Increasing access to medications for opioid use disorder (MOUD) is an effective measure for addressing rising opioid-related deaths.1 However, the limited supply of providers eligible to prescribe MOUD has been implicated as a barrier to treating opioid use disorder (OUD) despite increases in insurance coverage and an expansion of train­ing.2,3 Beginning with the passage of the Comprehensive Addiction and Recovery Act of 2016, nurse practitioners, cer­tified nurse midwives, and physician assistants, collectively referred to as advanced practice providers (APPs), have had a pivotal and cost-effective role in addressing these service gaps related to treating OUD and other substance use disor­ders (SUD), particularly in rural areas.4–6 However, cur­rent requirements for obtaining the Drug Enforcement Admin­istration (DEA) X-waiver required to prescribe buprenorphine products for OUD (hereafter, buprenorphine) may limit the capacity of APPs to treat OUD effectively, especially relative to physicians. Currently, APPs must complete at least 3 times as much training as physicians (24 hours versus 8 hours) to be eligible to apply for an X-waiver. Additionally, APPs are re­quired to prescribe under the purview of a physician with an X-waiver in the majority of states.7

A shortage or perceived lack of support from fellow pro­viders has also been cited as a known barrier to APPs treat­ing OUD.8,9 For states in which APPs are required to be supervised, such as North Carolina, waivered APPs seeking to prescribe buprenorphine are currently required by federal law to be supervised by a waivered physician. This effectively requires 2 providers to be waivered for an APP to prescribe buprenorphine, which explicitly places the ability for APPs to treat OUD at the discretion of their supervising physician. Furthermore, if treatment for OUD is not included as part of the collaborative practice agreement between an APP and their physician supervisor, then OUD treatment is implicitly excluded from their scope of work. There is evidence that these additional training and prescribing oversight require­ments remain a deterrent to APPs obtaining waivers and prescribing buprenorphine.10 These restrictions are in place even though research demonstrates that restrictions on APPs’ scope of practice are not associated with improve­ments in the quality of care delivered and that nurse practi­tioners in particular are capable of producing similar health outcomes to physicians while incurring less cost and health care utilization.11,12

Given the established barriers that APPs face with regard to treating OUD, there may be differences in the OUD treat­ment practice patterns of APPs relative to physicians. A comparative description of these practice patterns may pro­vide evidence to inform policies addressing barriers faced by APPs and may also serve as a baseline measurement for use in assessing the impact of any policy changes. The objec­tive of the analysis presented here is to identify and describe patterns in general substance use disorder (SUD) screening, referral, and treatment practices between APPs and physi­cians in North Carolina. In particular, given the prevalence of OUD and disparities in treatment access, we present simi­larities and differences in OUD-specific treatment practices between physicians and APPs.


This study used data from a cross-sectional survey of physicians and APPs in North Carolina that was admin­istered as part of a larger mixed-method study examining buprenorphine-dispensing attitudes and awareness among a range of stakeholders.13–17 An advisory group of physi­cians and APPs treating OUD assisted with survey design. Confidential surveys were distributed in August 2020 and September 2020 via email to all physicians, physician assis­tants, nurse practitioners, and certified nurse midwives licensed in North Carolina using mailing lists provided by the North Carolina Medical Board and the North Carolina Board of Nursing (28,853 total emails). From that list, 11,149 providers opened the email (38.6%), and 709 (2.46%) opened the survey. This response rate is in line with the average response rate for email marketing campaigns in the United States (2.5%), especially given that some email addresses on the provided lists may no longer be in use.18 Respondents who began but did not reach the end of the survey were excluded (n = 21). QualtricsTM was used to collect responses from the participants who clicked on the survey link in the email and reached the end of the survey (N = 332). This study was found to be exempt from review by the institutional review board of the University of North Carolina at Chapel Hill.

The survey included questions about licensure type, which we used to group providers as physicians or APPs. Physicians included medical doctors (MD) and doctors of osteopathic medicine (DO). Providers were grouped based on the differential requirements for obtaining the X-waiver. The survey also collected age and self-reported gender as basic demographic characteristics. At the suggestion of the provider advisory group, the survey began with a selection of general questions about practices related to SUD, includ­ing if and when providers screen for SUD, if and where they refer for SUD treatment, and if they currently treat SUD. OUD-specific questions followed the general SUD ques­tions, including whether they currently treat OUD, if they have completed an X-waiver training, if they received their waiver following the training, and if they currently prescribe buprenorphine. See Appendix 1 for the survey questions in their entirety.

We calculated chi-squared tests with a cutoff of p < .05 to determine significance for differences between physicians and APPs. For each question, participants who had miss­ing data were re-coded into a “Did Not Respond” category. Analyses and data management were conducted using SAS 9.4 (SAS Institute, Cary, NC).


Sample characteristics are shown in Table 1. Of the 332 respondents who completed the survey, 48.8% (n = 162) were APPs and 51.2% (n = 170) were physicians. Our sam­ple of APPs was primarily comprised of nurse practitioners (73.5%), followed by physician assistants (21.0%) and certi­fied nurse-midwives (< 1%). Physicians were almost entirely medical doctors (95.9%) with the remainder being doctors of osteopathic medicine (4.1%). The majority of APPs were aged 35–44 years (29.6%), while the majority of physicians were aged either 45–54 or 55–56 years (22.4% for both categories). This difference in the age distribution was sig­nificant (p = .0008). APPs predominantly self-reported their gender as female/feminine (75.9%), while most physicians reported their gender as male/masculine (52.4%). A por­tion of both physicians (2.94%) and APPs (2.47%) reported a gender other than masculine or feminine, such as nonbi­nary/genderqueer. This difference in gender identity by pro­vider type was significant (p < .0001).

Table 1.Sample Demographics
Total Physicians APPs
N (%) N (%) N (%) p-value
Total 332 – 170 (51.20) 162 (48.80)
MD 163 (49.10) 163 (95.88)
DO 7 (2.11) 7 (4.12)
NP 119 (35.84) 119 (73.46)
PA 34 (10.24) 34 (20.99)
CNM 9 (7.65) 9 (0.05)
Age .0008
25-29 10 (3.01) 1 (0.59) 9 (5.56)
30-34 32 (9.64) 16 (9.41) 16 (9.88)
35-44 76 (22.89) 28 (16.47) 48 (29.63)
45-54 77 (23.19) 38 (22.35) 39 (24.07)
55-64 63 (18.98) 38 (22.35) 25 (15.43)
65-74 41 (12.35) 30 (17.65) 11 (6.79)
75+ 5 (1.51) 4 (2.35) 1 (0.62)
Did not respond 28 (8.43) 15 (8.82) 13 (8.02)
Gender identity < .0001
Man/masculine 111 (33.43) 89 (52.35) 22 (13.58)
Woman/feminine 184 (55.42) 61 (35.88) 123 (75.93)
Other* 9 (2.71) 5 (2.94) 4 (2.47)
Did not respond 28 (8.43) 15 (8.82) 13 (8.02)
Patient rurality .0192
Mostly urban 47 (14.16) 28 (16.47) 19 (11.73)
Mostly rural 82 (24.7) 30 (17.65) 52 (32.10)
Mixed urban/rural 177 (53.31) 99 (58.24) 78 (48.15)
Did not respond 26 (7.83) 13 (7.65) 13 (8.02)

Demographics of physician and advanced practice providers in North Carolina responding to an email survey about substance use treatment practices. *The category ‘Other’ included non-binary/genderqueer, other, and prefer not to say.
CNM = Certified Nursing Midwife; DO = Doctor of Osteopathic Medicine; MD = Medical Doctor; NP = Nurse Practitioner; PA = Physician Assistant

APPs were significantly more likely than physicians to screen for SUDs (83.3% versus 69.4%, p = .0045) and to refer for SUD treatment (68.5% versus 48.8%, p = .0017; Table 2). There was no difference between APPs and phy­sicians regarding currently treating SUDs (37.7% versus 40.0%, p = .9083); however, of those who treat SUD, APPs are significantly less likely to treat OUD (54.1% versus 72.1%, p = .0343). A similar proportion of APPs and physi­cians completed X-waiver training (23.5% versus 29.4%, p = .6491), and there was no difference in the proportion of those who received their X-waiver after completing training (84.6% versus 96.0%, p = .4717). Of those who received the X-waiver, APPs and physicians were equally likely to be currently prescribing buprenorphine (75.8% versus 77.1%, p = .8900).

Table 2.Substance Use Treatment Practices by Provider Type
Total Physicians APPs
N (%) N (%) N (%) p-value
Total 332 (100) 170 (51.2) 162 (48.8)
Screen for SUD .0045
Did not respond 2 (0.6) 1 (0.59) 1 (0.62)
No 77 (23.19) 51 (30) 26 (16.05)
Yes 253 (76.2) 118 (69.41) 135 (83.33)
If YES, when do you screen for SUD?
First intake 163 (64.43) 80 (67.8) 83 (61.48) .2952
Based on family history 48 (18.97) 22 (18.64) 26 (19.26) .9009
Annual/well-person exam 85 (33.6) 38 (32.2) 47 (34.81) .6609
Every visit 81 (32.02) 37 (31.36) 44 (32.59) .8334
When they mention something that raises a concern 150 (59.29) 68 (57.63) 82 (60.74) .6150
Before prescribing a controlled substance 138 (54.55) 64 (54.24) 74 (54.81) .9267
Do you refer patients for SUD? .0017
Did not respond 5 (1.51) 4 (2.35) 1 (0.62)
It depends 77 (23.19) 46 (27.06) 31 (19.14)
No 56 (16.87) 37 (21.76) 19 (11.73)
Yes 194 (58.43) 83 (48.82) 111 (68.52)
If NO, why not? .7120
Did not respond 6 (10.71) 3 (8.11) 3 (15.79)
Do not refer for any SUD-related services 43 (76.79) 29 (78.38) 14 (73.68)
No SUD treatment options in my area 7 (12.5) 5 (13.51) 2 (10.53)
If YES, where?
Psychiatrist 119 (61.34) 53 (63.86) 66 (59.46) .5339
Behavioral health provider (e.g., Psychologist) 154 (79.38) 66 (79.52) 88 (79.28) .9676
Comprehensive treatment program 92 (47.42) 39 (46.99) 53 (47.75) .9165
Inpatient treatment 80 (41.24) 36 (43.37) 44 (39.64) .6012
Outpatient treatment 128 (65.98) 56 (67.47) 72 (64.86) .7048
Recovery programs (e.g., AA/NA) 120 (61.86) 48 (57.83) 72 (64.86) .3183
Do you currently treat SUD? .9083
Did not respond 10 (3.01) 5 (2.94) 5 (3.09)
No 193 (58.13) 97 (57.06) 96 (59.26)
Yes 129 (38.86) 68 (40) 61 (37.65)
If YES, do you treat OUD? .0343
No 47 (36.43) 19 (27.94) 28 (45.9)
Yes 82 (63.57) 49 (72.06) 33 (54.1)
Completed waiver training .6491
Did not respond 13 (3.92) 6 (3.53) 7 (4.32)
No 204 (61.54) 103 (60.59) 101 (62.35)
Started, but not completed 15 (4.52) 6 (3.53) 9 (5.56)
Yes 88 (26.51) 50 (29.41) 38 (23.46)
Don't know what this is 12 (3.61) 6 (2.94) 7 (4.32)
Waivered after training .4717
Did not respond 1 (1.14) 0 (0) 1 (2.63)
No 3 (3.41) 1 (2) 2 (5.13)
No, but intend to 2 (2.27) 1 (2) 1 (2.56)
Yes 81 (92.05) 48 (96) 33 (84.62)
Other 1 (1.14) 0 (0) 1 (2.63)
Are you currently prescribing buprenorphine for OUD? .8900
Yes 62 (76.54) 37 (77.08) 25 (75.76)
No 19 (23.46) 11 (22.92) 8 (24.24)

Substance use disorder practice patterns by provider type for providers in North Carolina. A Fisher’s exact test was used due to unmet assumptions for Chi-Square.
AA = Alcoholics Anonymous; APP = Advanced Practice Provider (Nurse Practitioner, Physician Assistant, or Certified Nurse Midwife); OUD = Opioid Use Disorder; NA = Narcotics Anonymous; SUD = Substance Use Disorder


APPs face several well-documented barriers to treating OUD relative to physicians. This study sought to describe differences in practice patterns between APPs and physi­cians to generate evidence for future policy development. Among respondents to our survey, APPs screen for SUDs more often, are more likely to refer patients for treatment, and are less likely to report treating OUD than physicians. APPs treat SUDs, complete X-waiver training, receive X-waivers, and prescribe buprenorphine as frequently as physicians. Our findings suggest that while there are some differences, APPs and physicians are generally similar in their practice patterns.

This study has methodological limitations worth con­sidering. First, this study was limited to providers in North Carolina, and thus may not generalize beyond North Carolina. Second, the sample of respondents was a purposive sample and may not be representative of all providers. Third, this study was a cross-sectional study using self-report survey data. As such, future research is needed to examine how these patterns may vary over time. Additionally, the use of self-report is important for understanding provider perspec­tives; however, it does introduce the potential for positive responder bias in which providers may attempt to appear more favorable in their responses. Future studies should seek to examine buprenorphine prescription patterns using prescription claims data. Finally, although the demographic differences we found between the provider types are consis­tent with national statistics, we acknowledge that it is possi­ble that they contribute to the differences in SUD treatment practices.10

Our findings—more frequent screening, more referrals, and less treating OUD among APPs—may be attributable to a combination of regulations and treatment approaches. We believe the differences we found in screening and referring patients for SUDs to be reflective of the supervision require­ments for APPs in North Carolina. Prior work has shown that nurse practitioners in states with supervision requirements refer patients in general practice more than nurse practi­tioners in states without those requirements.11 It seems reasonable that a similar difference would be found between APPs and physicians given that physicians have no such requirements. Future work to elicit the reasoning behind the rates of screening and referring is needed. The difference in treating OUD, which was not asked exclusively of those who received waivers for buprenorphine, may be reflective of practice differences in the use of other evidence-based interventions for OUD that do not rely on pharmacotherapy, such as contingency management.19 Future research is needed regarding other treatment options currently being used in order to understand this difference.


APPs are in a position to help address the impact of opi­oid-related deaths in this country and are fully capable of increasing MOUD prescribing in areas that currently lack and sorely need more treatment options. In North Carolina, the SUD- and OUD-specific treatment practice patterns are similar between APPs and physicians. Future research is needed to better understand the drivers of the differences identified in screening, referring, and treating OUD. The sim­ilar proportion of APPs and physicians who have completed X-waiver training, received an X-waiver, and are prescribing buprenorphine suggests that policy efforts to reduce barri­ers for APPs may result in a robust workforce of SUD treat­ment providers.

Financial support

This research was funded jointly by the University of North Carolina Eshelman School of Pharmacy and Research Triangle International, with a sub-award to Mountain Area Health Education Center. Portions of B.O. and C.H.'s salaries were funded by the sub-award during data collection for this study. P. H. is supported by an Eshelman Fellowship from the Eshelman School of Pharmacy, University of North Carolina at Chapel Hill. The funders had no role in the design, data collection, analysis, or interpretation of this study.


We wish to acknowledge Dr. Delesha Carpenter for her role in the development of the study design and obtaining funding for the project. With input from a provider advisory team, B.O. and C.H. proposed, developed, and implemented the survey used in the current study. PH conducted the analysis and lead in writing this manuscript. MR and CF provided critical feedback and aided in the interpretation of results. All authors contributed to the writing and revising of the manuscript.

Disclosure of interests

None of the authors report a conflict of interest.