Validity of International Classification of Diseases Diagnostic Codes for Opioid Use, Dependence and Abuse
The National Survey on Drug Use and Health estimated 2 million people aged 12 or older in the United States met criteria for Opioid Use Disorder (OUD). Opioid Use Disorder is often defined as the problematic use of opioids that causes clinically significant distress or impairment.
Veterans are one of the hardest hit groups by OUD. A 2018 study noted a sharp rise in OUD amongst Veterans, with diagnoses nearly tripling between the years of 2003 and 2017.
Doctors and researchers continue to gain a better understanding of the factors that lead to OUD, and are realizing the disorder is often not as “black and white” as currently categorized.
A new study, “Identifying Individuals with Opioid Use Disorder: Validity of International Classification of Diseases Diagnostic Codes for Opioid Use, Dependence and Abuse,” recently published in Drug and Alcohol Dependence, conducted a clinical chart review of 520 Veterans assigned ICD-9 or ICD-10 diagnostic codes for opioid use, dependence, or abuse from 2012-2017.
The study was led by Dr. Pooja Lagisetty, a research investigator at the Center for Clinical Management Research at the VA Ann Arbor Healthcare System.
ICD codes, or International Classification of Diseases codes, are simply codes for a diagnosis that a provider will enter into a patient’s chart after a visit that states why a patient was seen. For the purposes of estimating the number of patients with OUD and evaluating treatment outcomes, the ICD codes for opioid use, dependence, and abuse are all considered OUD diagnoses.
However, these ICD codes are applied inconsistently, and often fail to accurately convey the nature of a patient’s opioid use. Dr. Lagisetty’s study calls into question the binary distinction between a patient either having or not having an OUD diagnosis code, and instead proposes a spectrum of definitions that may better serve patients, and the clinicians that serve them.
“My worry, as a clinician, is that these codes are pretty imprecise, and there’re lots of folks that fall along the spectrum of opioid use,” Dr. Lagisetty said.
Therefore, Dr. Lagisetty and her colleagues reviewed the charts of Veterans assigned ICD codes for OUD, using information in their medical records to sort patients into four categories that more precisely describe gradations of opioid use. Those categories are:
- Patients with a high likelihood of Opioid Use Disorder;
- Patients exhibiting limited aberrant opioid use;
- Patients using opioids as prescribed without evidence of aberrant use;
- Insufficient Information.
The results of the study show that just under 60% of patients with an ICD code for OUD were categorized as having a high likelihood of opioid use disorder, while another 16.5% exhibited only limited aberrant opioid use. Nearly 20% were using prescribed opioid medication without signs of abuse, and almost 7% fell into the insufficient information category. According to the study, the numbers show that patients assigned the current ICD codes for opioid use, dependence, or abuse often lack documentation of meeting OUD criteria.
“It really demonstrates the broad spectrum of patients that we’re seeing. So, we often lump all opioid use under this large umbrella, and that’s what we found using these ICD codes. But what we’re really finding is that our Veterans fall along a much broader spectrum, and different parts of that spectrum may need to be treated in different ways.”
The results also highlight that policies and treatment guidelines stemming from ICD coding may be based on inaccurate classification of Veterans and potentially overestimating the number of individuals who have a high likelihood of OUD compared to other opioid-related diagnoses.
“What else can we put into the equation to more accurately predict OUD? How can we improve the accuracy of predicting patients with a high likelihood of OUD from 60% to 90% or 95%?” said Dr. Lagisetty. “The key here is that if we’re going to try to identify this population in large samples, if we’re going to try to monitor whether we’re providing good access to treatment for patients, we need to know where patients fall along this spectrum, especially if the treatments are different.”
Inaccurate classification has downstream implications particularly since treatments for moderate to severe OUD include medications such as buprenorphine, methadone, and naltrexone and addiction focused behavioral therapy. However, for individuals on long-term opioids for chronic pain without aberrant use, treatment options may be more focused towards pharmacologic and non-pharmacologic treatments for pain.
By increasing the accuracy of OUD prediction, doctors may be able to diagnose patients earlier in their disease course, tailor treatment to the patient, and hopefully prevent or reduce adverse events, such as opioid overdose. Further, Dr. Lagisetty said this will help clinicians decide if a patient will benefit more from seeing a pain specialist versus an addiction specialist, or vice versa.
Overall, the findings show the need for more robust methods of categorizing individuals with OUD, rather than relying on ICD coding alone.
“As a clinician, it really highlights how we need to be nuanced and Veteran-centered in our care because one size doesn’t fit all,” said Dr. Lagisetty.