For two decades, the number of Americans who die of drug overdoses each year has steadily increased, from less than 20,000 in 1999 to more than 80,000 in 2020. Studying trends in these related deaths drug, researchers from the University of California San Diego School of Medicine, San Diego State University (SDSU), and international collaborators designed and validated a prediction model for reporting countries at risk of future epidemics of overdose deaths. The aim of the open source tool is to predict and prevent deaths through the early deployment of public health resources.
The results were published on June 9, 2021 by Lancet Public Health.
âA big challenge for public health experts is determining which regions of the country are most at risk for future overdose epidemics. If we can predict where such epidemics may occur, then we will be empowered to intervene and prevent deaths from occurring, âsaid lead author Annick Borquez, PhD, epidemiologist and assistant professor in the Division of Infectious Diseases and UC San Global Public Health. Diego School of Medicine.
The opioid epidemic has been described as a triple overlapping wave of fatal overdoses from prescription opioids, heroin, and very potent synthetic opioids, including fentanyl. Investigators used this third wave to determine if a tool could be developed to predict and prevent deaths.
“This study provides a new, rigorously validated tool to inform policy planning in the context of emerging drug overdose epidemics and sets a new standard for the development of an evidence-based response to drug use epidemics.” said lead author Charlie Marks, MPH, Graduate Research Assistant, SDSU-UC San Diego Joint Doctoral Program in Interdisciplinary Substance Abuse Research.
Using data from the CDC from 2013 (when the fentanyl epidemic began) to 2018 (the most recent data available at the time), the research team designed and trained a retrospective statistical model to find patterns in the relationship between county characteristics and overdose deaths, then used the data to predict death rates over the next year. The predictions were then compared to the actual overdose death rates in each county.
âWe found that our approach made a substantial improvement in forecasting counties with high fatal overdose rates compared to a simple benchmark that relied solely on the previous year’s rates. We also found that the increase in overdoses in a neighboring county is very predictive of future overdoses in a given county, indicating that the overdose epidemic is spreading geographically, âMarks said.
The team developed the OD Predict Explorer web tool, a publicly accessible interface where time-bound model results are presented and available for research. Users can click on a map and compare what the model predicted from 2013 to 2018 versus what was observed and determine if the tool actually identified the counties with the highest overdose across the country, including including those experiencing new peaks.
âMany of the counties with the most deaths were concentrated in the Midwest and Northeast, where the fentanyl epidemic has hit the hardest, but a growing number of western states are also affected. There is this idea that this epidemic has been concentrated in rural areas where the epidemic of prescription opioids began, but the reality is that overdose deaths have also increased in cities. For example, San Francisco and San Diego have seen a sharp increase in fentanyl overdose deaths over the past year, âBorquez added.
Borquez and Marks note that all predictive models are limited by the datasets used to inform them and said that more current and available information on deaths, drug markets and seizures, and prescription data at levels counties and states are urgently needed.
“While our approach may be effective, it also requires that fatal overdose data from all counties in the United States be accessible and available for the current year, which unfortunately is not yet standard practice.” , Borquez said. “Our model will only be useful in predicting and preventing deaths if there is no delay in obtaining data from local and national agencies.”
âRefining the model and securing access to restricted data through broad collaborations will be the next steps to improve model performance. Imagine if we could develop predictive tools for substance use epidemics, similar to what has been developed to predict COVID-19 infections and deaths. “
Borquez predicted that it would take one to two years before the tool is refined enough to make real-time predictions at the national level based on the available data, but that a promising short-term path will be to apply in states that share recent overdose deaths. information.
âWe urgently need methods to predict where these overdoses and other epidemics may occur. Epidemics such as hepatitis and HIV are also linked to drug use and can also be prevented, âBorquez said. âInvesting in drug use monitoring and harm reduction infrastructure doesn’t solve one problem, it solves several problems at once. “
Co-authors included: Daniela Abramovitz, Gabriel Carrasco-Escobar, RocÃo Carrasco-HernÃ¡ndez, Natasha Martin, Steffanie Strathdee and Davey Smith of UC San Diego; Christl A Donnelly from the University of Oxford, Daniel Ciccarone from UCSF and Arturo GonzÃ¡lez-Izquierdo from University College London.
Funding was provided by the Avenir grant from the National Drug Abuse Control Institute (DP2DA049295).