Logistic Regression Model for Likelihood of Prescribing Opioids for Chronic Pain to Fewer Patients at 3 Months Relative to 12 Months Ago*
| Predictors | Adjusted Odds Ratio (95% CI) | P value |
|---|---|---|
| Number of patients prescribed opioids for chronic pain in last 3 months | 1.36 (1.23–1.49) | .001 |
| ADF knowledge score† | 0.99 (0.90–1.09) | .812 |
| Exposure to opioid educationalinformation in last 12 months‡ | 1.19 (1.06–1.32) | .002 |
| Recall of opioid promotion‖ | 1.08 (0.90–1.22) | .240 |
| Health care providers | ||
| SPs | 0.61 (0.43–0.87) | .006 |
| PAs | 0.69 (0.48–1.01) | .056 |
| NPs | 0.68 (0.47–0.99) | .042 |
| PCPs | Reference | |
| Race/ethnicity¶Hispanic | 1.27 (0.72–2.23) | .407 |
| Asian | 1.27 (0.86–1.88) | .228 |
| Black | 1.08 (0.46–2.53) | .861 |
| Other | 1.12 (0.60–2.08) | .726 |
| White | Reference | |
| Practice type | ||
| Clinics/urgent care centers | 0.84 (0.55–1.29) | .433 |
| Non-federal government clinics | 1.97 (1.12–3.49) | .021 |
| Federal government clinics | 0.67 (0.34–1.31) | .237 |
| Health maintenance organizations | 0.82 (0.45–1.49) | .512 |
| Community health centers | 0.87 (0.56–1.35) | .533 |
| Other | 0.70 (0.43–1.15) | .158 |
| Solo/group practice | Reference | |
| Sex | ||
| Male | 1.23 (0.93–1.63) | .140 |
| Female | Reference | |
| Age | 1.01 (0.99–1.02) | .062 |
ADF, abuse-deterrent formulations; CI, confidence interval; NPs, nurse practitioners; PAs, physician assistants; PCPs, primary care physicians; SPs, specialists.
↵* The outcome variable is based on responses to the question: Is the number of patients you prescribed opioids for chronic pain in the past 3 months relative to 12 months ago, the same, less, or more. Response options same and more were combined and compared to less.
↵† Correct responses to 5 ADF questions were summed to create a ADF knowledge score.
↵‡ Exposure to opioid educational information was created by summing “yes” responses for each source of opioid educational information.
↵‖ “Yes” responses to questions about seeing prescription drug promotions in the last 6 months related non-opioid analgesics, medication-assisted treatment for opioid deterrence, and opioid reversal agents were combined to create an overall recall measure of prescription drug promotions.
↵¶ Asian, Black, Other, and White race categories are Non-Hispanic.