“Currently most of the data that is collected for the opioid epidemic is collected in SAS,” said SAS Medical Director Steve Kearney. “We see a huge advantage in leveraging that nationally. If we can share this information, we can all learn from it.”
To that end, SAS recently launched an Opioid Analytics Users Group. The group aims to bring together leaders from the public sector, academia and industry to share best practices and develop new solutions to a drug epidemic that presently claims more than 130 American lives a day.
Data-driven approach
The analytics user group plans to take a deep dive into the data surrounding opioid use, starting with state and federal prescription drug records.That data could help researchers chart new approaches to fighting opioid abuse, but first it will have to be cleaned up. “Right now, there is no common identifier in that data set: No Medicaid number or insurance number, so the same individual can be John Smith or J.R. Smith or John R. Smith,” Kearney said. It's an issue common to most data sharing efforts.
In addition to addressing this kind of systemic record-keeping problem, the user group will be looking for novel ways to utilize the data that is currently available. For example, geographic data around usage patterns could help states and cities to better allot their treatment resources. “Where should we have treatment facilities? As a state looks at the Medicaid population, where can individuals receive treatment? The data can tell the story of where services are needed,” Kearney said.
Researchers also can look at data around illicit drug use, culled from police reports and other sources. “We can use the data to identify hot spots, to get rapid recovery teams in, to deliver services to those people — and also to identify the sources of these illicit drugs,” Kearney said.
Early examples
Even as the analytics user group ramps up its efforts, SAS already is working with various state agencies and other key partners to leverage data in support of the effort to contain the opioid epidemic.The New Jersey Attorney General’s office has been at the forefront of that effort. It recently launched an integrated, information-sharing dashboard called the Integrated Drug Awareness Dashboard (IDAD). The AG’s office is looking to analytics to target intervention initiatives and enhance public outreach.
The first phase of the project breaks down silos to combine law enforcement data with other sources of information, offering visibility into the number and types of prescription opioids being dispensed throughout the state, along with drug lab analysis data and the locations of heroin, fentanyl and other drug-related arrests.
“New Jersey has created an environment where they can look at the prescription drug side and also look at the illicit side. They can look at arrests and shootings and toxicology lab data to identify the hot spots for the various forms of fentanyl,” Kearney said. “They are using those for rapid response teams and other public health interventions. They can target their policies, their grant funding and other tools.”
SAS also has collaborated with federal agencies in support of state efforts.
In June 2018, for example, the U.S. Department of Health and Human Services Office of Inspector General (OIG) published a report showing more than 71,000 Medicare beneficiaries were at serious risk of opioid misuse or overdose. To help state fraud units leverage the data, OIG developed a toolkit that includes complimentary SAS programming code to perform data analysis on beneficiaries.
The OIG continues to help states expand their use of the data. Its in-depth analysis of opioid prescriptions in Ohio, for example, found that nearly 5,000 Medicaid beneficiaries received high amounts of opioids and more than 700 beneficiaries are at serious risk of prescription opioid misuse or overdose. It also identified nearly 50 prescribers that it said warranted investigation.
In addition to supporting these government-wide efforts, SAS also has been collaborating with academia. It recently partnered with graduate students from Carnegie Mellon University's Heinz College of Information Systems and Public Policy to leverage data mining and machine learning capabilities around Medicare data.
That research is intended to help the Centers for Medicare and Medicaid Services identify people at risk of opioid overdose, and also to help uncover Medicare provider fraud.
Looking ahead, the analytics user group plans to focus on Medicaid recipients in 10 states where those individuals have been identified as being especially at risk for addiction. The group will also look at best practices in culling relevant data from disparate data silos, especially data concerning criminal drug activities.
“We want to find better ways to tackle the illicit use,” Kearney said. “If there are states that are doing a good job of that, finding new ways to share that information, the user group will want to look at that.”
