May 9, 2016
Palm Springs, Calif.—A risk stratification and mitigation tool is helping to reduce the risk for opioid-related adverse events at the Department of Veterans Affairs (VA). According to data presented at the 2016 annual meeting of the American Academy of Pain Medicine, the Stratification Tool for Opioid Risk Mitigation (STORM) is a safer, more effective approach to managing pain, with and without opioids, compared with alternative risk mitigation strategies.
“Using STORM, researchers found three types of patients who are high risk,” said Jodie Trafton, PhD, director, VA Program Evaluation Resource Center, Central Office of Mental Health Operations. “Patients with mental health and substance use disorder risk factors, patients with multiple medical comorbidities, and patients [on] high doses and/or polypharmacy. … This differentiation of populations has allowed us to focus and tailor our risk mitigation approach and alter treatment strategies to prevent opioid-related adverse events.”
As Dr. Trafton explained, the challenges faced by VA providers and their patients are vast. In addition to the approximately 500,000 new veterans in pain, there are the numerous older veterans afflicted with diseases of aging. “It’s a big challenge for us,” said Dr. Trafton, who noted that the pain experience by veterans is often more complex than that of the general public. “In the military, opioids are often used to keep people functioning and doing their jobs.”
Dr. Trafton and her colleagues developed STORM with clinical data available in VA administrative records and Centers for Disease Control and Prevention data regarding deaths and cause of deaths. With simple predictors that could be rapidly calculated from the data—mental health comorbidities, substance use disorders, medical comorbidities, opioid dose, coprescribed sedatives and information about prior adverse events—the tool generates a real-time estimate of patients’ risk, resulting in a score between 0 and 1. An individual with a 0.5 score on the model, for example, would have a 50% chance of having an adverse event in the time window studied.
STORM incorporates two separate models: The first looks at one-year risk for poisoning or suicide-related events; the second includes a three-year time window of these events in addition to accidents, falls or drug-induced conditions (e.g., psychosis).
“We’ve taken a large-scale, holistic approach to risk, focusing on broad categories of behavioral risk in the population exposed to opioids,” said Dr. Trafton.
To show the benefits of using this approach, Dr. Trafton noted alternative risk management strategies that have been employed at the VA and other health care systems.
“By looking at patients with coprescribed sedative medications, you would have to screen or intervene on 185,000 patients to identify slightly less of the population than STORM,” she said. “Only 2.7% of patients with coprescribed sedative medication had any adverse events, barely above the actionable average rate in the VA.”
Predictors of Opioid-Related Adverse Events
Not surprisingly, the strongest predictor of a poisoning or suicide-related event is having had one in the past. “Although this probably seems obvious,” said Dr. Trafton, “recent studies have shown that over 90% of patients have no changes to their treatment plans after these events, which suggests that people aren’t intervening consistently in that population.”
Patients who have had substance abuse detoxification or inpatient mental health treatment also were found to be at very high risk in the model, although patients with opioid use disorder weren’t at higher risk compared with those with other substance abuse disorders, Dr. Trafton noted.
“Almost all mental health disorders contributed to risk, not just [post-traumatic stress disorder] and depression, which have been focused on in the past,” she reported. “These patients really warrant close clinical follow-up.”
Other adverse events—emergency department visits, falls or accidents—also were strong predictors. Even if those events are not particularly related to overdose or suicide risk, said Dr. Trafton, clinicians should be focusing on them.
On the prescription side of the equation, researchers found that patients on short-acting opioid medications had the same risk as those on long-acting formulations, whether they were acutely on those medications or had been on them for more than 90 days. “Paying attention to only the patients on long-acting opiates doesn’t make a lot of sense,” she explained. “You need to pay attention to patient risk right from the start.”
Dr. Trafton also noted a 1.4 times increased risk with coprescription of sedatives. Risk was only slightly elevated, however, with increased levels of the morphine-equivalent daily dose. Patients who had been recently prescribed other classes of evidence-based but sedating pain medications also were at higher risk. “The more different drug classes the patient had been prescribed in the last year, the greater their risk of overdose,” said Dr. Trafton.
STORM provides an individual patient report that includes basic demographic details as well as information about their relevant medications. The report also provides a list of risk mitigation strategies specifically applicable to the patient while tracking prior implementation strategies.
“This allows us to quickly identify interventions that might be clinically appropriate,” said Dr. Trafton.
Finally, the model includes hypothetical risk scores of opioid dose reductions of 50%, 90% and 90% with discontinuation of sedative.
“This hypothetical assessment helps to differentiate patients for whom prescriptions are major contributors to adverse events versus those whose risk is primarily due to other factors, allowing us to focus and tailor our risk mitigation approach,” she concluded.
Dr. Trafton and her colleagues are working with pain experts and end users to refine the tool moving forward. STORM is currently available nationally for use in VA facilities.
Moderator of the session, Rollin M. Gallagher, MD, MPH, clinical professor of anesthesiology and critical care at the University of Pennsylvania, in Philadelphia, and national program director for pain management in the VA Health System, said STORM has the potential to change the trajectory of a tragic public health problem.
“This is very exciting,” said Dr. Gallagher. “We’re really getting to the point where we can differentiate groups of patients, determine what their individual risks are, and move forward with targeted intervention and treatment planning.”
—Chase Doyle
