As the opioid crisis continues to worsen, claiming the lives of more than 130 people each day in the U.S., the healthcare industry needs to dig deep in order to determine the role that prescription medications play. After all, about 80 percent of people who use heroin first misused prescription opioids.
Currently no one solution exists that can effectively address the entirety of the opioid crisis. It will take unprecedented collaboration across industry stakeholders if we are to manage and slow this epidemic both now and in the future. Big data and analytics can help tackle the challenge by identifying and evaluating entities who contribute to potential risk, or who are themselves at-risk, within the various points of the healthcare ecosystem.
A critical lens that is often missed is the need to analyze data from traditional health sources and combine this information together with non-medical sources of data from public records. Public records provide insights into at-risk entities, behaviors, and connections – offering a unique view of individuals as they engage throughout the healthcare system. Outputs of this risk evaluation provide astounding insights otherwise unavailable when dealing only with healthcare or non-healthcare data separately: they identify social groups and other “networks” of schemers who work together to perpetuate the dangerous cycle of drug availability and abuse. An analysis of each data set by itself is simply not comprehensive enough to reveal the entire spectrum of opioid abuse patterns: lacking the relationship component that drives all diversion tactics.
Relationships and risks
Relationships are the foundation on which our communities are built, driving people’s behaviors, dependencies and, unfortunately, schemes. According to the Centers for Disease Control and Prevention, drug diversion – the transference of legally controlled and prescribed substances from one individual to another – is the number one avenue for opioid abuse. By uncovering relationships, associations, and affiliations among providers, patients, and pharmacies, healthcare stakeholders can identify players who contribute to risks or who are at personal risk of prescription drug abuse and fraud. These individuals may be behaving either intentionally or inadvertently, but nonetheless warrant further evaluation.
Patients, or health plan members, are the most at-risk group in fraudulent schemes and this group includes individuals who are new to taking opioid-type medications, or the opioid naive. Patients who receive multiple opioid prescriptions are also placed at greater risk due to the cumulative effect that these medications have. Other risky patients include those who may be intentionally abusing prescriptions, partaking in recreational drug use, or reselling drugs on the black market.
Prescribers are another source of potential risk due to unknowing, irresponsible, or fraudulent prescribing behaviors. Writing prescriptions to friends and family members is an immediate red flag, as is prescribing excessive quantities of certain drug types across many patients. Physicians may also unknowingly prescribe lower quantities to high-risk patients, ultimately putting these individuals at-risk.
Pharmacies are a third source of potential risk as fraudsters can target them with counterfeit scripts, or, when they lack sufficiently robust patient and provider screening, become targeted as easy-to-fill locations. By serving high volumes of patients who seek a certain type of substances or by filling prescriptions not associated with a corresponding medical condition, pharmacies become inadvertent participants in the propagation of fraudulent schemes.
Uncovering patterns in big data
However, intelligence isn’t lacking on these risk-prone players. There is a massive amount of transactional information about patients, providers, and pharmacies and their respective roles in the opioid epidemic. By analyzing large quantities of prescription data, in combination with public records data, stakeholders have an opportunity to detect which providers are engaged in high-dose script writing, instances where opioids are being prescribed to large social or family groups, and when prescribing has occurred to patients with a high-risk for potential abuse, among others.
Coupled with analytics, this transaction data can also surface situations where a large number of prescription seeking patients for a particular pharmacy originate from a single physician, or even where prescriptions are written without a corresponding doctor or hospital visit. Data can reveal “frequent flyers” or “doctor shoppers,” patients who go to one or multiple providers for high-risk drugs within a short period of time.
Patients seen exhibiting this behavior are likely supporting a drug habit or seeking to divert drugs. Since they typically act quickly and pay with cash, they can be tough to catch without a supporting technology that tracks—and flags—suspicious steps. Proper identity resolution forms the foundation of this process: it is critical to identify the correct individual even when he or she uses an identity variation. For example, many systems will fail to catch that Richard Grape, Ricky Grape, and Rick Grape with slightly altered inputs are all the same person; an error that could result in the receipt of additional prescriptions.
Visualization, an important tool in the big data and analytics arsenal, can help stakeholders quickly see relationships that identify interconnected entities and allows them to focus on the social groups of interest. These networks of entities who work together to drive widespread drug diversion, may exist completely outside of the scope of healthcare data. Relationships may include family members, associates, colleagues, roommates, members of social organizations, joint owners, and businesses that individuals frequent, to name but a few. When public record insights are coupled with healthcare interactions, patterns quickly emerge identifying entities and clusters of potential risk and abuse.
Once these at-risk or high-risk entities are confirmed, healthcare stakeholders can bring these insights together in order to mitigate drug diversion and non-medical opioid use throughout the ecosystems’ various workflows. Consider the value of answering the following questions about a patient within an identified social network:
- Is she part of a social group whose members are trending toward excessive total Morphine Equivalent Dose (MED) calculations? – Tracking daily patient MED totals can help identify individuals who are potentially at risk for overdose, abuse, or diversion.
- Is he receiving multiple prescriptions that cause him to exceed the daily safe MED?
- Is she within the network of a patient receiving the same drug type?
- Is he receiving prescriptions from a provider within his social network?
- Is she filling a script from a prescriber that services high-risk patients?
- Is he receiving multiple prescriptions within a short timeframe?
- Is she part of an upward trending high-risk social group?
In cases where answers to these questions merit further investigation, it is often determined that the social networks involved are working with pill mills to acquire opioids for non-medical reasons. Pill mill behaviors typically involve providers, clinics, and pharmacies that fill specific high-risk prescriptions frequently and without proper due diligence. Through healthcare claims and public records data, the stakeholders who need access to at-risk or high-risk intelligence can gain significant visibility into the key offenders and potential collateral of excessive prescribing.
It’s important to note that these data insights surface not only knowing participants, but may include prescribers and pharmacies who are inadvertently participating in pill mill activities. These entities can be targeted by large networks of collaborating patients who have organized together to obtain large quantities of appropriately dispensed high-risk drugs. Identifying these instances provides opportunities for education as well as further screening.
Scoring and sharing
As we begin to work together and fight the opioid epidemic, it is critical to identify and evaluate those at-risk and those who are sources of risk, regardless of which doctor, pharmacy, provider, or health plan had identified them first. Each party can benefit from one another’s lessons learned during historical and ongoing interactions. It is only through collaboration that stakeholders can share insights to detect and prevent risk at an industry level.
Imagine a future state in which every healthcare stakeholder would immediately benefit from the collective intelligence of their peers and counterparts, using this information to prevent, detect, and mitigate further behaviors that threaten patient health and industry integrity. By securing and sharing insights, we will not only reduce the unknown risks, but also increase our potential to overcome this crisis.