RESUMEN
OBJECTIVE: Clinicians and policymakers have been wrestling with the appropriateness and safety of opioid therapy during the opioid crisis. Policy and clinical decisions have often been made without much current data on trends in drug use in patients with pain. Thus, we evaluated definitive urine drug test (UDT) results in patients being treated for pain to see if those taking their prescribed opioids were less likely to be positive for the primary illicit drugs currently driving overdose deaths: cocaine, heroin, fentanyl, and methamphetamine. DESIGN, SETTING, AND PATIENTS: A cross-sectional study of UDT results from January 1, 2015 to September 30, 2021, from 600,000 patient specimens submitted for testing by pain management specialists. INTERVENTIONS: UDT by liquid chromatography-tandem mass spectrometry as ordered by the treating clinician. MAIN OUTCOME MEASURES: Presence of other substances stratified by whether a patient's prescribed opioid was found. RESULTS: The presence of cocaine, heroin, fentanyl, and methamphetamine for the total population was low (<5 percent). Of the 347,092 patients prescribed opioids, 76 percent (n = 264,961) were positive on UDT for their prescribed opioid ("consistent"). Compared to patients without their prescribed opioid present ("inconsistent"), patients consistent with therapy were 54 percent (incidence rate ratio (IRR) 1.54, 95 percent confidence interval (CI) 1.47-1.59) less likely to be positive for cocaine, 47 percent [IRR 1.47, 95 percent CI 1.34-1.57] less likely to be positive for heroin, and 35 percent [IRR 1.35, 95 percent CI 1.24-1.45] less likely to be positive for methamphetamine, p < 0.001. Differences between the groups for fentanyl were not significant. CONCLUSIONS: Overall positivity rates for cocaine, heroin, fentanyl, and methamphetamine were low. Patients with prescribed opioid present were less likely to be positive for cocaine, heroin, or methamphetamine. Patterns of substance use within this pain management population should be used to inform ongoing policy decisions.