How Health IT Tools are Working to Reduce the Prescription Opioid Epidemic

Jan. 29, 2018
Facing a growing prescription opioid crisis, healthcare provider and payer organizations are looking to leverage predictive data analytics to identify patients at high risk of an opioid overdose or addiction to intervene before an adverse event occurs.

U.S. President Donald Trump earlier this month vowed to declare the U.S. opioid crisis a “national emergency,” pledging to commit more funding and attention to the issue. And, while President Trump has yet to officially declare a state of emergency, a move that could help remove barriers and would enable the federal government to devote more funding to address the drug epidemic, healthcare organization leaders have been trying to combat the worsening opioid epidemic for several years now.

According to the Drug Enforcement Administration (DEA), deaths from prescription drug abuse have outpaced those from cocaine and heroin combined in the last 14 years. According to the Centers for Disease Control (CDC), overdose deaths involving prescription opioids have quadrupled since 1999. Further, the CDC reports that, today, nearly half of all U.S. opioid overdose deaths involve a prescription opioid. In 2015, more than 15,000 people died from overdoses involving prescription opioids. What's more, each day, more than 1,000 people are treated in emergency departments for not using prescription opioids as directed.

Many healthcare and health IT organizations have called for increased use of prescription drug monitoring programs (PDMPs) and for federal leaders to ensure that PDMPs are interoperable across state lines, in order to reduce instances of prescription drug and opioid abuse and addiction. Last year, Health IT Now wrote a letter urging the U.S. Food and Drug Administration to leverage health IT tools, in combination with physician education and training, to combat the opioid crisis.

While education and training for healthcare providers is one tool that the FDA has, Health IT Now said that training will be “reinforced and exponentially more effective when that education is paired with comprehensive and real-time data of the patient’s controlled substance prescription history. “In order for those pieces to coordinate, PDMPs need to be interoperable across state lines, real-time and within the workflow of prescribers and dispensers.”

A report from Shatterproof, a national substance use disorder prevention organization, released a report back in March that found that PDMPs can be an effective and valuable tool to help identify and prevent prescription drug misuse. However, in most states, prescriber participation is very low, which compromises the effectiveness of the clinical tool. That report called on state legislatures to require doctors to use state-run databases to track patients’ history of opioid and sedative prescriptions in an effort to address the growing opioid abuse problem in the U.S.

Health information exchange (HIE) organizations throughout the country also are taking steps to enhance data exchange specifically in the area of prescription information to help providers address opioid misuse and abuse. As an example, the Nebraska Health Information Initiative (NeHII), a statewide HIE, and the Nebraska Department of Health and Human Services, are taking steps to capture state prescription information and deliver it to Nebraska’s enhanced PDMP. According to NeHII, the inclusion of complete prescription data in the PDMP database enhances patient safety and helps combat the abuse of opioids by giving providers and pharmacists a more complete picture of a patient’s medication history, allowing opioid use to be addressed in the overall context of the patient’s care plan while also highlighting diversion or other patterns of abuse.

Beyond PDMPs, many healthcare provider and payer organizations are looking to leverage advanced health IT tools, such as data analytics and predictive modeling, to identify risk factors that can put patients at high risk for a prescription opioid overdose or addiction in order to intervene before an adverse event occurs.

Virginia Premier Health Plan, a managed care insurance organization owned by VCU Health and based in Richmond, Va., plans to implement a newly developed predictive analytics tool that predicts a patient’s likelihood of experiencing an overdose. The tool, called the Venebio Opioid Advisor (VOA), was developed by Venebio, a Richmond-based life sciences consultancy.

The company conducted numerous published opioid studies, with the support of the National Institute on Drug Abuse (NIDA) of the National Institutes of Health (NIH), and, as a result of that research, researchers were able to identify the most common risk factors associated with an opioid overdose and developed predictive algorithms that can predict the likelihood of a patient experiencing an unintentional overdose from a prescription opioid.

Barbara Zedler, M.D., Venebio’s chief medical officer and a lead researcher on several of the studies, says many unintentional overdoses occur not because of excessive dosage, but from other factors like age, concomitant medications and pre-existing health conditions that can increase the risk of certain individuals treated with prescription opioids. The risk-screening tool analyzes these risk factors and predicts a patient’s likelihood of a life-threatening opioid overdose.

“It is tragically ironic that prescription opioid overdose deaths have increased nearly four-fold to almost 50 per day in a time when we have all the data we need to effectively target at-risk patients and address their personal risk factors,” Zedler says “With VOA, we can leverage data that already resides in virtually every electronic medical record (EMR) system and in every payer’s claims data warehouse to identify even hard-to-find patients at risk and help reduce their risk of a prescription opioid overdose.”

Venebio’s tool, VOA, is a clinical decision support tool that quantifies a patient’s likelihood of experiencing a life-threatening overdose from a prescription opioid, determines a personalized risk factor profile for each patient and provides clinicians with individualized guidance regarding interventions to reduce the patient’s risk of overdose. Speaking to the development of the predictive tool, Zedler says, “As a physician, I wanted to be sure [the tool] was very practical, quick to operationalize and could be done right there at the point of care by a physician or other healthcare professional.”

Zedler says Venebio has conducted retrospective validation studies in populations as large as 18 million opioid users and has published four peer-reviewed studies, two of which demonstrate VOA’s ability to predict with 90 percent accuracy the likelihood of a patient experiencing an unintentional overdose from a prescription opioid. Initial adopters of the VOA include health plans and Medicaid-managed care plans in Virginia and New York. The company also is working with an EMR vendor to integrate the tool into its EMR software.

Virginia Premier Health Plan serves about 200,000 members in more than 100 Virginia counties and operates as a Medicaid plan as well as a plan for people are dually eligible for both Medicaid and Medicare. Javier Menendez, vice president, pharmacy operations, Virginia Premier Health Plan, said the organization is currently using analytics tools to identify patients at risk for opioid addiction in order to intervene and connect those patients to needed services. However, he notes that the organization’s current tools are more retrospective, while the VOA offers more sophisticated clinical predictive algorithms.

“For us, it’s all about data integration from a clinical perspective to tie in all the different data claim entry points, like hospital, medical, labs, pharmacy and, through all those, put together our target of finding people at risk for opioid addiction. That may include people that have had previous claims for overdose, or abuse or toxicity. And, in addition to above average utilization or what would be considered high utilization of these controlled substances, we put those two together, and we pretty much know that this person is at high risk for this. And, then we see what we can do to intervene to prevent the addiction, or even to engage to connect the patient to recovery services,” he says.

He continues, “Patient safety is our highest priority, and the risks associated with prescription opioids, at any dose, are substantial. This [VOA] tool will give us an earlier point of intervention to identify those patients that are at risk through some predictive modeling. The Venebio Opioid Advisor offers us the best option available to identify those patients at risk of a prescription opioid overdose and take preventive action before any harm can occur.”

Many physicians already are using predictive screening assessments, where patients are asked about factors in their histories known to be associated with opioid abuse, in an effort to offset addiction risk. However, asking questions has its limitations. Some healthcare IT leaders contend that natural language processing (NLP) technology can be used to find patterns in patient data that point to opioid abuse risk.

NLP is a technology that allows providers to gather and analyze unstructured data, such as free-text notes, and NLP technologies are gaining traction in healthcare as there is an explosive growth of unstructured clinical data available in electronic health records. By leveraging NLP technology text mining, physicians and clinicians may find risk factors for opioid abuse hidden in unstructured data, Elizabeth Marshall, M.D., director of clinical analytics at Linguamatics, a U.K.-based NLP-based text-mining software provider, says.

“With regard to addiction and mental health, it can be difficult for general physicians, such as family practice physicians, especially out there in rural areas, to actually capture information that indicates that a patient is at risk. They are probably asking the right questions, but they might not be asking these questions all within the same visit with the patient. It’s a matter of collecting this information, throughout the timeline of the patient, and if you apply natural language processing and use a predictive model, like the opioid risk tool which has already been validated, you can find information throughout the patient’s timeline that is specific to that risk tool,” Marshall says, referring to the screening tool available at that can be used with adult patients in primary care settings to assess risk for opioid abuse or misuse.

“When it comes to technology and NLP and artificial intelligence, I don’t think we should remove the human factor, but it can enhance our capabilities as humans,” she says.