Estimating the Number of Opioid Overdoses in British Columbia Using Relational Evidence with Tree Structure Publication type: PREPRINT
Mallory J Flynn, Paul Gustafson, Michael A. Irvine, M.J. Flynn, P. Gustafson, M.A. Irvine · 2025
Read Estimating the Number of Opioid Overdoses in British Columbia Using Relational Evidence with Tree Structure Publication type: PREPRINT by Mallory J Flynn, Paul Gustafson, Michael A. Irvine, M.J. Flynn, P. Gustafson, M.A. Irvine (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2506-21024.txt.
Introduction
Mallory Flynn, Paul Gustafson, and Michael Irvine tackle a hidden count problem in a 2025 preprint. They want to know how many opioid overdoses took place in British Columbia in a set period. Many cases are hard to count, yet the total guides health funds and policy choices.
Care data in the province form a tree, with each path showing steps a person took after an overdose. The authors compare two ways to work back from those paths to the full total. One blends path based estimates with weights, while the other builds a full model that states doubt at each level.
Catalog record
| CREATOR | Mallory J Flynn, Paul Gustafson, Michael A. Irvine, M.J. Flynn, P. Gustafson, M.A. Irvine |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 46 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2506-21024
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Estimating the Number of Opioid Overdoses in British Columbia Using Relational Evidence with Tree Structure Publication type: PREPRINT Mallory J Flynn, Paul Gustafson, Michael A. Irvine, M.J. Flynn, P. Gustafson, M.A. Irvine Abstract In many fields, populations of interest are hidden from data for a variety of reasons, though their magnitude remains important in determining resource allocation and appropriate policy. In public health and epidemiology, linkages or relationships between sources of data may exist due to intake structure of care providers, referrals, or other related health programming. These relationships often admit a tree structure, with the target population represented by the root, and paths from root-to-leaf representing pathways of care after a health event. In the Canadian province of British Columbia (BC), significant efforts have been made in creating an opioid overdose cohort, a tree-like linked data structure which tracks the movement of individuals along pathways of care after an overdose. In this application, the root node represents the target population, the total number of overdose events occurring in BC during the specified time period. We compare and contrast two methods of estimating the target population size - a weighted multiplier method based on back-calculating estimates from a number of paths and combining these estimates via a variance-minimizing weighted mean, and a fully Bayesian hierarchical model. keywords Bayesian modeling; multiplier method; population estimation; tree; overdose; opioid † † copyright-year: 2025 † † corresponding: Mallory J Flynn, † † citation: ), , , . † † copyright: , and , and ( † † year: 2025 † † ctitle: Estimating the Number of Opioid Overdoses in British Columbia Using Relational Evidence with Tree Structure † † cjournal: Preprint † † cvolume: 2025;01:1–21 Mallory J Flynn et al \presentaddress Department of Statistics, Faculty of Science, University of British Columbia, 2207 Main Mall, Vancouver, BC, V6T 1Z4, Canada 0 0 footnotetext: Abbreviations: MM, multiplier method; WMM, weighted multiplier method; ATD, acute toxicity death 1 Introduction The drug toxicity public health emergency continues to be one of the most significant public health challenges of the 21st century, with more than 44,000 Canadians dying from opioid overdose between January 2016 and December 2023, with rapid increases in drug overdoses and deaths throughout North America ???? . While in many jurisdictions increased numbers of drug toxicity deaths were driven by prescription opioids ? , in British Columbia a major contributor to the increasing rate of drug toxicity deaths was the introduction of fentanyl and fentanyl analogues to the illicit drug supply ? . During the COVID-19 pandemic, opioid-related overdoses again increased sharply throughout, with an average of 22 overdose deaths per day observed between January and December 2023, compared with eight deaths per day in 2016 and 12 deaths per day, nationally, in 2018 ? . Approximately 89% of these deaths occurred in BC, Alberta, and Ontario alone, and a similar percentage of accidental, apparent overdose toxicity deaths are occurred among individuals between the ages of 20 and 59 years ? - significantly lower than the 82 year average life expectancy of Canadians ? . In British Columbia, there has been a large increase in the number of deaths due to opioid-related overdose ?? and a corresponding yearly drop in life expectancy between 2014 and 2018 attributed to these deaths ?? . Although research estimates suggest prevention methods have aided in averting deaths due to opioid toxicity ?? , the total numbers of overdose events remain unclear, and in British Columbia, the number of attributed to opiates, in particular, is conjectured to be dramatically under-represented by healthcare data alone ? . Overdose events are not fully captured in health administrative data sources, due in part to bystander administration of naloxone, persons not seeking medical care, and miscoded data. As such, the true number of overdoses is under-reported, and it is probable that there is a much higher burden of overdose when non-fatal overdose events are considered alongside mortality data ?? . Synthesis of available data sources, including existing surveillance data and estimates from the literature, is required to produce indirect statistical estimates of the total number of overdoses ? . A comprehensive and scalable method to estimate the overall number of opioid overdose events in the population is instrumental in improving the health of communities by providing important information to decision makers and clinicians to inform policies and resource allocation, programs, intervention strategies, and clinical care. The Public Health Agency of Canada (PHAC) and the BC Centre for Disease Control (BCCDC) have created a dataset that is very well suited for addressing this issue ? , which links individuals existing in multiple health datasets (including ambulance data, emergency room data, hospital admissions, coroners data, and others) and tracks their progression through a tree of possible treatment pathways after an overdose event (see Figure 1 ). It is important to note that while all of the nodes in Figure 1 emanating from the lower child of the root node contain counts, many of these counts are simply summations of the observed counts of descendents (e.g., node B is a summation of leaf descendents), with parent nodes themselves unobserved. In addition, measured nodes are only partially observed; estimates of the proportion of individuals missing have been determined through preliminary investigations of the missingness expected in these sources of data. The traditional multiplier method is a popular means of target population size estimation in public health and epidemiology, as it is applicable to existing data such as health administrative data. When knowledge about multiple sub-populations of the target population is available, the weighted multiplier method (WMM) is an extension of the traditional methodology that allows for multiple sources of evidence to be utilized and synthesized into a single estimate of the target population, which is represented in the root of the tree that is constructed by the combined evidence ? . The pathways of care elucidated by the Provincial Overdose Cohort (POC) data regarding the overdose crisis in British Columbia presents an ideal application for the WMM due to the clear creation of a tree structure with multiple known sub-populations which are mutually exclusive. It represents pathways of care after an overdose event; while one major arm of the tree is well-informed using healthcare administrative datasets, the other major arm is largely unobserved, as these nodes represent those individuals that did not access healthcare services after an overdose event. Here, we apply both the WMM and a Bayesian hierarchical model to this data structure in order to estimate the total number of overdose events over the given time period, and compare and contrast the estimates provided by each method. In a companion paper, we describe the underlying theory behind both the multiplier-based methodology and the hierarchical Bayesian model ? . In another companion paper, we describe software packages developed to ease implementation of the novel multiplier-based methodology, as well as a hierarchical Bayesian model adaptable to any general tree structure ? . The POC data are owned by the BC Provincial Health Authorities and the Ministry of Health and stewarded by the BCCDC ????? . Access to data provided by the Data Steward(s) is subject to approval, but can be requested for research projects through the Data Steward(s) or