More than 42,000 Americans died from opioid overdoses in 2016, according to the Centers for Disease Control and Prevention. In June 2018, the US Department of Health and Human Services Office of Inspector General (OIG) published a report revealing more than 71,000 Medicare beneficiaries were at serious risk of opioid misuse or overdose and over 300 prescribers exhibited questionable opioid prescribing to those thousands of beneficiaries. These numbers are alarming.
According to the HHS, drug overdose deaths and opioid-involved deaths continue to increase in the US. Deaths from drug overdose are up across the entire demographic spectrum. The epidemic has been declared a national emergency.
It’s fair to say we have a complicated journey ahead in this tragic manifestation of multiple causes. And the faster we apply data-driven decision making to curb this upward trajectory of substance and opioid-related deaths, the more lives we can save.
How can data analysis help?
The OIG is at the vanguard of fighting the surge in substance use disorder driven by opioids. While much attention has been focused on improving state-level prescription drug monitoring programs (PDMPs), the OIG is empowering organizations around the US with the tools they need to analyze Medicaid claims to identify beneficiaries at risk and provide them with the assistance they need.
The OIG has made available a toolkit for Medicaid, private health plans and Medicare Part D sponsors, which includes complimentary SAS® programming code to perform data analysis on beneficiaries. At its core, the toolkit provides step-by-step instructions to help Medicaid programs and Medicaid Fraud Control Units identify beneficiaries at risk of misuse or overdose.
Using Complimentary SAS® Code to Calculate Opioid Levels and Identify Patients at Risk of Misuse or Overdose
The US HHS Office of Inspector General Toolkit
Beyond simple stats to predictive analytics
In many places, documenting and analyzing prescription Medicare, Medicaid and other health data is still performed manually or may be limited by available analysis skills and tools. This only provides static reporting of one moment in time. But the methodology and software code from OIG goes beyond this old way of data gathering. The OIG toolkit provides a critical foundation on which states, attorneys general, and other public and private partners can begin to curb the opioid crisis by uncovering valuable data on anomalous and suspicious trends.
Turning the tide against substance use disorder, one state at a time
The OIG did not stop with just sharing the SAS code. They began, and are continuing audits of state Medicaid programs to highlight the benefits of applying analytics to Medicaid claims to identify patients at risk. In July 2018, the OIG published “Opioids in Ohio Medicaid: Review of Extreme Use and Prescribing” – which is an in-depth analysis of opioid prescriptions within Ohio’s Medicaid program based on the SAS code. The OIG found that:
- Nearly 5,000 Medicaid beneficiaries received high amounts of opioids.
- More than 700 beneficiaries are at serious risk of prescription opioid misuse or overdose.
- Nearly 50 prescribers warranted investigation.
The OIG recommended that Ohio review these beneficiaries because they may be receiving poorly coordinated care, seeking drugs to sell, or been victims of stolen identification numbers. In the Ohio report, the OIG further advised that it will conduct additional analyses of opioid use and payment for treatment in other state Medicaid programs. With new insights into substance use disorder among its Medicaid population, Ohio can begin to act. And other states will follow.
Subsequently, the OIG announced a more in-depth, state-level analysis of states’ oversight of opioids to review prescribing and monitoring of opioid use in selected states, along with associated policies, data analytics, programs, outreach and other efforts.
With the toolkit and its SAS code, the OIG has provided the starting point for states. The methodologies can help states detect Medicare and Medicaid beneficiaries at risk of addiction and identify prescribers operating as suspected pill-mills to investigate in more depth – such as the case of the largest health care fraud takedown ever in the US.
States need the ability to continuously monitor behavior to more quickly intervene and proactively investigate potential misuse by prescribers, beneficiaries, manufacturers and distributors. A multifaceted approach to combating the opioid crisis includes aggregating and analyzing opioid-related data scattered across numerous government agencies, taking advantage of more sophisticated methods that include rule-based code such as in the OIG Toolkit – and more advanced data integration, anomaly detection and predictive analytics. Pattern recognition, machine learning and artificial intelligence can help analysts working in a range of disciplines.
Extract the vital opioid-prescribing insights hidden in data to enable harm reduction
Once states uncover at-risk individuals and noncompliant prescribers, they will have to take further action. The toolkit provides relatively basic, yet valuable, analysis of claims data that produces a SAS data set as output. Analysts can export this output to a variety of accessible formats (e.g., Microsoft Excel). But what about gaining a fuller understanding of the contributing factors?
Prescribers can’t rely on PDMP opioid prescription data alone because it doesn’t paint a full picture. A multifaceted approach to combating the opioid crisis includes aggregating and analyzing opioid-related data scattered across numerous government agencies. Now, analysts can take advantage of more sophisticated methods that include rule-based code such as in the toolkit – and more advanced data integration, anomaly detection and predictive analytics.
To be more effective in combating the epidemic, states need the ability to continuously monitor behavior to more quickly intervene and proactively investigate potential misuse by prescribers, beneficiaries, manufacturers and distributors. Pattern recognition, machine learning, anomaly detection and artificial intelligence can help analysts working in a range of disciplines to:
- Uncover early signs of addiction by identifying patients with inappropriate amounts or combinations of opioid prescriptions.
- Anticipate and deter drug trafficking by more quickly identifying suspicious prescribing and dispensing patterns.
- Coordinate treatment by providing patient and drug insights directly to physicians and prescribers so they can make well-informed prescription decisions at the point of patient care.
Besides putting harm reduction programs in place, states must use every tool in their arsenal to curtail and fight the opioid epidemic, including the growing challenge of illicit drugs and heroin. It’s a moral imperative for public servants. Now we have a remarkable opportunity to use insights from data analysis to deploy proven methodologies and technologies to combat our skyrocketing opioid crisis. Our fellow citizens’ lives are on the line.
Prescription and illicit drug awareness dashboard
SAS recently partnered with a state government to implement an integrated dashboard that allows officials to monitor prescription and illicit drug trends. By integrating multiple law enforcement and public health data sources, the dashboard allows for data sharing and reporting to support informed decision-making and informed efforts (like these below) to help save lives.
- Public health
- Provides analytically driven decisions on whether a comprehensive response by EDs, EMS providers and investigative resources is warranted by identifying overdose spikes and their magnitude.
- Provides a mechanism to immediately notify recovery coaches and human services mobile-response vans to provide free naloxone and training in affected areas.
- Identifies and prioritizes responses using AI and machine learning to identify where to allocate resources by identifying drug trends, patterns and anomalies.
- Investigations support and investigative case management
- Helps prioritize investigative targets by identifying data linkages to uncover shadow organizations and individuals providing legal (and illegal) opioids that lead to fatal and non-fatal overdoses.
- Emergency drug scheduling
- Provides the data necessary to decide and justify which psychoactive substances required emergency scheduling.
- Inform policy development
- Provides awareness and understanding of the presence and prevalence of drugs, the impact of drug abuse on specific demographics in specific locations.
Read more on how you can curtail the substance and opioid crisis using data analytics:
- International Institute for Analytics: Data and Analytics to Combat the Opioid Epidemic: Physicians, patients, policymakers, pharma companies and more must work together to stem the opioid epidemic. The key is to tap into data to understand what’s happening and reduce addiction and deaths.
- International Institute for Analytics: How Advanced Analytics Can Prevent Medicaid Fraud: Medicaid fraud is prevalent, costly and difficult to prevent. With a combination of more integrated data and advanced analytics, state agencies can turn the tables on fraudsters by accelerating the transition from detection to prevention.
- Improve child welfare through analyticsWith tremendous potential for child welfare agencies to use data and analytics to prevent child abuse and improve outcomes for children and families, child welfare advocates discuss the benefits of using data and establishing a data-driven culture to advance practice and policy.
- Analytics for prescription drug monitoring: How to better identify opioid abusePrescription drug monitoring programs (PDMPs) are a great start in combating abuse of prescription drugs, but they could be doing much more. Better data and analytics can inform better treatment protocols, provider education and policy decisions – and save lives.
- 10 ways analytics can make your city smarter From child welfare to transportation, read 10 examples of analytics being used to solve problems or simplify tasks for government organizations.