Recipients of opioid prescriptions with a longer duration and that have a higher tablet quantity may be at risk of progressing from acute opioid use to chronic opioid use, according to the results of a study published in PLOS Digital Health.
Investigators from Stanford University sourced data for this study from the Digital Health Cooperative Research Center. A mixture of Medicaid and fee-for-service claims collected between 2015 and 2019 in 6 states were evaluated for evidence of opioid prescriptions among individuals who were opioid naive, which was defined as an individual having had no opioid prescription in the 2 months prior to the prescription of interest. Risk for chronic opioid use, which was defined as opioid use for the subsequent 9 months, was assessed on the basis of patient and prescription characteristics using a machine learning approach. The patient population was split into training (80%) and validation (20%) cohorts.
The study population consisted of 180,000 individuals, the mean age of which was 39.18 (standard deviation [SD], 13.36) years; 63.3% were women, and 3.8% were documented as having substance use disorder. Acute opioid use was reported among 70.1% of the patient population, and chronic opioid use was reported among 29.9%. Stratified by opioid use status, all demographic characteristics differed between groups (all P <.0001) except for gender and median household income.
The daily mean morphine milligram equivalent (MME) for the initial opioid prescription was 29.40 (SD, 26.68) mg, the prescription length was 12.94 (SD, 14.67) days, the number of tablets per prescription was 32.81 (SD, 40.53), and most (98.90%) prescriptions were for short-acting opioids. Patients with chronic opioid use were prescribed higher doses (mean, 32.98 vs 27.87 MME; P <.001), a longer duration of treatment (mean, 16.15 vs 11.57 days; P <.001), more pills (mean, 46.94 vs 26.77 tablets; P <.001), and more were given long-acting opioids (2.55% vs 0.47%; P <.001) compared with patients with acute opioid use, respectively.
Patients who progressed to chronic opioid use were more likely to receive prescriptions for hydromorphone, methadone, long-acting morphine, long-acting oxycodone, and tramadol (all P ≤.0003) compared with patients with acute opioid use.
The best fit predictive model had an area-under-the-receiver operating characteristic curve of 0.80, accuracy of 0.75, precision of 0.57, and recall of 0.65. This model found that the most robust predictors for progressing from acute to chronic opioid use were tablets per prescription, low back pain, emergency department claim, hypertension, prescription claim from an independent laboratory, gabapentin, elective claim, prescription length, chronic pulmonary disease, age, tramadol, depression, oxycodone, tablet opioid formulation, and urgent care claim.
This study may have been limited by only selecting a subset of comorbidities as potential predictors in the models.
Investigators of this study found that the risk for progressing from acute to chronic opioid use was increased for opioid-naive individuals who received opioid prescriptions of a longer duration and with more tablets. The study authors commented, “The ability to identify high-risk prescription features is important for policymakers and payers who currently target strategies for addressing the contribution of prescription opioids to the opioid epidemic.”
Disclosure: Multiple study authors declared affiliations with biotech, pharmaceutical, and/or device companies. Please see the original reference for a full list of authors’ disclosures.
This article originally appeared on Clinical Pain Advisor