Predictive Analytics Model Forecasts Drug Overdose Deaths

June 14, 2021 - Researchers developed a statistical predictive analytics model that studies patterns in drug-related fatality data and can identify which counties are at high risk for future fatal overdoses, according to a study published in Lancet Public Health.

The model was developed by researchers at San Diego State University (SDSU), University of California San Diego School of Medicine, and other collaborators. They used Centers for Disease Control and Prevention (CDC) data collected between 2013 and 2018 to develop and train the model and found that it could forecast fatality rates for the next year by county.

"This study provides a novel, rigorously validated tool to inform policy planning in the context of overdose epidemics driven by emerging drugs and sets a new standard for the development of a data-driven response to drug use epidemics," said Charlie Marks, MPH, one of the study’s authors and a graduate research assistant at SDSU-UC San Diego’s joint doctoral program in Interdisciplinary Research in Substance Abuse, in a press release.

Drug overdose deaths have quadrupled since 1999, according to the CDC. Over 70 percent of 2019 drug overdose deaths involved opioids. Illegally made fentanyl, a synthetic opioid, set off the third wave of the opioid epidemic in 2013. The CDC says that fentanyl is 50 to 100 times more potent than morphine, and since 2013 there has been a significant spike in overdose deaths in the US.

Researchers used county-level data to assess drug markets, demographic considerations, healthcare access, and the geographical spread of overdose as predictors, the study stated. The predictive analytics model outperformed the benchmark model, successfully predicting overdoses and ranking counties by risk rate.

READ MORE: NIH to Leverage Data Management for Opioid Crisis Research

"A big challenge for public health experts is figuring out which parts of the country are at greatest risk of future overdose outbreaks,” said Annick Borquez, PhD, one of the study’s author and an epidemiologist and assistant professor at UC San Diego School of Medicine, in the press release. “If we can predict where such outbreaks may happen, then we will be empowered to intervene and stop deaths from occurring.”

The CDC does not report county-level data regarding the number of overdose deaths to the public if the county’s total was less than 10 to protect the privacy of effected individuals. However, the study’s authors were given access to the CDC’s full database with specific county rates to inform the prediction model.

In the final model, researchers analyzed everything from high school graduation and unemployment rates to opioid prescription rates and urgent care facility access. The model was designed to identify patterns between the county characteristics and death rates.

"We found that our approach brought substantial improvement to predicting counties with high fatal overdose rates compared to a simple benchmark that relied on past year rates alone,” Marks continued.

“We also found that increased overdoses in a neighboring county are very predictive of future overdoses in a given county, indicating that the overdose epidemic spreads geographically.”

READ MORE: Predictive Analytics Model Anticipates Pediatric Blood Clots

The researchers also developed an interactive public web application called ODPredict Explorer where users can explore the model’s effectiveness. Of course, the predictive model is only as good as the data used to train it, researchers noted. They called for more up-to-date and readily available drug overdose fatality data, as the most recent data was released in 2018.

"While our approach can be effective, it also requires that fatal overdose data from all the counties in the U.S. be accessible and available for the current year, which unfortunately is not yet standard practice," Borquez continued in the press release. "Our model will only be useful in predicting and preventing deaths if there is no lag in getting data from local and national agencies."

"Further model refinement and securing access to restricted data through broad collaborations will be next steps to improve model performance. Imagine if we could develop predictive tools for substance use epidemics, similar to what was developed to predict COVID-19 infections and deaths."

Borquez estimated that the tool will need a few more years to be able to make real-time predictions, but it shows significant promise for predicting overdose fatalities on a national level in the near future.

The opioid epidemic continues to cause devastation across the country. But new developments provide hope for curbing overdose rates. CareSource recently announced a partnership that provides a medication disposal powder to neutralize active ingredients in unused medications, in an attempt to avoid opioid misuse.

However, COVID-19 has caused access issues to naloxone, a new study revealed. Naloxone prescription rates have declined significantly over the past year, potentially meaning that those who need it most are having trouble gaining access to critical prescriptions and care.