Everyone knows the terrible toll the opioid epidemic takes on our society. It doesn’t matter where you live and how wealthy or educated you are, no one is immune. I watched as a colleague tried to help a child struggling with addiction, spending countless hours of worry and thousands of dollars, only to have the child succumb to an accidental opioid overdose. The emotional and financial cost on individuals and families caused by opioid addiction and overdoses is immense and impacts everyone.

Chances are you or your family know someone who has become addicted to drugs (namely opioids), quit, relapsed, and even overdosed.

The numbers tell the story. According to the CDC, in 2020, nearly 92,000 people died from drug overdoses, making it a leading cause of injury-related death in the U.S. Of those deaths, nearly 75 percent involved a prescription or illicit opioid. And these drug overdose deaths continue to increase.

Deaths from prescription opioids like oxycodone and hydrocodone increased by 17 percent from 2019 to 2020. Unfortunately, opioid problems aren’t just the realm of drug dealers on the streets. The origin of many opioid use disorders can be traced to legal opioids prescribed by physicians.

Opioid addiction is a complex problem, and solutions require research to understand how we got here. Research of this nature requires robust data to track patterns across many patients over a period of years. A database with these characteristics is hard to come by.

Stanford researchers and graduate students, led by Dr. Tina Hernandez-Boussard, set out to develop a predictive model to identify risk factors for how non-opioid users (opioid “naïve “patients) become chronic users. Stanford leveraged a unique research database of millions of de-identified Medicaid paid claims. Select Medicaid clients approved using this data for health care research, knowing it would be used to help address significant problems like the devastating opioid crisis.

Stanford combed the database using state-of-the-art machine learning techniques to develop the models.

In the end, the research results were both instructive and actionable.

The Stanford researchers concluded that a patient’s first experience with an opioid prescription is the biggest factor determining potential chronic use.

An initial population of 180,000 de-identified Medicaid recipients from six states with opioid prescriptions formed the study group. This group was opioid “naïve,” meaning they received no prescriptions for opioids for at least six months prior to the date of the opioid prescription used as the index for the study.

Of the opioid naïve patients in the study group, 29.9 percent had evidence of receiving additional opioid prescriptions 3 to 9 months following the initial prescription (the definition of chronic opioid use).

Key factors for chronic opioid use determined by the model included details of the initial prescription, number of pills, duration of the prescription, and types of opioids prescribed.

The results of the study are exciting. As a physician, I know there is a challenging line between appropriately controlling pain and putting patients at risk. These data-driven results provide more clarity on how to use opioids responsibly while decreasing the risk of addiction. The results of the models developed in this study are actionable in that they identify high-risk prescribing patterns and at-risk patients, enabling outreach and education.

The research also provides valuable insight for care management and population health programs to improve outcomes for society’s most vulnerable populations. Translating research from leading academic institutions like Stanford into real-world use cases that solve important health care problems is key to creating a more equitable and effective system for all.

Gary Call is a physician and health care executive.

Image credit: Shutterstock.com