Your privacy, your choice

We use essential cookies to make sure the site can function. We also use optional cookies for advertising, personalisation of content, usage analysis, and social media.

By accepting optional cookies, you consent to the processing of your personal data - including transfers to third parties. Some third parties are outside of the European Economic Area, with varying standards of data protection.

See our privacy policy for more information on the use of your personal data.

for further information and to change your choices.

You are viewing the site in preview mode

Skip to main content
Download PDF

Impact of an opioid use disorder medication implementation intervention on hospitalization and emergency department utilization in the Veterans Health Administration

BMC Psychiatry volume 25, Article number: 288 (2025) Cite this article

Abstract

Background

It is important to evaluate how medication for opioid use disorder (MOUD) implementation interventions impact downstream outcomes, however little is known about impact on hospitalization and emergency department (ED) utilization. Stepped Care for Opioid Use Disorder Train the Trainer (SCOUTT) is a national United States Veterans Health Administration (VHA) effort initially implemented at 18 facilities that increased MOUD receipt in primary care, mental health, and pain clinics. This evaluation assessed SCOUTT’s impact on hospitalization and ED utilization.

Methods

This evaluation used a controlled interrupted time series analysis. We extracted electronic health record data for patients with OUD and ≥ 1 visit in an intervention clinic (N = 35) or matched comparison clinic (N = 35) in the post-implementation year. We examined monthly measures of hospitalization and ED utilization in the pre-implementation (9/1/2017-8/31/2018) and post-implementation (9/1/2018-8/31/2019) years. Segmented regression models estimated pre-post immediate and trend changes in intervention relative to comparison clinics, adjusting for pre-implementation trends and patient characteristics. Sensitivity analyses were conducted among patients with ≥ 1 visit in both the pre-/post-implementation years, and post-hoc secondary analyses were conducted among patients with OUD and ≥ 1 other SUD vs. OUD only.

Results

Patients with OUD in both intervention (N = 7,488) and comparison (N = 7,558) clinics had a mean age of 53 years, and the majority were male, White, and not married. During the pre-implementation period, hospitalization and ED utilization increased over time in both intervention and comparison clinics; during the post-implementation period, hospitalization and ED utilization decreased over time in intervention clinics and remained stable in comparison clinics. There was no significant difference in pre-post changes between intervention and comparison clinics for most analyses. In sensitivity analyses the pre-to-post decrease in monthly trend for ED visits was larger in intervention clinics and, in secondary analyses, the pre-to-post decrease in monthly trend for hospitalizations was larger in intervention clinics among patients with OUD and ≥ 1 other SUD.

Conclusions

This evaluation did not find evidence that SCOUTT substantially impacted hospitalization or ED utilization relative to comparison clinics, though there may have been positive impacts for patients with longer engagement in SCOUTT clinics and patients with OUD and ≥ 1 other SUD.

Peer Review reports

Introduction

There are multiple effective medications to treat opioid use disorder (MOUD), and receipt of these medications reduces risk of overdose and death [1,2,3]. However, in the United States the majority of people with OUD do not receive MOUD [4]. In response to this treatment gap and an ongoing overdose crisis [5, 6], a variety of new care models and implementation interventions have been developed to increase MOUD provision across multiple U.S. healthcare settings, including statewide models and national healthcare system initiatives [7,8,9,10]. Prior research suggests that many U.S. patients experience barriers to accessing care in substance use disorder (SUD) specialty settings [11, 12], therefore these interventions often focus on increasing provision of buprenorphine and naltrexone in general medical settings such as primary care.

To fully evaluate the impact of MOUD implementation interventions, it is important to examine whether these interventions affect downstream outcomes—such as hospitalization and emergency department (ED) utilization—in addition to MOUD receipt. Many studies have examined the patient-level association between MOUD receipt and hospitalization and/or ED utilization among patients with OUD, with most finding that receiving MOUD is associated with lower hospitalization and ED utilization [13,14,15,16,17,18,19]. However, fewer studies have examined whether MOUD implementation interventions have impacted these outcomes at a broader population level in participating clinics (rather than examining if receiving MOUD impacted outcomes for a given patient) [20,21,22]. This type of analysis provides valuable insight for clinical and healthcare system leaders concerned with improving population-level outcomes through implementation and quality improvement interventions. Whether an MOUD implementation intervention impacts population-level utilization outcomes likely depends on multiple factors, including how much it increases MOUD receipt, among which patients (i.e., among patients who would have been likely to have hospitalization and ED utilization without MOUD), and to what extent patients initiated on MOUD are retained in treatment.

The few studies that have examined this question have found mixed results [20,21,22]. One observational study of buprenorphine for OUD implemented in primary care found that patients with OUD in intervention facilities had fewer hospitalizations than those in non-intervention facilities but found no difference in ED utilization [20]. Another observational study of addiction pharmacotherapy and recovery coaching integrated into primary care found that patients with OUD in intervention facilities had fewer inpatient days and ED visits than those in non-intervention facilities, but found no difference in mean number of hospitalizations [21]. A cluster-randomized implementation-effectiveness trial of nurse care management for OUD in primary care found no differences in days of acute care utilization (hospitalizations, ED, and/or urgent care) between patients in intervention and non-intervention facilities [22].

The Veterans Health Administration (VHA) is the largest provider of OUD treatment in the U.S [23]. The Stepped Care for Opioid Use Disorder Train the Trainer (SCOUTT) initiative is a national VHA effort to implement MOUD provision in primary care, mental health, and pain care settings [24]. A prior evaluation found evidence that the SCOUTT initiative increased MOUD receipt among patients with OUD in intervention vs. comparison clinics [25]. However, it is unknown whether the SCOUTT initiative affected downstream utilization outcomes, and to our knowledge no evaluations of VHA MOUD national initiatives have assessed impact on these outcomes. This evaluation aimed to assess the impact of the SCOUTT initiative on hospitalizations and ED utilization among patients with OUD.

Methods

Setting and evaluation design

SCOUTT is a national VHA initiative that was initially implemented in 35 intervention clinics (primary care, mental health, and pain) at 18 VHA facilities across the U.S., one from each regional network. SCOUTT was initiated on September 1, 2018, during an in-person conference attended by teams from the 18 VHA facilities. The conference, subsequent trainings, and ongoing external practice facilitation were aimed to help the teams implement evidence-based models of MOUD care within the three types of clinics in their facilities. SCOUTT implementation is described in more detail in prior publications [24, 25].

The present evaluation used a controlled interrupted time series design; comparing an intervention group to a control group allowed us to examine whether changes in outcomes from pre- to post-implementation were likely due to the intervention as opposed to another contemporaneous factor [26]. We examined whether SCOUTT implementation impacted trends in hospitalization and ED utilization for patients seen in intervention clinics relative to patients seen in comparison clinics. Implementation began 9/1/2018; the pre-implementation period was defined as the year prior to implementation (9/1/2017-8/31/2018) and the post-implementation period was defined as the year following implementation (9/1/2018-8/31/2019). Thirty-five matched comparison clinics in other VHA facilities were selected based on their similarity to intervention clinics with respect to clinic-level metrics using nearest-neighbor matching, including volume of patients with OUD, pre-implementation MOUD prescribing trends, and facility complexity (see supplemental materials in Hawkins et al., 2021 for a detailed description of comparison clinic selection) [25]. Per VHA policy (VHA Program Guide 1200.21) this evaluation was designated as a quality improvement project and received a non-research determination from the VHA Office of Mental Health and Suicide Prevention, and therefore did not require institutional review board oversight.

Data source and sample

Electronic health record data including patient demographic information, diagnoses, healthcare utilization, and medications were extracted from the VHA Corporate Data Warehouse. The sample included patients who were ≥ 18 years old, had ≥ 1 visit in an intervention or comparison clinic in the post-implementation period, and had an OUD diagnosis documented during either the pre- or post-implementation periods assessed via International Classification of Diseases 10th Revision codes. Because SCOUTT reached mostly primary care patients, and patients who have OUD identified in primary care may be less likely to have repeated visits with confirmatory diagnostic codes than in other clinic settings, we used a broad definition of OUD requiring ≥ 1 diagnostic code to ensure that all patients potentially eligible to receive MOUD were included (aligned with prior studies of MOUD in primary care) [27, 28].

Measures

Outcomes

Primary outcomes included the monthly proportion of patients with hospitalization (proportion with ≥ 1 VHA inpatient hospitalization during each month) and the monthly proportion of patients with ED utilization (proportion with ≥ 1 VHA ED visit during each month). We secondarily examined different types of hospitalization (psychiatric or medical/surgical) using patient treatment file (PTF) codes to classify hospitalization type, as well as counts of inpatient hospitalizations and ED visits during each month.

Covariates

Baseline patient characteristics were measured during the pre-implementation period; these included age, race, ethnicity, sex, VHA service-connected disability rating ≥ 50%, housing instability, marital status, anxiety, posttraumatic stress disorder, depression, serious mental illness (bipolar, psychoses, and/or schizophrenia), non-opioid/non-tobacco drug use disorder, alcohol use disorder, and Charlson comorbidity score [29]. Clinic type (primary care, mental health, pain) was adjusted for to account for practice differences across these settings. To account for whether patients had an opportunity to be impacted by the intervention and for receipt of other SUD services, two binary, time-varying (monthly) covariates were also included: (1) having ≥ 1 visit in the intervention or comparison clinic, and (2) having ≥ 1 visit to a VHA SUD specialty clinic.

Analyses

We described patient characteristics across intervention and comparison clinics and described outcomes during the year prior to and year following implementation across intervention and comparison clinics. For the controlled interrupted time series analyses, we used segmented regression models to estimate changes in monthly outcomes from the pre- to the post-implementation period (immediate level change and trend change) in intervention and comparison facilities. Models included terms for month since the beginning of pre-implementation period (months 1-24) to estimate the pre-implementation trend, month since implementation (months 13-24) to estimate change in trend after implementation, and a pre/post-implementation indicator (0 = pre [months 1–12], 1 = post [months 13–24]) to estimate immediate change after implementation, adjusted for all covariates described above [26, 30]. Models comparing the difference in pre-post changes between intervention and comparison clinics additionally included interaction terms between facility type and month since the beginning of pre-implementation, month since implementation, and pre-/post-implementation indicator. We used generalized estimating equations models to account for correlation among multiple observations for the same patient, with a fixed effect for facility to additionally account for facility-level correlation. Models examining binary outcomes used a logit link, and models for analyses examining secondary count outcomes used a negative binomial link. Models were fit with robust standard errors and an autoregressive (AR1) correlation structure. We graphed the marginal adjusted prevalence of outcomes among patients in intervention and comparison clinics over each month.

To determine if results were different for patients with longer engagement in SCOUTT clinics, we conducted a sensitivity analysis that repeated interrupted time series analyses for primary outcomes among a subsample of patients with ≥ 1 visit during the pre-implementation period and ≥ 1 visit during the post-implementation period. Because a prior analysis found that the SCOUTT initiative may not have increased MOUD provision among patients with co-occurring non-opioid SUDs [31], we also conducted a post-hoc secondary analysis that repeated the main analyses among patients with OUD and ≥ 1 other SUD and among patients with OUD only to determine if results differed between these groups. Analyses were conducted with Stata version 18 software [32].

Results

Sample characteristics and outcome prevalence

Patient characteristics in intervention (N = 7,488) and comparison (N = 7,558) clinics are presented in Table 1. In both intervention and comparison clinics, mean age was about 53 years, and the majority of patients were male, White, and not married. Over half had a VHA service-connected disability rating ≥ 50%, and about one-quarter had housing instability. Non-opioid SUD diagnoses, anxiety, serious mental illness and Charlson comorbidities were slightly less common in intervention than comparison clinics, while posttraumatic stress disorder was slightly more common. In intervention facilities, 56% of patients were seen at primary care clinics, 31% at mental health clinics, and 14% at pain clinics, while in comparison facilities 48% of patients were seen at primary care clinics, 35% at mental health clinics, and 17% at pain clinics. As reported in the prior evaluation, 6% of patients in intervention clinics and 4% of patients in comparison clinics received MOUD during the pre-implementation year, and 14% of patients in intervention clinics and 6% of patients in comparison clinics received MOUD during the post-implementation year [25].

Table 1 Characteristics of VA patients with OUD seen in SCOUTT intervention or comparison clinics
Full size table

Descriptive summaries of outcomes during the year prior to and the year following implementation are presented in Table 2. During the year prior to implementation, over one-quarter of patients had ≥ 1 hospitalization (28.2% in intervention clinics and 27.5% in comparison clinics), and nearly half of patients had ≥ 1 ED visit (49.9% in intervention clinics and 48.8% in comparison clinics). During the year following implementation, about one-third of patients had ≥ 1 hospitalization (32.8% in intervention clinics and 35.1% in comparison clinics), and over half of patients had ≥ 1 ED visit (54.3% in intervention clinics and 53.5% in comparison clinics).

Table 2 Descriptive summaries of outcomes during year prior to and year following implementation
Full size table

Interrupted time series results

Hospitalizations

The adjusted monthly proportion of patients with ≥ 1 hospitalization across intervention and comparison clinics is presented in Fig. 1. In intervention clinics, this proportion increased over time during the pre-implementation period (from 3.7% in month 1 to 5.4% in month 12; adjusted odds ratio [aOR] for pre-implementation trend: 1.03, 95% confidence interval [CI]: 1.01–1.04, p < 0.001), and decreased over time during the post-implementation period (from 5.3% in month 13 to 5.1% in month 24; the monthly trend changed significantly from pre- to post-implementation (aOR 0.98, 95% CI: 0.96–0.99, p < 0.001). In comparison clinics, this proportion increased over time during the pre-implementation period (from 3.6% in month 1 to 4.9% in month 12; aOR for pre-implementation trend 1.02, 95% CI: 1.01–1.03, p < 0.001), and remained relatively stable during the post-implementation period (5.3% in month 13 to 5.7% in month 24; the monthly trend did not significantly change from pre- to post-implementation (aOR 0.99, 95% CI: 0.98-1.00, p = 0.148).

In models comparing intervention and comparison clinics, the change in the monthly proportion with hospitalization immediately after implementation (from the last month of the pre-period to the first month of the post-period) did not significantly differ between intervention and comparison clinics (aOR 0.93, 95% CI: 0.82–1.05, p = 0.235), and the change in the monthly trend from pre- to post-implementation did not significantly differ between intervention and comparison clinics (aOR 0.98, 95% CI: 0.96-1.00, p = 0.068). Results from analyses separately examining psychiatric and medical/surgical hospitalizations, and monthly counts of hospitalizations, are presented in Appendix A; controlled interrupted time series findings for all secondary hospitalization outcomes were similar to the main analyses.

Results from sensitivity analyses repeating the main analyses among patients with ≥ 1 clinic visit during the pre-implementation period and ≥ 1 visit during the post-implementation period are presented in Appendix B; controlled interrupted time series findings for hospitalization were similar to the main analyses. Results from post-hoc secondary analyses repeating the main analyses among patients with OUD and ≥ 1 other SUD and among patients with OUD only are presented in Appendix C; controlled interrupted time series findings for hospitalization were similar to the main analyses except that among patients with OUD and ≥ 1 other SUD, the decrease in the monthly trend from pre- to post-implementation was larger in intervention vs. comparison clinics (aOR: 0.97, 95% CI: 0.95–0.99, p = 0.028).

Fig. 1
figure 1

Adjusted monthly proportion of patients with ≥ 1 hospitalization in intervention and comparison clinics; the vertical line represents the start of implementation

Full size image

ED utilization

The adjusted monthly proportion of patients with ≥ 1 ED visit across intervention and comparison clinics is presented in Fig. 2. In intervention clinics, this proportion increased over time during the pre-implementation period (from 9.3% in month 1 to 11.2% in month 12; aOR for pre-implementation trend: 1.01, 95% CI: 1.01–1.02, p < 0.001), and decreased over time during the post-implementation period (from 11.1% in month 13 to 10.5% in month 24; the monthly trend changed significantly from pre- to post-implementation (aOR 0.98, 95% CI: 0.97–0.99, p = 0.001). In comparison clinics, this proportion increased over time during the pre-implementation period (from 9.6% in month 1 to 11.0% in month 12; aOR for pre-implementation trend 1.01, 95% CI: 1.00-1.01, p = 0.024), and remained relatively stable during the post-implementation period (from 11.2% in month 13 to 11.0% in month 24; the monthly trend did not significantly change from pre- to post-implementation (aOR 0.99, 95% CI: 0.98-1.00, p = 0.104).

In models comparing intervention to comparison clinics, the change in the monthly proportion with ED utilization immediately after implementation (from the last month of the pre-period to the first month of the post-period) did not significantly differ between intervention and comparison clinics (aOR 0.98, 95% CI: 0.90–1.07, p = 0.643), and the change in the monthly trend from pre- to post-implementation did not significantly differ between intervention and comparison clinics (aOR 0.99, 95% CI: 0.98-1.00, p = 0.172). Results from analyses examining monthly counts of ED visits are presented in Appendix A; controlled interrupted time series findings for this secondary outcome were similar to the main analyses.

Results from sensitivity analyses repeating the main analyses among patients with ≥ 1 clinic visit during the pre-implementation period and ≥ 1 visit during the post-implementation period are presented in Appendix B; controlled interrupted time series findings for ED visits were mostly similar to the main analyses, except that the decrease in the monthly trend from pre- to post-implementation was larger in intervention vs. comparison clinics (aOR: 0.98, 95% CI: 0.97–0.99, p = 0.037). Results from post-hoc secondary analyses repeating the main analyses among patients with OUD and ≥ 1 other SUD and among patients with OUD only are presented in Appendix C; controlled interrupted time series findings for ED visits were similar to the main analyses.

Fig. 2
figure 2

Adjusted monthly proportion of patients with ≥ 1 emergency department visit in intervention and comparison clinics; the vertical line represents the start of implementation

Full size image

Discussion

The present evaluation did not find evidence that the SCOUTT initiative substantially impacted hospitalizations or ED utilization among patients with OUD. In sensitivity analyses, the change in monthly trend from pre- to post-implementation for ED visits was larger in intervention vs. comparison clinics among the subsample of patients who had ≥ 1 visit during the pre-implementation period and ≥ 1 visit during the post-implementation period, suggesting that there may have been some positive impact on this outcome for patients with longer engagement in SCOUTT clinics. These findings align with a recent implementation-effectiveness trial of primary care-based nurse care management for MOUD, which did not find differences in acute care utilization between intervention and non-intervention facilities [22].

In general, monthly trends in both hospitalizations and ED visits changed from increasing to decreasing over time in intervention clinics, and from increasing over time to relatively stable over time in comparison clinics. It is possible that any positive impact SCOUTT may have had on these outcomes was joined by the impact of other factors that were also affecting comparison clinics around the same time as the intervention, such as broader efforts to address OUD and overdose (e.g., efforts to change opioid prescribing, increase MOUD, and increase overdose education/naloxone distribution) [23, 33, 34]. While the present evaluation cannot determine the reasons for these changes, it may be that a combination of multiple VHA initiatives and/or other factors (e.g., an overall increasing awareness of opioid risks) positively affected hospitalization and ED utilization among VHA patients with OUD. It also may be possible that more follow-up time is needed for differences in these outcomes to emerge. Importantly, all data from this evaluation precede the COVID-19 pandemic which may have altered later trends in these outcomes.

Post-hoc secondary analyses suggested that hospitalization trends may have improved in intervention relative to comparison clinics among patients with OUD and ≥ 1 other SUD. This finding was somewhat surprising, given that a prior analysis found that SCOUTT may not have increased MOUD receipt among this group [31]. It is unclear why SCOUTT may have positively impacted hospitalizations among this group, and further evaluation is needed to understand potential mechanisms. For example, if SCOUTT increased clinician knowledge and awareness of SUDs more generally, it may have helped engage patients with multiple SUDs in care and/or prevent hospitalizations related to non-opioid SUDs (e.g., alcohol-related hospitalizations). However, this cannot be determined from the present evaluation.

There may be ways in which SCOUTT and other MOUD implementation efforts can increase their overall impact on downstream outcomes. An increased focus on linking patients at high risk for hospitalization and ED care to MOUD care could increase the impact of the initiative on these outcomes. This might include providing more support for providers to prescribe MOUD to complex patients (e.g., clearer guidelines, increased collaboration with SUD specialists) [35], improved handoff processes from primary care, mental health, and pain settings to SUD specialty care settings, and/or expanding the initiative to inpatient and ED settings. Additionally, a prior evaluation found that although SCOUTT increased MOUD receipt in intervention clinics relative to comparison clinics, retention on MOUD was lower in intervention clinics than comparison clinics [25]. As retention is important for preventing adverse outcomes, a greater focus on supporting retention may be needed in MOUD implementation efforts. Further implementation work, including qualitative investigation, is needed to better understand factors that may be limiting impact on downstream outcomes and how to address them (e.g., barriers to and facilitators of reaching complex patients and supporting retention). Finally, increasing focus on harm reduction strategies in MOUD implementation interventions, especially strategies that may reduce the risk of adverse outcomes, may also increase their impact on hospitalization and ED care outcomes. These include increasing distribution of harm reduction supplies such as naloxone (an opioid overdose reversal medication), sterile syringes, and fentanyl testing supplies, and providing overdose prevention education [36].

There are limitations to this evaluation. Although we adjusted for patient-level covariates and the pre-implementation trend in outcomes and included a control series, there may have been unmeasured confounding (e.g., patient baseline clinical severity or provider factors at intervention vs. comparison clinics) or residual bias from external factors occurring at the same time as the intervention [26]. Intervention clinics may have varied in experience with MOUD prior to the intervention (though matched comparison clinics were selected based on pre-implementation trends in MOUD prescribing), and may have implemented SCOUTT later than the launch date due to variation in the length of time needed to complete local training and other implementation activities. Additionally, future evaluations might examine whether findings differed for different types of clinics in stratified analyses. Though comparison clinics were matched to intervention clinics on multiple characteristics, intervention clinics may have had other differences from comparison clinics that could be related to care outcomes among their patients (e.g., willingness to participate in a quality improvement initiative might be related to overall quality of care). We used a broad definition of OUD to ensure that all patients potentially eligible to receive MOUD were included, however some prior studies have used more conservative definitions [2], and it is possible that our sample included some patients who were misdiagnosed. Our data could not definitively determine whether patients were actively receiving treatment during the entire month, or whether hospitalizations and ED visits were OUD-related -- while OUD may contribute to these outcomes in multiple ways (e.g., overdoses, infections, increased risk of cardiovascular and psychiatric events) [37,38,39,40], some hospitalizations/ED visits may not have been directly related to patients’ OUD. Additionally, these data do not capture ED visits or hospitalizations that occurred outside of the VHA system. Finally, the generalizability of these findings may be limited in non-VHA settings.

Conclusions

This evaluation examined the impact of a national initiative to increase MOUD provision in VHA primary care, mental health and pain clinics on hospitalization and ED utilization. We did not find evidence that the initiative substantially impacted hospitalization or ED utilization relative to comparison clinics, though there may have been positive impacts for patients with longer engagement in SCOUTT clinics and patients with OUD and ≥ 1 other SUD. Overall trends suggested that there was improvement in these outcomes over the period when SCOUTT was implemented in both intervention and comparison clinics, which may have resulted from a combination of multiple VHA initiatives and/or other factors. MOUD implementation initiatives might seek to increase their impact on hospitalization and ED utilization by increasing focus on reaching higher risk patients and incorporating harm reduction strategies.

Data availability

Data are not publicly available due to institutional rules regarding data sharing.

Abbreviations

aOR:

Adjusted odds ratio

CI:

Confidence interval

ED:

Emergency department

MOUD:

Medication for opioid use disorder

OUD:

Opioid use disorder

SCOUTT:

Stepped Care for Opioid Use Disorder Train the Trainer

SUD:

Substance use disorder

VHA:

Veterans Health Administration

References

  1. Volkow ND, Jones EB, Einstein EB, Wargo EM. Prevention and treatment of opioid misuse and addiction: A review. JAMA Psychiatry. 2019;76(2):208–16.

    PubMed  Google Scholar 

  2. Wakeman SE, Larochelle MR, Ameli O, Chaisson CE, McPheeters JT, Crown WH, et al. Comparative effectiveness of different treatment pathways for opioid use disorder. JAMA Netw Open. 2020;3(2):e1920622.

    PubMed  PubMed Central  Google Scholar 

  3. Krawczyk N, Mojtabai R, Stuart EA, Fingerhood M, Agus D, Lyons BC, et al. Opioid agonist treatment and fatal overdose risk in a state-wide US population receiving opioid use disorder services. Addiction. 2020;115(9):1683–94.

    PubMed  PubMed Central  Google Scholar 

  4. Substance Abuse and Mental Health Services Administration. Key Substance Use and Mental Health Indicators in the United States: Results from the 2020 National Survey on Drug Use and Health. (HHS Publication No. PEP21-07-01-003, NSDUH Series H-56). Rockville, MD: Center for Behavioral Health Statistics and Quality, Substance Abuse and Mental Health Services Administration. Accessed March 17, 2022 at https://www.samhsa.gov/data/sites/default/files/reports/rpt35325/NSDUHFFRPDFWHTMLFiles2020/2020NSDUHFFR1PDFW102121.pdf. 2021.

  5. Hedegaard H, Miniño A, Spencer M, Warner M. Drug Overdose Deaths in the United States, 1999–2020. NCHS Data Brief No. 428. Centers for Disease Control and Prevention. Accessed January 7, 2022 from: https://www.cdc.gov/nchs/products/databriefs/db428.htm. 2021.

  6. Ahmad F, Cisewski J, Rossen L, Sutton P. Provisional drug overdose death counts. National Center for Health Statistics. Accessed January 21, 2023 at: https://www.cdc.gov/nchs/nvss/vsrr/drug-overdose-data.htm. 2023.

  7. Campbell CI, Saxon AJ, Boudreau DM, Wartko PD, Bobb JF, Lee AK, et al. PRimary care opioid use disorders treatment (PROUD) trial protocol: a pragmatic, cluster-randomized implementation trial in primary care for opioid use disorder treatment. Addict Sci Clin Pract. 2021;16(1):9.

    PubMed  PubMed Central  Google Scholar 

  8. Miele GM, Caton L, Freese TE, McGovern M, Darfler K, Antonini VP, et al. Implementation of the hub and spoke model for opioid use disorders in California: rationale, design and anticipated impact. J Subst Abuse Treat. 2020;108:20–5.

    PubMed  Google Scholar 

  9. Korthuis PT, McCarty D, Weimer M, Bougatsos C, Blazina I, Zakher B, et al. Primary Care-Based models for the treatment of opioid use disorder: A scoping review. Ann Intern Med. 2017;166(4):268–78.

    PubMed  Google Scholar 

  10. Hagedorn H, Kenny M, Gordon AJ, Ackland PE, Noorbaloochi S, Yu W, et al. Advancing Pharmacological treatments for opioid use disorder (ADaPT-OUD): protocol for testing a novel strategy to improve implementation of medication-assisted treatment for veterans with opioid use disorders in low-performing facilities. Addict Sci Clin Pract. 2018;13(1):25.

    PubMed  PubMed Central  Google Scholar 

  11. Timko C, Schultz NR, Britt J, Cucciare MA. Transitioning from detoxification to substance use disorder treatment: facilitators and barriers. J Subst Abuse Treat. 2016;70:64–72.

    PubMed  PubMed Central  Google Scholar 

  12. Pating DR, Miller MM, Goplerud E, Martin J, Ziedonis DM. New systems of care for substance use disorders: treatment, finance, and technology under health care reform. Psychiatric Clin. 2012;35(2):327–56.

    Google Scholar 

  13. Cotton AJ, Lo K, Kurtz FB, Waldbauer L. Extended-release buprenorphine outcomes among treatment resistant veterans. Am J Drug Alcohol Abus. 2022;48(3):334–7.

    CAS  Google Scholar 

  14. Samples H, Williams AR, Crystal S, Olfson M. Impact of Long-Term buprenorphine treatment on adverse health care outcomes in medicaid: the impact of longer treatment on health care outcomes for opioid use disorder within a key population of medicaid enrollees. Health Aff (Millwood). 2020;39(5):747–55.

    PubMed  Google Scholar 

  15. Tkacz J, Volpicelli J, Un H, Ruetsch C. Relationship between buprenorphine adherence and health service utilization and costs among opioid dependent patients. J Subst Abuse Treat. 2014;46(4):456–62.

    PubMed  Google Scholar 

  16. Williams AR, Samples H, Crystal S, Olfson M. Acute care, prescription opioid use, and overdose following discontinuation of long-term buprenorphine treatment for opioid use disorder. Am J Psychiatry. 2020;177(2):117–24.

    PubMed  Google Scholar 

  17. Ronquest NA, Willson TM, Montejano LB, Nadipelli VR, Wollschlaeger BA. Relationship between buprenorphine adherence and relapse, health care utilization and costs in privately and publicly insured patients with opioid use disorder. Subst Abuse Rehabilitation. 2018:59–78.

  18. Schwarz R, Zelenev A, Bruce RD, Altice FL. Retention on buprenorphine treatment reduces emergency department utilization, but not hospitalization, among treatment-seeking patients with opioid dependence. J Subst Abuse Treat. 2012;43(4):451–7.

    PubMed  PubMed Central  Google Scholar 

  19. Lo-Ciganic WH, Gellad WF, Gordon AJ, Cochran G, Zemaitis MA, Cathers T, et al. Association between trajectories of buprenorphine treatment and emergency department and in‐patient utilization. Addiction. 2016;111(5):892–902.

    PubMed  Google Scholar 

  20. Hsu Y-J, Marsteller JA, Kachur SG, Fingerhood MI. Integration of buprenorphine treatment with primary care: comparative effectiveness on retention, utilization, and cost. Popul Health Manage. 2019;22(4):292–9.

    Google Scholar 

  21. Wakeman SE, Rigotti NA, Chang Y, Herman GE, Erwin A, Regan S, et al. Effect of integrating substance use disorder treatment into primary care on inpatient and emergency department utilization. J Gen Intern Med. 2019;34:871–7.

    PubMed  PubMed Central  Google Scholar 

  22. Bobb JF, Idu AE, Qiu H, Yu O, Boudreau DM, Wartko PD, et al. Offering nurse care management for opioid use disorder in primary care: impact on emergency and hospital utilization in a cluster-randomized implementation trial. Drug Alcohol Depend. 2024;261:111350.

    PubMed  Google Scholar 

  23. Wyse JJ, Gordon AJ, Dobscha SK, Morasco BJ, Tiffany E, Drexler K, et al. Medications for opioid use disorder in the department of veterans affairs (VA) health care system: historical perspective, lessons learned, and next steps. Subst Abus. 2018;39(2):139–44.

    PubMed  PubMed Central  Google Scholar 

  24. Gordon AJ, Drexler K, Hawkins EJ, Burden J, Codell NK, Mhatre-Owens A, et al. Stepped care for opioid use disorder train the trainer (SCOUTT) initiative: expanding access to medication treatment for opioid use disorder within veterans health administration facilities. Subst Abus. 2020;41(3):275–82.

    PubMed  PubMed Central  Google Scholar 

  25. Hawkins EJ, Malte CA, Gordon AJ, Williams EC, Hagedorn HJ, Drexler K, et al. Accessibility to medication for opioid use disorder after interventions to improve prescribing among nonaddiction clinics in the US veterans health care system. JAMA Netw Open. 2021;4(12):e2137238.

    PubMed  PubMed Central  Google Scholar 

  26. Lopez Bernal J, Cummins S, Gasparrini A. The use of controls in interrupted time series studies of public health interventions. Int J Epidemiol. 2018;47(6):2082–93.

    PubMed  Google Scholar 

  27. Lapham G, Boudreau DM, Johnson EA, Bobb JF, Matthews AG, McCormack J, et al. Prevalence and treatment of opioid use disorders among primary care patients in six health systems. Drug Alcohol Depend. 2020;207:107732.

    CAS  PubMed  Google Scholar 

  28. Cantone RE, Garvey B, O’Neill A, Fleishman J, Cohen D, Muench J, et al. Predictors of medication-assisted treatment initiation for opioid use disorder in an interdisciplinary primary care model. J Am Board Fam Med. 2019;32(5):724–31.

    PubMed  PubMed Central  Google Scholar 

  29. Charlson M, Szatrowski TP, Peterson J, Gold J. Validation of a combined comorbidity index. J Clin Epidemiol. 1994;47(11):1245–51.

    CAS  PubMed  Google Scholar 

  30. Wagner AK, Soumerai SB, Zhang F, Ross-Degnan D. Segmented regression analysis of interrupted time series studies in medication use research. J Clin Pharm Ther. 2002;27(4):299–309.

    CAS  PubMed  Google Scholar 

  31. Frost MC, Malte CA, Hawkins EJ, Glass JE, Hallgren KA, Williams EC. Impact of an intervention to implement provision of opioid use disorder medication among patients with and without co-occurring substance use disorders. J Subst Use Addict Treat. 2023;155:209175.

    CAS  PubMed  PubMed Central  Google Scholar 

  32. StataCorp. Stata statistical software: release 18, college station. TX: StataCorp LLC; 2023.

  33. Oliva EM, Christopher ML, Wells D, Bounthavong M, Harvey M, Himstreet J, et al. Opioid overdose education and Naloxone distribution: development of the veterans health administration’s National program. J Am Pharm Assoc (2003). 2017;57(2S):S168–79. e4.

  34. Lin LA, Bohnert ASB, Kerns RD, Clay MA, Ganoczy D, Ilgen MA. Impact of the opioid safety initiative on opioid-related prescribing in veterans. Pain. 2017;158(5):833–9.

    PubMed  Google Scholar 

  35. Frost MC, Soyer EM, Achtmeyer CA, Hawkins EJ, Glass JE, Hallgren KA, et al. Treating opioid use disorder in veterans with co-occurring substance use: A qualitative study with buprenorphine providers in primary care, mental health, and pain settings addict. Sci Clin Pract. 2023;18(1):26.

    Google Scholar 

  36. Mueller SR, Walley AY, Calcaterra SL, Glanz JM, Binswanger IA. A review of opioid overdose prevention and Naloxone prescribing: implications for translating community programming into clinical practice. Subst Abus. 2015;36(2):240–53.

    PubMed  PubMed Central  Google Scholar 

  37. Ronan MV, Herzig SJ. Hospitalizations related to opioid abuse/dependence and associated serious infections increased sharply, 2002–12. Health Aff (Millwood). 2016;35(5):832–7.

    PubMed  Google Scholar 

  38. Gan WQ, Buxton JA, Scheuermeyer FX, Palis H, Zhao B, Desai R, et al. Risk of cardiovascular diseases in relation to substance use disorders. Drug Alcohol Depend. 2021;229:109132.

    PubMed  Google Scholar 

  39. Bohnert ASB, Ilgen MA. Understanding links among opioid use, overdose, and suicide. N Engl J Med. 2019;380(1):71–9.

    PubMed  Google Scholar 

  40. Martins SS, Keyes KM, Storr CL, Zhu H, Chilcoat HD. Pathways between nonmedical opioid use/dependence and psychiatric disorders: results from the National epidemiologic survey on alcohol and related conditions. Drug Alcohol Depend. 2009;103(1–2):16–24.

    PubMed  PubMed Central  Google Scholar 

Download references

Acknowledgements

Not applicable.

Funding

This work was supported by the VHA Office of Mental Health and Suicide Prevention, the VHA Center of Excellence in Substance Addiction Treatment and Education, and the VHA Health Services Research and Development Quality Enhancement Research Initiative (Partnered Evaluation Initiatives 18–203 and 19 − 001). This material is based upon work done as part of the VHA Advanced Fellowship Program in HSR supported by the Office of Academic Affiliations, US Department of Veterans Affairs. The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; or decision to submit the manuscript for publication. The opinions expressed in this work are the authors’ and do not necessarily reflect those of the institutions, funders, the Department of Veterans Affairs, or the United States Government.

Author information

Authors and Affiliations

  1. Health Systems Research (HSR) Center of Innovation for Veteran-Centered and Value-Driven Care, Veterans Affairs (VA) Puget Sound Health Care System, 1660 S Columbian Way, Seattle, WA, 98108, USA

    Madeline C. Frost, Carol A. Malte, Emily C. Williams, Aline Lott, Anissa N. Danner & Eric J. Hawkins

  2. Department of Health Systems and Population Health, University of Washington School of Public Health, 3980 15th Ave NE, Seattle, WA, 98195, USA

    Madeline C. Frost & Emily C. Williams

  3. Center of Excellence in Substance Addiction Treatment and Education, VA Puget Sound Health Care System, 1660 S Columbian Way, Seattle, WA, 98108, USA

    Carol A. Malte, Amy J. Kennedy, Andrew J. Saxon, Hildi J. Hagedorn, Aline Lott, Anissa N. Danner & Eric J. Hawkins

  4. Division of General Internal Medicine, University of Washington School of Medicine, 325 Ninth Ave, Seattle, WA, 98104, USA

    Amy J. Kennedy

  5. Department of Psychiatry and Behavioral Sciences, University of Washington School of Medicine, 1959 NE Pacific Street, Seattle, WA, 98195, USA

    Andrew J. Saxon & Eric J. Hawkins

  6. Informatics, Decision-Enhancement, and Analytic Sciences Center, Health Systems Research, VA Salt Lake City Health Care System, 500 Foothill Dr, Salt Lake City, UT, 84148, USA

    Adam J. Gordon

  7. Program for Addiction Research, Clinical Care, Knowledge, and Advocacy, Department of Internal Medicine, University of Utah School of Medicine, 30 N Mario Capecchi Dr., Salt Lake City, UT, 84112, USA

    Adam J. Gordon

  8. Center for Care Delivery & Outcomes Research, Health Services Research & Development, Minneapolis VA Health Care System, 1 Veterans Dr, Minneapolis, MN, 55417, USA

    Hildi J. Hagedorn

  9. Department of Psychiatry, University of Minnesota, 606 24th Ave S, Minneapolis, MN, 55454, USA

    Hildi J. Hagedorn

  10. Center of Excellence in Substance Addiction Treatment and Education, Corporal Michael J. Crescenz Philadelphia VA Medical Center, 3900 Woodland Ave, Philadelphia, PA, 19104, USA

    Ryan S. Trim

Authors
  1. Madeline C. Frost
    View author publications

    Search author on:PubMed Google Scholar

  2. Carol A. Malte
    View author publications

    Search author on:PubMed Google Scholar

  3. Amy J. Kennedy
    View author publications

    Search author on:PubMed Google Scholar

  4. Andrew J. Saxon
    View author publications

    Search author on:PubMed Google Scholar

  5. Adam J. Gordon
    View author publications

    Search author on:PubMed Google Scholar

  6. Hildi J. Hagedorn
    View author publications

    Search author on:PubMed Google Scholar

  7. Emily C. Williams
    View author publications

    Search author on:PubMed Google Scholar

  8. Ryan S. Trim
    View author publications

    Search author on:PubMed Google Scholar

  9. Aline Lott
    View author publications

    Search author on:PubMed Google Scholar

  10. Anissa N. Danner
    View author publications

    Search author on:PubMed Google Scholar

  11. Eric J. Hawkins
    View author publications

    Search author on:PubMed Google Scholar

Contributions

MCF, CAM, and EJH designed the evaluation and planned the analyses. CAM extracted the data and conducted statistical analyses. MCF led preparation of the manuscript. All authors assisted with interpretation of findings, contributed important intellectual content and revisions to the manuscript, and read and approved the final manuscript.

Corresponding author

Correspondence to Madeline C. Frost.

Ethics declarations

Ethics approval and consent to participate

Per VHA policy (VHA Program Guide 1200.21) this evaluation was designated as a quality improvement project and received a non-research determination from the VHA Office of Mental Health and Suicide Prevention, and therefore did not require institutional review board oversight. The evaluation used secondary electronic health record data and therefore did not consent participants.

Consent for publication

Not applicable.

Competing interests

AJS receives royalties from UpToDate, Inc. No other disclosures are reported by the authors.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Supplementary Material 1

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Frost, M.C., Malte, C.A., Kennedy, A.J. et al. Impact of an opioid use disorder medication implementation intervention on hospitalization and emergency department utilization in the Veterans Health Administration. BMC Psychiatry 25, 288 (2025). https://doi.org/10.1186/s12888-025-06722-6

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1186/s12888-025-06722-6

Keywords

BMC Psychiatry

ISSN: 1471-244X

Contact us