Researchers at Penn's Perelman School of Medicine have developed an overdose prediction tool for cocaine and other stimulants.
The model, which was published in an article in the journal JAMA Health Forum, is designed to assess overdose risk for patients with substance use disorder based on demographics, medical history, and other factors. The researchers created the tool in response to a long-term trend of increasing stimulant-involved overdose deaths in the United States.
Lead author and 2024 Epidemiology Ph.D. graduate Tuhina Srivastava spoke to Penn Medicine about the motivations behind the project.
“Too often, the response to people with substance use disorder is reactive or even punitive, so we believe this provides a potential step toward minimizing or eliminating that,” Srivastava told Penn Medicine. “It’s classified as a chronic disease and should be treated as such.”
According to the Centers for Disease Control and Prevention, stimulant-involved overdose deaths rose from 12,122 in 2015 to 59,725 in 2023. In Philadelphia, 70% of overdose deaths in 2023 involved stimulants.
“Although some progress [has] been made in reducing opioid-related deaths in recent years, there is still a great deal of work to be done,” co-author and 2021 Health Policy Master of Science graduate Rebecca Arden Harris told Penn Medicine.
The tool was trained on de-identified data from almost 71 million Americans enrolled in Medicaid, including people with a history of stimulant overdose and those without. It has proved accurate in tests, scoring over a nine out of 10 on a statistical accuracy model.
The researchers analyzed multiple risk factors, including occurrences of previous overdoses, higher poverty levels in a person’s surroundings, and the number of people living in a household. They considered scenarios involving cocaine-involved overdose and other stimulant-involved overdose with or without an opioid. Still, the top predictor for all stimulant overdoses was prior substance use disorder.
RELATED:
Perry World House hosts panel to discuss the future of global public health
Penn first years, Residential Services discuss Quad renovations in DP interviews
Knowledge of previous substance use disorder and other predictors “may help providers identify who might especially benefit from extra resources,” similar to preventive measures for other health conditions, such as heart attacks and diabetes, according to co-author and Medical School professor Cheryl Bettigole.
Recent trends are encouraging for overdose deaths involving stimulants and opioids alike. Preliminary data for 2024 suggest a decline in both categories nationally and in Philadelphia. The researchers hope that their tool can play a role in further lowering stimulant overdose deaths.
With the team’s model, care providers can identify those at risk of overdose early, allowing them to provide patients with resources such as cognitive behavioral therapy, naloxone — the medicine used to reverse an opioid overdose — and individualized, incentive-based recovery programs.
The researchers hope that the model’s transparency will “build trust among clinicians and public health officials.”
