Accidental overdose admissions and premature death: A large population-based data-linkage cohort study in New South Wales, Australia (PAVLOVA 4)

Why clinicians should know about it

• Picked for Epidemiology (paper of the day, 27 October 2026): Population-linked cohort quantifies mortality after accidental overdose

Abstract

BACKGROUND AND AIMS: People with substance use disorders have reduced life-expectancy; however, risks of premature death from injuries and natural causes following an accidental overdose are not well characterised. This study aimed to estimate risk of premature death (by cause and over time), and disease burden following hospitalised accidental overdose. DESIGN: An observational study using the Poisoning And enVenomation Linkage to evaluate Outcomes and clinical Variation in Australia (PAVLOVA) cohort. SETTING: A population-based data-linkage including admissions from all public and private hospitals in New South Wales, Australia (population ~8 million), 2011-2020. PARTICIPANTS: 14 036 individuals [median age 43.5 years, interquartile range (IQR) = 29.4-61.8 years, 46.7% female, n = 6557] with an index non-fatal accidental poisoning, as defined by hospital coding, followed for a median of 3.9 years (IQR = 1.7-6.3). We restricted our analysis to people with accidental poisonings with prescription and illicit drugs that are associated with non-medical use and dependence, including opioids and benzodiazepines. MEASUREMENTS: Standardised mortality ratio (SMR) by cause (natural causes stratified by body system, and external causes stratified by intent and mechanism), SMR over time (post-index poisoning), years of life lost (YLL) by cause, age and sex. SMR and YLL were calculated based on expected values generated using official statistics. RESULTS: Of the 14 036 people included, there were 2588 deaths. Death was more likely in individuals with accidental poisoning compared with the general population [SMR = 3.5, 95% confidence interval (CI) = 3.4-3.7]. SMRs were highest for external causes (SMR = 12.8, 95% CI = 11.5-14.1), including accidental poisoning (SMR = 59.5, 95% CI = 51.7-67.3) and suicide (SMR = 10.5, 95% CI = 8.2-12.9). Death was also more likely across all natural causes, particularly infectious (SMR = 7.0, 95% CI = 5.1-8.9) and respiratory diseases (SMR = 6.2, 95% CI = 5.4-6.9). Median YLL per death was 21.2 (IQR = 10.0-35.6) and 15.2 (IQR = 7.2-30.2) for males and females, respectively. The main contributors to YLL were natural causes (64.0%): neoplasms (19.5%), circulatory (12.2%) and respiratory disease (11.1%). The remainder of YLL was from external causes (36.0%): mainly accidental poisoning (22.4%). SUMMARY: People in New South Wales, Australia, who experience an accidental overdose appear to have an increased risk of death over subsequent years, but particularly in the first month, compared with the general population; however, accidental poisoning accounts for only 22% of years of life lost, with death from natural causes being much more common.

Abstract as published, via PubMed.