisis. Using a case study of PWH in BC, who are disproportionately affected, we examined all deaths occurring between 2015 – 2020 and adapted a published method to classify unknown-cause deaths as likely drug-related.

methodsFrom the STOP HIV/AIDS cohort (a linkage of administrative healthcare and clinical data for PWH in BC), we categorized all deaths in the cohort between 2015 – 2020 into three cause of death groupings: unknown-cause (ICD-10 coded “R99”), drug- related (X41-X42; X44; Y11-Y12; Y14), and non-drug related causes (all other deaths). We described and compared decedents’ demographics, residence and clinical characteristics including lifetime drug use history and drug-related healthcare contact occurring 12 and 24 months preceding death. We estimated the percentage of unknown-cause deaths that met the following criteria for likely drug-related death classification: a) aged 20–64 at death, b) a “pending”, “accident”, or “undetermined” manner of death; and c) a history of unregulated drug use (drug-related healthcare contact or clinical data).

resultsOf 849 deaths among PWH in BC between 2015 – 2020, 79 were unknown-cause (9%), 155 drug-related (18%), and 615 non-drug-related (72%). The unknown-cause deaths closely resembled confirmed drug-related deaths regarding several sociodemographic characteristics (sex, age, residence, location of death) and drug-related healthcare contact in the 12 months preceding death (75% of unknown-cause and 72% of drug-related deaths, respectively). Applying the classification criteria, 86% of unknown-cause deaths were classified as likely drug-related, increasing the estimated number of drug-related deaths from 155 to 223 (by 44%).

conclusionsOur findings indicate many deaths of unknown cause among PWH in BC between 2015 – 2020 were likely drug-related, suggesting an underestimation of the true burden of the unregulated drug toxicity crisis in this population based on standard mortality records. This approach may be further examined in non-PWH cohorts and validated using updated mortality data.