Variation in Nevada Primary Care Clinicians' Use of Urine Drug Testing to Mitigate Opioid Harm.

Introduction: The prescription opioid epidemic led to federal, state, and health system guidelines and policies aimed at mitigating opioid misuse, including presumptive urine drug testing (UDT). This study identifies whether a difference exists in UDT use among different primary care medical license types.Methods: The study used January 2017-April 2018 Nevada Medicaid pharmacy and professional claims data to examine presumptive UDTs. We examined correlations between UDTs and clinician characteristics (medical li-cense type, urban/rural status, care setting) along with clinician-level measures of patient mix characteristics (proportions of patients with behavioral health diagnoses, early refills). Adjusted odds ratios (AORs) and pre-dicted probabilities (PPs) from a logistic regression with a binomial distribution are reported. The analysis included 677 primary care clinicians (medical doctors [MD], physician assistants [PA], nurse practitioners [NP]).Results: Of those in the study, 85.1 % of clinicians did not order any presumptive UDTs. NPs had the highest proportion of UDT use (21.2 % of NPs), followed by PAs (20.0 % of PAs), and MDs (11.4 % of MDs). Adjusted analyses showed that being a PA or NP was associated with higher odds of UDT (PA: AOR: 3.6; 95 % CI: 3.1-4.1; NP: AOR: 2.5; 95 % CI: 2.2-2.8) compared to being an MD. PAs had the highest PP for ordering UDTs (2.1 %, 95 % CI: 0.5 %-8.4 %). Among clinicians who ordered UDTs, midlevel clinicians had higher mean and median UDT use (PA and NP mean: 24.3 % vs. MDs: 19.4 %; PA and NP median: 17.7 % vs. MDs: 12.5 %).Conclusion: In Nevada Medicaid, UDTs are concentrated among 15 % of primary care clinicians who are frequently non-MDs. More research should include PAs and NPs when examining clinician variation in mitigating opioid misuse.