Government agencies and public health organizations are working to develop new tools to share pertinent health care data among providers and, in some cases, to predict the path of a user about to return to a life of opioid abuse.
State and local agencies have been wrestling the devastation caused by growing opioid abuse for some time, but the issue has lately been getting attention from the federal government.
In late March, President Donald Trump turned to New Jersey Gov. Chris Christie to lead the President’s Commission on Combating Drug Addiction and the Opioid Crisis, which will help develop life-saving addiction prevention, treatment and reentry initiatives.
Christie has first-hand knowledge of the depth of the problem. Drug overdose deaths rose nearly 22 percent in New Jersey between 2014 and 2015, largely due to opioids, the governor said. In 2014, 47,055 Americans died from drug overdoses. Of those deaths, 61 percent involved the use of opioids, according to the Centers for Disease Control and Prevention.
On April 19, Health and Human Services Secretary Tom Price announced that his agency would provide $485 million in grants to help states and territories combat opioid addiction through prevention, treatment and recovery services.
But on the front lines, local government agencies, states and public health organizations are working to develop new tools designed to share pertinent health care data among providers, and in some cases to predict the path of a user about to return to a life of opioid abuse.
“You can also use outlier and anomaly detection to identify individuals or practices” that are behaving outside normal parameters, said Kevin Bingham, who leads Deloitte Consulting’s practice helping local government tackle the opioid crisis.
Using simple geospatial analysis, the firm works with states to identify urgent care centers, to which some patients travel between 150 and 200 miles to acquire opioids. “By using those types of tools you see much more effective fighting of opiate pill-mills,” Bingham said. “That’s a simple analysis -- not even complicated.”
Another simple step to cut down on opioid abuse: require that physicians check their patients’ existing prescriptions before writing an order for an opioid. According to the journal Addictive Behaviors, states that checked previous prescriptions showed a 56 percent reduction in the odds that patients would “doctor shop,” or seek physicians with loose prescription policies.
While such treatments and tools are useful, public health providers say, they don’t always bring to bear the full potential of technology to address the opioid threat, including analytics, communitywide information-sharing networks and uses of big data to spot problems occurring across public health safety nets.
SAS Inc. is one of a handful of firms working to integrate those tools by developing a common, easy-to-use platform that can be rapidly adopted to address opioid and other public safety challenges across the country, including threats to the safety of children.
In the public health care community, opioid abuse often puts children in danger of abuse or neglect, which can in turn restart dormant cases of drug abuse. Consequently, IT solutions for identifying one threat may overlap the other.
“The idea of all of this is to put the power of very deep analytics, pattern recognition and anomaly detection to use by an investigator in a government agency working in a range of disciplines, whether that’s child safety or drug addiction and opioid abuse,” said Steve Bennett, director of the global government practice at SAS.
That's also the idea for the firm’s new platform, SAS’s Visual Investigator, built on SAS Viya, an open, cloud-based analytics and visualization architecture. “We wanted to keep all the things SAS has done for the last 40 years and have all of those built into the system, but not in a way where you have to be a data scientist or a Ph.D. to use them,” he said.
Bennett said the Investigator platform is designed for use in professional communities by those who are not necessarily IT experts but who could benefit from analytics built for their discipline.
“Maybe alerts and systems are built into a workflow that a government investigator is already comfortable following and using,” Bennett said. “The science is there but the system is designed for people who are investigators.”
Some of the inspiration for developing the tool came from Bennett’s experience in government, where he spent 12 years as a senior executive in the Department of Homeland Security before coming to SAS a year ago.
Bennett compared the Visual Investigator platform to “a digital incident wall” that automatically performs the difficult cognitive tasks of synthesizing and connecting information. It means investigators “spend a lot less time trying to wrangle all that information to one place.”
Analytics vs. opioids
New Hanover County, N.C., is considering using the SAS Visual Investigator technology to help tackle a county addiction epidemic where opioids account for up to 30 percent of the interventions carried out by the county’s Department of Social Services.
The system, designed to integrate data from criminal justice and public health databases into DSS data, alerts agency workers to relevant changes to risk patterns affecting children in their charge.
"The intersection of opioid abuse and child abuse and neglect is threatening to overwhelm the system," said DSS Assistant Director Wanda Marino. "Case workers, investigators and law enforcement need to understand their data to intervene quickly and try and break this tragic cycle.”
Marino, an 18-year veteran at DSS, said Visual Investigator is not yet operational in New Hanover County, but use cases for the technologies are clear. Many of those cases involve spotting the intersection of child safety and opioid abuse.
“In our world today, the opioid epidemic has a huge impact on the safety of children,” she said, adding that “a lot of children are in homes where care givers or guardians that provide for them are using drugs, including alcohol.”
Another tool for investigators are prescription drug monitoring programs that use analytics to flag “pill mills” and “doctor shopping,” providers and practices with lax requirements for obtaining prescriptions.
In contrast to analytics-based strategies to tracking prescription drug abuse, some firms are taking a broader, “care-management approach” to confronting the opioid threat.
Salt Lake City-based Eccovia Solutions offers a “care coordination management platform” that helps bridge gaps between providers in the health care arena and those more focused on social health concerns such as homelessness, chronic disease and behavioral health.
The firm’s ClientTrack care coordination platform is a software-as-a-service solution that offers states, counties and other health and human services organizations access to information about patients as well as care within the social health arena.
“The solution was designed originally to be able to provide a robust case management solution to organizations that specialize in treating these types of clients or patients,” said Eccovia Chief Program Officer Dario Benavidez. “And we’ve been down a maturation path with the application over the last 20 years.”
In March, Eccovia said it was picked by the Los Angeles County Department of Health Services to provide an integrated care coordination platform for the county’s five-year Whole-Person Care pilot program whose goal is to build an integrated health system that delivers coordinated, “wrap-around” services to 50,000 of the most vulnerable, high-risk individuals living in the county.
The program will bring together access to resources from 11 social service groups across the county, the firm said, providing coordinated services to patients who might be homeless or those fighting mental illness or substance abuse disorders.
“This care coordination platform will provide the technology infrastructure for Whole Person Care-LA and enable our teams to connect the most vulnerable LA county residents to services,” said Dr. Clemens Hong, director of Whole Person Care, Los Angeles.
“The idea is have a communitywide network of care with information that’s shared appropriately to help a patient,” he said.
Meanwhile, ClientTrack has also been used in a pilot program for fighting substance abuse that was set up by Harbor Homes, a nonprofit that offers health care and shelter support to low-income persons in Nashua and southern New Hampshire.
The Nashua “Safe Station” program reported a 34 percent decrease in substance abuse disorders since November 2016 and saved an estimated $1.9 million by diverting drug abuse patients from hospital emergency rooms to the program’s overnight respite facilities.
Samantha Cassista, quality improvement specialist for Harbor Homes, called the program “an important component to addressing substance use disorders in Southern New Hampshire.”
The Nashua program comprises six non-profit organizations with “distinct individual missions that all work cohesively to create a one-stop shop of support and services for those organizations,” Cassista said.
Those services revolve around environmental health, homelessness and substance abuse, she said. “Coordinating multiple agencies that are providing diverse services for low-income individuals is not only complex, but there are unique challenges for integrated care,” she added.
Pulling together and managing resources for a single homeless client for example might prompt coordination among parts of all six of the non-profits. To manage the logistics, Harbor Homes and the other partners rely on the ClientTrack Care Coordination platform, SharePoint and Office 365.
Because all six members use the ClientTrack platform, Nashua leaders can track the progress of participants from when they enrolled in the Safe Station program to when they are seen by a licensed drug abuse counselor.
“It’s helpful and it works, but frankly our state has only until this year started to really encourage and create support in the community for trying collectively into becoming more of an [health information exchange] entity – and we’re working toward that,” she said.
For individuals and families dealing with addiction, mobile technology shows promise as a common tool for dealing with the opioid menace as it jumps from person to person.
In the fall of 2016 the Food and Drug Administration held a Naloxone App Challenge that awarded cash prizes for low-cost, scalable mobile apps that help opioid users, their friends and families and first responders identify and react to an overdose by administering naloxone, a medication that reverses the effects of opioid overdose.
The winning app, OD Help, connects potential opioid overdose victims with a crowdsourced network of naloxone carriers that can be tailored for rural or urban use. It features an optional interface with a breathing monitor that detects when a victim’s breathing rate is dangerously low, a sign of an opioid overdose, and alerts a nearby naloxone carrier of the potential overdose. The app also provides instructions on how to correctly diagnose an overdose and administer naloxone and helps contact emergency medical services when help is required.
The federal Substance Abuse and Mental Health Services organization app competition focused on supporting people in recovery from opioid use. Winning entries included an app that incentivized users with rewards for positive responses and one that leveraged chatbots for advice on what to do in situations when users were considering relapsing.
Artificial intelligence and other machine learning techniques may also eventually be helpful in tracking and managing opioid abuse, but so far they have had limited use, according to analysts.
IBM has been testing its Watson cognitive computing system to help develop treatment plans for cancer patients, but its work inside the health care field is in early stages.
“I don’t think I’ve seen anything to suggest that they’ve done that with the opiate epidemic, where it has the ability to take a patient that comes in on the fly and use all kinds of AI,” Deloitte’s Bingham said.
“Getting to machine learning will be exciting, because that brings a predictive aspect to this,” he said. “So it’s going to take a little while.”
Editor's note: This article was changed April 21 to correct a reference to Visual Investigator.
NEXT STORY: Digging into government spending data