Missed opportunities? An evaluation of potentially preventable poisoning deaths

Introduction: Although most poisoning deaths are not preventable with current medical technology, in some cases different management decisions may have prevented fatal outcomes. Objective: This study aims to review reported poisoning-related deaths for preventability to provide insight to improve future care. Methods: Fatality abstracts published in the US National Poison Data System (NPDS) Annual Reports (2008-2012) were analyzed. Preventability was graded using a Likert scale of 1 (definitely non-preventable) to 6 (definitely preventable). Two medical toxicologists screened all cases. Cases deemed definitely not preventable (score 1) by both reviewers were excluded from further review and considered to be "non-preventable". All cases considered at least possibly preventable by either screener were reviewed by a multidisciplinary panel of 5 physicians for preventability scoring. Differences were resolved by consensus. Cases determined to be "preventable" (scores 4-6) were characterized by type of improvement issue involved (diagnosis, treatment, monitoring, other) and recurring scenarios. Results: Of 390 published abstracts, 78 (20.0%) deaths were considered at least possibly preventable by at least one screener. Of these, 34 (8.7%) deaths were determined to be "preventable" by the panel. Inter-observer agreement by weighted kappa analysis was 0.58 for screening, 0.24 for preventability, and 0.44 for specific aspects of care. The most common toxicants were salicylates (n = 9), opioids (n = 4), toxic alcohols (n = 3), fluoride containing product (n = 3), and bupropion (n = 3). The most common improvement opportunities involved treatment and monitoring. Discussion: Most of the ingested substances in preventable deaths have delayed GI absorption or require metabolic activation to produce a delayed effect (such as salicylates, opioids, and toxic alcohols), and therefore provide an opportunity for early recognition and successful interventions. Most improvement opportunities are clearly described in the literature but may be not recognized. Conclusions: Based on an analysis of published NPDS data, a considerable number of poisoning-related deaths reaching medical attention may be preventable. The most common scenarios involved in potentially preventable poisoning fatalities related to monitoring and treatment. Salicylates and opioids were the most common agents involved in preventable deaths.