An empty hospital hallway with a stretcher

Photo: Shutterstock

Can Better Data Predict Opioid Overdoses and Slow Infectious Disease Rates?

Tufts University School of Medicine teams and collaborators are running multiple projects that seek to reduce overdoses and the spread of infections, such as HIV and hepatitis C, in people who use drugs

In much the same way that meteorologists have become more sophisticated in forecasting the weather, researchers at Tufts University School of Medicine, led by professor Thomas J. Stopka in the Department of Public Health and Community Medicine, are using their research to better predict spikes in opioid-related overdoses and address HIV and hepatitis C infections.

Opioid-related overdoses continue to be at crisis levels in communities across the United States, with more than 75,673 fatal overdoses in the 12-month period ending in April 2021. Stopka and his colleagues believe that using the predictive power of their research to study the patterns of overdoses and heightened infections in those using drugs can save lives.

Their goal is to provide timely alerts to people who have substance use disorders, their families, public health, public safety, and social services agencies in hopes of blunting the number of deaths and serious illnesses that result when a new drug hits and disrupts the illicit drug supply.

In a situation where every hour counts, research by Stopka and co-principal investigators Shikhar Shrestha and Jennifer Pustz, N19, MG19, is helping the City of Lowell, Massachusetts, cut in half the time it takes to identify an uptick in opioid overdoses and to alert those in danger more effectively. Shrestha is an assistant professor and Pustz was a researcher in the Department of Public Health and Community Medicine when the study began. New predictive model research may make it possible to identify risks of overdoses at the level of individual ZIP codes in the future, Stopka adds.

Professor Thomas J. Stopka in the Department of Public Health and Community Medicine at Tufts University School of Medicine

Professor Thomas J. Stopka in the Department of Public Health and Community Medicine at Tufts University School of Medicine. Jake Belcher for Tufts University

Another Tufts study is identifying ways to decrease the risk of overdose among people recently released from jail or prison. And yet another effort, along the Interstate-91 corridor in New Hampshire and Vermont, is developing new interventions that the researchers hope will slow the spread of HIV and the hepatitis C virus among people who inject drugs.

Earlier Alerts Slowing Overdoses in Lowell

Drug overdoses increase when a new adulterant or cut is introduced into substances already being sold on the streets. The greatest risk occurs when fentanyl is in the mix. Fentanyl was first introduced in Eastern Massachusetts around 2013. Then it moved to southeastern Massachusetts and arrived later in western Massachusetts.

Stopka and colleagues saw an uptick in overdoses in Lowell, Massachusetts, and the Merrimack Valley when fentanyl arrived, replacing heroin as the drug of choice for many people with substance use disorders. More recently, xylazine, an animal tranquilizer also called Tranq, has been in the news as it makes its way into the drug supply. Mixed with fentanyl, it can put people to sleep for hours, lessening the frequency of their need. (Often, they don’t seek out drugs just for a high, but rather to not feel sick from withdrawal). But it can result in fatal drug poisonings and naloxone (Narcan®) may not reverse its effects.

Shrestha, Pustz, and Stopka, along with colleagues from the University of Massachusetts Lowell, and city agencies and non-profits, are focused on better understanding opioid alert programs. These programs are used by police departments, fire departments, public health officials and others who seek to warn people with substance use disorders when trouble is brewing in the drug supply.

This chart from the Lowell opioid overdose monitoring system shows spikes of opioid overdoses dating back to 2011.

This chart from the Lowell opioid overdose monitoring system shows spikes of opioid overdoses dating back to 2011. Courtesy of Thomas Stopka

In one such effort, the researchers observed that when 18 overdoses occurred in Lowell over a period of 72 hours, an alert system would be triggered that would include a press release to the news media, postings on social media, and other traditional routes to communicate news to the public. “But the messages weren’t getting to the people using drugs, their families, and peers soon enough,” Stopka says.

Using EMS data from Trinity EMS, the major ambulance service in Lowell, the team is developing and testing predictive models that could signal a problem in as quick as 36 hours that there has been a spike in overdose deaths. The models aim to help Lowell officials get the alerts out faster—through traditional media, social media, and other channels—to reach people who are at risk of overdose. They also aim to reach organizations working with people who use drugs and peer groups, including syringe services programs, methadone maintenance programs, soup kitchens, and shelters.

Shikhar Shrestha, assistant professor in the Department of Public Health and Community Medicine at Tufts University School of Medicine

Shikhar Shrestha, assistant professor the Department of Public Health and Community Medicine. Photo: Anna Miller / Tufts University

“The goal of the study is to not just reduce the time it takes to detect a spike in opioid overdose, but also to understand how alert messages are communicated among stakeholders and people with substance use disorders and how they act on the shared information,” says Shrestha. “The research will help us develop and share a streamlined protocol that reduces the time to detect overdose spikes and then share the messages efficiently in the affected community.” 

Predicting Overdoses by ZIP Code

Stopka and fellow principal investigator Cici Bauer, associate professor at the University of Texas Health Sciences Center in Houston, along with Olaf Dammann, professor and vice chair of the Department of Public Health and Community Medicine at the School of Medicine, are working to develop spatiotemporal prediction models relying on vast amounts of data to forecast fatal opioid-related overdoses at a granular level—as detailed, perhaps, as particular ZIP codes. This work is also being done with researchers at Boston University, the University of Massachusetts Lowell, the Massachusetts Department of Public Health, as well as a community advisory board comprised of experts from state and local public health departments, healthcare facilities, community-based agencies, and harm reduction programs.

“Current forecasting models focus on predicting overdoses on a large geographical scale, such as states or counties,” Stopka says. “They lack the granularity that could really help local public health officials guide decisions and resources.”

“In the same way that meteorologists try to predict where hurricanes and tornadoes are going to hit, we want to develop predictive models so people can make themselves safe by responding before the big hurricane of overdoses hits,” he concludes. The goal is to be able to inform public health responses that could flood an area experiencing an overdose spike with safe syringes, Narcan, and medications for opioid use disorder to help people begin treatment.

The Jail-to-Community Transition

Another effort by researchers at the School of Medicine is focused on improving the transition from jail or prison to the community for incarcerated individuals with opioid use disorders. People released from jail or prison are 120 times more likely to overdose on opioids than the general population in Massachusetts. A recent study by Stopka and fellow principal investigators from the University of Massachusetts—Elizabeth Evans, professor at UMass Amherst, and Peter Friedmann, professor at UMass Baystate—suggests at least three ways a new, jail-based opioid use disorder treatment program in seven county jails in Massachusetts can be improved to reduce that risk.

“If someone starts treatment with methadone or another opioid disorder treatment while in jail, and then is released without proper planning, they may relapse before getting continued treatment back in the community,” Stopka explains. “And if they attempt to use opioids at levels comparable to what they previously used, at a time when their tolerance is much lower, overdoses often result.”

But giving incarcerated individuals bridge doses of medication to cover them until they can connect with a community treatment program—and making sure they have a cellphone to contact that program to make an appointment—are key to preventing overdoses during the transition.

“The justice system in Massachusetts is at the forefront of public health innovation,” says Evans. “And we’re learning whether a program like this is effective, how it can be implemented, and what it costs, which is information that can help legislators know whether to continue the program or replicate it elsewhere in other states.”

From left: Thomas Stopka, Jon Kelley of Trinity EMS, Shikhar Shrestha, and Jennifer Pustz at the Owl Diner in Lowell, Massachusetts.

From left: Thomas Stopka, Jon Kelley of Trinity EMS, Shikhar Shrestha, and Jennifer Pustz at the Owl Diner in Lowell, Massachusetts. Photo: Courtesy of Thomas Stopka

Using a Van to Slow HIV and Hepatitis C

Stopka and Friedmann are also multi-principal investigators for one of eight groups nationally participating in a program funded by the National Institutes of Health to see if improved access to diagnostic and treatment services can slow infectious diseases associated with drug use in rural communities. In this study, the Tufts and UMass Chan-Baystate team is working along the I-91 corridor in Western New Hampshire and Vermont to test a mobile telemedicine-based hepatitis C virus treatment to assess its effectiveness among rural populations of people with substance use disorders.

“While some people may think of the opioid epidemic as a predominantly urban problem, the hepatitis C and overdose rates are also elevated in rural areas when accounting for population density. The cause is a lack of access to treatment and prevention services,” says Stopka.

“What we’ve learned over the years is that in both urban and rural areas, drugs are more accessible and less expensive than drug treatment. And when people are ready to get into drug treatment programs, supply of treatment slots rarely outpace demand. Hepatitis C virus treatment and overdose prevention services and drug treatment are less available and are further away in rural areas.”

In the I-91 corridor project, using a van, the team offers screening for HIV and hepatitis C. The van is equipped with internet services connected to telemedicine clinicians at Dartmouth-Hitchcock Medical Center in New Hampshire. “The goal is to demonstrate the effectiveness of bringing needed hepatitis C treatment and harm reduction services to rural areas where the need is great and lack of transportation is a major barrier to care,” says Friedmann.

Some of those in the study who test positive are randomly assigned to enhanced usual care that includes care navigation and referral for treatment and follow-up at the nearest provider, which can be several towns away. They are also offered free sterile syringes and referrals, if they want, to harm-reduction programs, which use humility and compassion to engage with people who have substance use disorders to improve their overall wellbeing.

Those who are randomized to the experimental group get immediate advice via telemedicine from Better Life Partners clinicians, and prescriptions for anti-viral medications to treat their hepatitis C infection. Those in the experimental arm of the study can then return to the van—near where they normally hang out—for follow up over at least 8-12 weeks, rather than needing to travel a distance for care and follow-up.

“Our team is exceptional. They’re listening and providing support every week,” says Stopka. “We are adjacent to other agencies that clients visit anyway, such as homeless shelters and places that provide meals. We have identified strong local partners who are fixtures in the community, and from whom we learn every day.”

Results of this study are still a year away. But Stopka is hopeful.

“If we can decrease the pool of infected people, that means they, in turn, are infecting fewer other people,” he says. “That’s a huge win. Together with the work we are doing to predict when and where overdoses are happening, we can begin to imagine concrete ways to really slow the devastation caused by opioids in the years ahead.”

Citation: Research reported in this article was supported by the National Institutes of Health’s National Institute on Drug Abuse, the Substance Abuse and Mental Health Services Administration, and the Centers for Disease Control and Prevention under awards 4UG3-UH3 DA044830-03, UM1DA04912, UG1DA050067, 1NU1ROT000018-01-00. Complete information on authors, funders, and conflicts of interest for these projects are available in the published papers here.

Disclaimer: The content is solely the responsibility of the authors and does not necessarily represent the official views of the funders.


Back to Top