How analytics can spot data relationships indicating opioid misuse
Big data and analytics can help the nation’s healthcare system fight the opioid epidemic by identifying relationships that are indicative of abuse or misuse, says Rick Grape, director of market planning at LexisNexis, a research and risk management firm serving multiple industries, including healthcare.
But it will take cooperation from several healthcare sectors including providers, payers and pharmacy management firms. These entities must understand the need to purchase data, verify it and then share the data to among each other to determine if a patient is at high risk for abuse or a pharmacy is at high risk to engage in drug diversion.
Socialization factors, supported with analytics, also will play a role, Grape notes. “Data wrapped in analytics can offer a unique view of how people are engaging with the healthcare system and the communities in which they live,” he explains. “Those data insights then enable extrapolation to identify relationships that are often complex and indicative of abuse or misuse of opioids. Big data and analytics can help uncover the unknown relationships that drive the widespread proliferation of opioids via fraudulent activity.”
A hurdle, Grape explains, is knowing if these entities lack identity management tools and expertise to know who the at-risk entities are. “We bring a social lens to show how all the entities interact,” he says.
Providers, payers and PBMs need to look at prescription histories, criminal records and entities that move around to different offices. Analytics can identify patients who are getting prescriptions in a city other than their own, or a doctor who is filling prescriptions in an apartment complex and all the patients are going to that doctor at the apartment.
That doctor may not just be filling prescriptions for his or her patients, but for patients’ mothers, fathers, sisters and brothers.
Another area to focus on is the dose level of opioids being prescribed, Grape warns. The dosage should be under 100 milligrams, but some PBMs are finding dosages of as much as 120 milligrams, which is excessive and also could indicate that patients are shopping around for doctors who will give the higher amount.
The challenge for IT professionals is to identify patients who may be getting opioids from several sources but it’s difficult to really know if the right patient is being looked at, Grape says. “I can be Richard Grape, or Rick Grape, or Ricky Grape.”
The same problem exists when an entity wants to sanction a physician who has not practiced opioid management responsibly, he adds. “How do you know that these sanctions are on the same doctor?”
Despite the roadblocks, data mining can help mitigate patient risk and abuse by identifying groups of patients or doctors that are more interconnected than other groups. But there are no guarantees, Grape acknowledges. “Providers, insurers, pharmacists and social groups can all be helpful, but also can be part of the risk.”