New Screening Tool Identifies 80% of Patients at High Risk of Rx Opioid Abuse
/By Pat Anson, PNN Editor
A new opioid screening tool is being introduced that claims to predict – with 80% accuracy – which patients receiving an opioid prescription are likely to have an overdose or get an opioid abuse diagnosis within the next six months.
Geneia, a healthcare data firm based in Pennsylvania, is integrating the opioid screening tool into its analytics platform, which is used by hospitals, insurers, employers and physician groups providing healthcare to over 7 million patients.
Contrary to popular belief, abuse is relatively rare in patients prescribed opioids, so it’s a bit like finding a needle in a haystack. Less than one percent (0.7%) of patients on opioids suffer an overdose or abuse their medication, according to Geneia.
But the consequences of opioid abuse are so serious medically, legally and financially to both patients and providers that healthcare organizations are eager to identify patients who are most at risk.
"An ounce of prevention is worth a pound of cure," said Fred Rahmanian, Geneia’s chief analytics and technology officer. "Knowing in advance who is at high-risk for an opioid abuse diagnosis or overdose enables physicians to choose alternative pain management strategies."
For example, a hospital using Geneia’s screening tool might give a surgery patient deemed to be “high-risk” a non-opioid pain reliever like acetaminophen for post-operative pain.
‘It’s All Data Driven’
Opioid screening tools have been used for years to help providers identify high-risk patients. Usually they involve questionnaires in which a patient’s medical history is assessed, along with their personal and family history of alcohol and drug use. The risk assessment might also include genetic testing or asking if a patient suffered sexual abuse as a minor.
Geneia’s screening tool is different. It only looks at the diagnostic codes used in a patient’s medical and pharmacy claims – known as ICD-10 codes – without getting any direct input from providers about patient history. The company’s algorithms search the diagnostic codes for 22 different variables that may indicate a patient may be prone to opioid abuse.
“It’s all data driven,” says Geneia CEO Heather Lavoie, who downplays the importance of questionnaires about patient history.
“Oftentimes you don’t get accurate information. Even about family history of substance use and abuse,” she told PNN. “People are not good historians. There’s often a lot of misinformation, so you’re not always getting an accurate response to questionnaires. Obviously, sometimes you do, but not universally. And so, the benefit is that we can, with very little information, be highly predictive.”
“Glad to see there are efforts to help identify people at increased risks of abuse. We need accurate and easy to use predictive criteria for those at increased risk of abuse and overdose,” says PNN columnist Dr. Lynn Webster, a pain management expert who developed the first “opioid risk tool” about 15 years ago.
“Much of the accuracy of tools depends on definitions. Unfortunately, there are multiple definitions of abuse, misuse and addiction. This makes it difficult to understand what is actually being assessed and predicted with any tool, including this one.”
Webster is puzzled why Geneia’s screening tool omits a patient’s personal history and genetic testing, and says it may rely too heavily on diagnostic codes.
“Unfortunately, few physicians know how to diagnose abuse from misuse and addiction. For example, many providers still believe that if someone experiences withdrawal they are addicted, which is not necessarily true,” said Webster. “The bottom line is that if the tool is based on ICD-10 diagnoses it relies on the clinician’s ability to make accurate diagnoses, most of whom have never been trained to make such diagnoses.
“I am surprised that genetic evaluation is not considered in their tool since the vulnerability of opioid addiction is mostly genetic. The balance of the vulnerability to addiction comes from environmental factors like social-economic despair and mental health disorders.”
Another thing that makes Geneia’s screening tool unique is that it can be used with children and adolescents. Opioid risk tools are typically only used for patients 18 years of age or older.
High False Positive Rate
Geneia tested its screening tool against a database of several million insurance claims, and says it accurately identified 80 to 88 percent of patients who are likely to have an opioid abuse diagnosis or an overdose.
But analytics are not foolproof. Geneia’s screening tool has a false positive rate of nine percent – meaning about one in every ten patients will be flagged inappropriately as high-risk.
“The false positive rate, depending on the data-set, is pretty high because it’s a very rare event rate. It’s really hard to pick the needle out of the haystack,” explained Jasmine McCammon, a data scientist who designed Geneia’s screening tool.
On the flip side of opioid abuse, the screening tool also identifies patients who are at not high risk, giving reassurance to doctors who are uncertain about whether a patient should be given opioids.
Regardless of what the screening tool finds, CEO Lavoie says the company’s analytics should never be used as a diagnostic device.
“Predictive models shouldn’t be used in place of strong medical practice. They really are to help advise and to provide contextual information, so they (doctors) can work with the patient more directly. It’s in no way a surrogate for good care, but it does help provide additional caution,” Lavoie said.
