Future Tense

Why Data About the Opioid Epidemic Is So Unreliable

And how to fix it.

Justin, a participant in a class on opioid overdose prevention by nonprofit Positive Health Project, practices with naloxone on Tuesday in New York City.

Spencer Platt/Getty Images

Headlines about the opioid epidemic come with often staggering reports of the numbers of deaths, of overdoses, and of lives saved by Naloxone. According to data released from the Centers for Disease Control and Prevention, there were 52,404 total deaths in 2015, or 144 drug overdose deaths per day. Overdoses are now considered the leading cause of death of people under the age of 50, according to a New York Times analysis using preliminary data.

As staggering as those numbers are, though, there are many reasons to believe the numbers we have are unreliable. One recent study estimated that due to variations from state to state in filling out death certificates, opioid deaths may be underreported nationally as much as 24 percent. If that is true, it’s dangerous: It means that we aren’t fully grasping what is already considered an epidemic or responding appropriately. To help fight this epidemic, we need numbers that are accurate and reflective of the current moment. Community-based coalitions can have a stronger impact if they have access to timely, accurate data that reflect the situation on the ground.

In April, the CDC presented a report explaining that opioid-related deaths might be underestimated for several reasons. For one thing, many autopsies show pneumonia as the cause even when the toxicology report shows a high level of opioids in the body. Furthermore, coroners’ guidelines state that a death can only be classified as an overdose if the toxicology report shows a certain blood level. That may seem reasonable, but drug levels can drop fairly quickly after death. If the medical examiner doesn’t do the autopsy soon enough, the toxicology report may not be accurate. In addition, rural counties faced with strained budgets don’t always do toxicology reports due to cost, and without a toxicology report, a death can’t be labeled as an opioid overdose.

Lastly, lots of people die in ways that are related—but difficult to formally connect—to opioid addiction, like suicide or car accidents caused by driving while under the influence. More than 1,000 families affected by the opioid epidemic have contributed to the Celebrating Lost Loved Ones interactive map, a project I created to help break stigma and raise awareness about the epidemic after I lost my little brother JT to the opioid epidemic. He was the first person to be listed on the map.

Read some of the biographies on the Celebrating Lost Loved Ones site, and you’ll see that many of the deaths don’t officially fall into the overdose category. Some of the families believe that their lost loved ones weren’t officially labeled overdose deaths because of their community’s strong desire to avoid any reporting of drugs. Regardless of the reasons why, it is clear our nation isn’t getting a full picture of this epidemic no matter how high the numbers may seem.

One way to help tackle this problem would be for county and state health departments to start examining “opioid related deaths”—for instance, if the toxicology report shows traces of opioids in the system of a person who died by suicide or if recently used drug paraphernalia was found at the scene. Some agencies are starting to examine data in this way. For instance, Oakland County, Michigan, has created an opioid-related deaths map. While it still may not be comprehensive, it gives local officials a far more realistic picture of the state of opioid abuse. Duplicating that could be difficult for county governments with larger populations and massive numbers of autopsy reports, but perhaps they could use technology to comb through all these records to look for signs that opioids may have been involved.

We also suffer from out-of-data information. Thanks to a lengthy process, it takes a while for an opioid death to be counted in statistics. Typically, a local county coroner or medical examiner will report data to the state health agency, which in turn passes the information to the CDC. The most recent data on the CDC website contains information on overdose deaths from 2015.

But concerned citizens, nurses, elected officials, law, and health staff from government agencies need real-time data (or close to it) to support their efforts of education, prevention, and treatment. Having better information could help direct resources where and when they are most needed. To be exact, communities need access to three key datasets that could be monitored and mapped locally in real time:

  1. Death data from the local coroner or medical examiner.

  2. Overdose data from law enforcement agencies and their record management systems. Some overdoses may result in a death, though many cases people survive. What’s critical here is the location where these ODs are happening.

  3. Data about the number of naloxone saves (that is, people resuscitated using naloxone, an anti-overdose medication) from first responders such as fire departments, police or sheriff departments, and emergency medical services. Communities could also crowdsource this data. Naloxone is now readily available from pharmacies, and many local task forces offer trainings on how to administer. If the community helps track where naloxone saves are coming from, it will help fill in data gaps to complement what first responders are capturing.

Of course, we have to be careful with this data due to sensitivity and privacy concerns. But there are many ways to abstract the data for public awareness. For instance, Tri-County Health in Colorado creates heat maps. You can also aggregate by a geographic boundary such as a ZIP code or municipality—the information doesn’t have to be as specific as a street address. Being too guarded about sharing the data won’t create the actionable information a community needs.

Addressing this nationwide crisis will require collaboration between citizens, local officials, health departments, and law enforcement. Knowing where and how to respond to treatment and awareness efforts will not happen efficiently unless these individuals can see up-to-date, accurate data streams.

Of course, there are other societal issues related to the opioid epidemic. Rising homelessness and people entering and leaving jails are related to the increase in opioid use. As communities address the opioid epidemic, they can also share their information with other projects such as Data-Driven Justice.

Technology and data have helped solve and address many social problems in recent years. But local governments still aren’t doing enough data sharing about health-related issues. Down the road, we will face another epidemic that seems unlikely, even impossible right now. Laying down a practice of using data in a transparent and up-to-date fashion will arm agencies to respond and collaborate better in the future. Communities across the nation are struggling with the opioid epidemic, and getting control of their own data will be a huge support in their mission to save lives.

This article is part of Future Tense, a collaboration among Arizona State University, New America, and Slate. Future Tense explores the ways emerging technologies affect society, policy, and culture. To read more, follow us on Twitter and sign up for our weekly newsletter.