The Cost of Implementing and Sustaining the Massachusetts Model

ABSTRACT

Objectives: To identify and value resources required to implement and sustain the Massachusetts model of office-based addiction treatment (MA Model) in the Primary Care Opioid Use Disorders Treatment trial (NCT03407638) using a nurse care manager (NCM) to support medication for opioid use disorder in primary care settings.

Study Design: A site-specific microcosting analysis was conducted via activity-based costing. Guided by a structured costing instrument, we conducted semistructured interviews with relevant personnel and assigned nationally representative costs.

Methods: Data came from 6 health care systems. Costs were categorized as fixed start-up, time dependent, or variable and estimated as annual per-clinic and per-patient costs for implementation and sustainment phases.

Results: Mean implementation cost (ie, year 1 fixed start-up, time-dependent, and variable) was $238,888 per clinic ($3185 per patient); each subsequent year cost $229,676 ($3062 per patient), assuming 75 patients per month and 29% new patient case mix. Mean onetime fixed start-up costs were $9212 per clinic and included supplies and training. Time-dependent costs were $70,446 per clinic and included rent and meetings. Variable costs were $159,229 per clinic and included NCMs’ and prescribers’ clinical duties. On average, NCMs spent 1967.6 hours on MA Model–related work per year (26.2 hours per patient). In sensitivity analyses, costs varied drastically with patient caseload, provider mix, and new patient case mix.

Conclusions: Fixed start-up and time-dependent costs were minimal. Variable costs were 66.7% of implementation costs and 69.3% of costs annually afterward. The primary cost driver was NCM time conducting MA Model–related work. The additional value of the model will depend on associated downstream outcomes. These results may be helpful to health care systems considering implementing the MA Model.

Am J Manag Care. 2026;32(4):e110-e117. https://doi.org/10.37765/ajmc.2026.89923

Takeaway Points

• This study estimated the costs required to implement and sustain the Massachusetts model of office-based addiction treatment (MA Model) for opioid use disorder (OUD) in primary care settings.
• Variable costs were 66.7% of total year 1 costs and 69.3% of costs annually afterward. Costs were mostly associated with time spent by nurse care managers on MA Model–related work. Given that variable costs were the majority, it would be more linear and predictable to find the marginal cost of expansion.
• The cost of the MA Model was on average $3062 per patient per year, comparable with reimbursement for OUD medications in opioid treatment program settings. Additional value will depend on associated downstream outcomes.

The opioid epidemic continues to be a major issue in the US, with an estimated 5.7 million persons experiencing opioid use disorder (OUD)¹ and 81,806 fatal overdoses involving opioids in 2022.² Consequently, the annual economic cost to the US exceeded $960 billion in 2023 US$, more than $100 billion of which is incurred by the health care sector.³ Medications for OUD (MOUD), including buprenorphine, methadone, and naltrexone, are highly effective and safe⁴; however, less than 20% of individuals in need (1.1 million) received MOUD in 2022, highlighting missed opportunities for treatment.¹ In addition to reduced mortality and increased health-related quality of life, MOUD has been linked to downstream cost offsets associated with use of health care and criminal-legal resources.^5-7

The primary care setting provides an opportunity to engage patients in MOUD treatment because buprenorphine and naltrexone for extended-release (ER) injectable suspension (Vivitrol)—the only FDA-approved ER formulation for OUD—can be provided in these settings.⁸ Common concerns regarding providing MOUD in primary care settings, including lack of resources and scope, can be addressed by strategies such as the Massachusetts model of office-based addiction treatment (hereafter referred to as the MA Model). The MA Model utilizes a full-time nurse care manager (NCM) to support office-based MOUD treatment (specifically buprenorphine and ER naltrexone) in collaboration with a primary care provider (PCP). With a sole focus on patients with OUD, NCMs can triage patients and coordinate care, enabling prompt treatment and more efficient use of clinical resources.^9-11

Despite evidence of effectiveness, uncertainty regarding the resources and costs to implement and sustain an intervention such as the MA Model can serve as a barrier to interested clinics. In this study, we conducted a detailed, site-specific microcosting analysis alongside the Primary Care Opioid Use Disorders Treatment (PROUD) cluster-randomized implementation trial (NCT03407638), as in our other studies.^12-16 Additionally, we created a customizable and publicly available budget impact tool for interested clinics to estimate relevant expenditures according to their existing resources.^13,14,17

METHODS

The PROUD Trial Overview

We will briefly discuss the design of the PROUD trial in the context of the resources required to implement and sustain the MA Model; additional information on the study has been published elsewhere.¹⁸

The PROUD trial is being conducted in 6 diverse health systems across the US. The trial required sites to hire a full-time NCM (ie, 36-40 hours/week) and identify a backup NCM and 3 PCPs willing to prescribe buprenorphine. PCPs included physicians, physician assistants (PAs), and nurse practitioners (NPs). The NCM and backup NCM were provided training and weekly technical assistance (TA) by the Boston-based TA team, which developed the MA Model and provided guidance to sites on implementation.^10,11 Both the NCM and backup NCM were registered nurses, and the TA team consisted of 2 NPs.

The MA Model training consisted of 2 in-person phases: a meeting in Boston and a local on-site training. The first phase consisted of a 2-day training in Boston where NCMs shadowed the TA team. The second phase entailed 1 TA team member visiting each trial site, where they observed trainees and PCPs in action and provided on-site guidance and an all-clinic training presentation. The NCM also provided ad hoc trainings to health care system staff, typically on MA Model orientation and MOUD. Additionally, the NCMs jointly attended weekly virtual meetings with the TA team to review clinical cases, receive support, and share findings. The NCMs could also schedule ad hoc meetings with the TA team and receive support via email.

In the typical MA Model workflow, eligible patients (usually referred or self-referred) first met with the NCM for a Diagnostic and Statistical Manual of Mental Disorders (Fifth Edition) OUD screening. Interested patients were then evaluated by a PCP to confirm OUD diagnosis, assess MOUD appropriateness, and schedule MOUD initiation. Patients received at least weekly follow-up visits with the NCM, who also provided care coordination and navigation, either in person or via phone.

Microcosting Overview

Interview data were obtained from 4 of the 6 primary care clinics in the PROUD trial. Data on the remaining clinics were provided by the study lead, implementation monitoring, and TA teams. We utilized a microcosting approach to determine all resources required to implement and sustain the MA Model from the clinic’s perspective. Relevant resources were then (1) valued according to real-world costs faced by clinics; (2) categorized as fixed start-up, time dependent, or variable (defined in the next paragraph); and (3) used to estimate the costs of both the implementation and sustainment phases (defined later), per clinic and per patient. As per established guidelines, the mean per-patient cost is a critical component of the economic evaluation.^12,19 From a budgetary standpoint, the mean cost allows decision makers to estimate the total cost of incorporating an intervention by multiplying the number of patients anticipated; it also provides an estimate of the amount that would need to be sought from payers to avoid incurring a loss. We conducted the microcosting analysis according to a 3- to 5-year timeline to align with typical real-world budgets²⁰ and utilized nationally representative unit costs to enhance generalizability.²¹ A budget impact tool was developed to supplement the data from the microcosting analysis. The tool provides a user-friendly method for organizing, analyzing, and reviewing the resource requirements of each site according to type (fixed start-up, time dependent, variable) and their needs during each phase.

Measures

Fixed start-up resources were defined as onetime costs required only for intervention start-up. Time-dependent resources were defined as those that were required on a recurring basis but were fixed in the interim, meaning they did not vary with the number of patients seen. Variable resources were defined as those required for each patient treated. Labor was valued using nationally representative mean wages and fringe rates derived from the US Bureau of Labor Statistics according to occupation.^22-25 NCM office supplies costs were derived from the US General Services Administration (GSA),²⁶ and the cost of space was derived from the Code of Federal Regulations²⁷ and a commercial real estate report.²⁸ The mean costs associated with trainings were derived from the GSA.^29,30 All unit costs are presented in Table 1 in 2023 US$.^22,25-30

The implementation phase was defined as the period between the start of the MA Model in each of the intervention clinics, including all planning activities, and the intervention steady state, which was assumed to occur 12 months following treatment initiation of the first MA Model patient.^13,14 Thus, the implementation phase included all fixed start-up, time-dependent, and variable resources utilized over the first 12 months of the program. The sustainment phase was defined as the period following steady state and included mean annual time-dependent and variable resource utilization.³¹

Data Collection and Analysis

Resources were identified using an activity-based costing approach,³² guided by the widely used Drug Abuse Treatment Cost Analysis Program instrument to ensure that all intervention-relevant resources were captured.³³ The data collection process consisted primarily of semistructured interviews with site investigators and staff. Each site’s relevant resources were subsequently organized and valued within a budget impact tool, which was adapted from a prior model developed by our team and tailored according to the PROUD intervention design.¹³ Resources were categorized based on whether they qualified as fixed start-up, time-dependent, or variable as well as whether they were required by the site for the implementation and/or sustainment phase. The tool is preloaded with nationally representative unit cost data for labor, medications, training, and infrastructure and is designed to estimate the annual total and per-participant costs for each study phase, which are displayed on the dashboard (eAppendix Figure [available at ajmc.com]). All monetary values are presented in 2023 US$. All analyses were conducted in Excel (Microsoft Corp). The Advarra Institutional Review Board (IRB) (CR00480315) and Weill Cornell Medicine IRB (20-03021596) approved this study.

Sensitivity Analysis

Based on information obtained from the microcosting analysis, the assumptions used for the primary analysis were as follows: (1) Physicians would be the on-site PCP instead of PAs or NPs; (2) 3 PCPs would be available; (3) an average monthly per-clinic caseload was 75 patients; (4) each month’s patient caseload consisted of 29% new patients; and (5) backup NCMs would share resources with the main NCM. The customizable budget impact tool allows the end user to alter each of these assumptions. One-way sensitivity analyses were conducted by varying the provider mix, number of PCPs, number of patients per month, and new patient case mix.

RESULTS

Resources included labor, supplies, space, and time spent on MA Model–related work. Fixed start-up resources included office supplies and trainings. Time-dependent resources included space and weekly TA meetings. Variable resources included time spent interacting with patients and care coordination.

The implementation and annual sustainment costs associated with the MA Model are presented in Table 2. The implementation cost (ie, total cost associated with the first year of the MA model) was estimated to be $238,888 per clinic, with each subsequent year’s cost being $229,676. The mean per-patient cost for the first year of the MA Model was estimated to be $3185, with each subsequent year having an estimated mean cost of $3062 per patient.

Fixed onetime start-up costs (Figure 1) were estimated to be $9212 per clinic and were primarily training related. Boston-based training costs were estimated to be $5737 and included lodging, meals, round-trip airfare, and trainee and trainer time. Local site visit training costs were estimated to be $2640 and included PCP time along with the same costs as the Boston-based training. Ad hoc training cost was estimated to be $30. Additional fixed start-up costs were $805 for the NCM’s office equipment, including a desk, phone, computer, and miscellaneous supplies.

The annualized mean time in each phase associated with the MA Model are as follows: Time spent on the Boston-based training included 16 hours each for the NCM and backup NCM trainees and 8 hours for the 2 TA team members. Time spent on the local site visit training included 4 hours for the NCM and backup NCM trainees, 6 hours for the TA team member, and 2 hours for the PCPs. Time spent on additional training by the NCM was estimated to be 0.5 hours per year.

Annual time-dependent costs (Figure 2) were estimated to be $70,446 per clinic and were driven primarily by office space rent ($45,384). Annual costs associated with the weekly TA calls were estimated to be $25,062, with $5696 for the trainees’ time and $19,366 for the trainers’ time.

Time spent on the weekly TA calls for the trainees was 4 hours per month for preparation and 4 hours per month for the call itself. Time spent on the weekly TA calls for the TA team was 4 hours per month and 2 hours per month for preparation. The TA team also incurred 4 hours per month of ad hoc TA meetings.

Variable costs (Figure 3) were estimated to be $159,229, which were primarily driven by MA Model delivery, representing 66.7% of all implementation costs and 69.3% of costs in each subsequent year. The primary variable cost driver was time spent by the NCM on MA Model–related work, which was $116,743 per year. The annual cost of time spent by PCPs was estimated to be $42,486.

For the mean total time spent on MA Model–related work, the NCM spent 1967.6 hours per year (26.2 hours per patient), in contrast to the 275.6 hours spent per year by the PCP. On average, the team spent 16.1 hours on each new patient per year and 10.5 hours on each established patient per year.

The time spent on MA Model–related work was as follows: The initial visit with each patient was estimated to take 1 hour for the NCM and 0.75 hours for the PCP. For the NCMs, medication initiation, including observation time (if in clinic), was estimated to take 2 hours per patient. Additional face time with each patient was estimated to take 0.75 hours. On a weekly basis, overall additional phone time with patients and patient navigation were estimated to be 2 hours and 4.5 hours, respectively. The total time associated with each patient’s follow-up visits with the NCM was estimated to be 0.75 hours per month. Each patient’s follow-up visit with the PCP was estimated to take 0.5 hours and occurred every 4 months.

Additional one-way sensitivity analyses demonstrated that our results were robust. Varying PCP type resulted in a substantial change to annual variable costs, ranging from $138,984 (NP) to $159,229 (physician), but a minimal change to cost per patient (implementation range, $2905-$3177; sustainment range, $2792-$3062). Varying target monthly patient caseload also resulted in substantial changes to the cost of the MA Model, with 50 patients per month resulting in $106,153 in annual variable costs and 100 patients per month resulting in $367,963 in annual variable costs. Likewise, varying new patient case mix resulted in substantial changes to variable cost, with 13% new patients per month resulting in $103,043 in annual variable costs and 58% new patients per month resulting in $261,067 in annual variable costs. However, varying the number of providers and whether office resources were shared did not result in changes to variable costs and resulted in minimal changes to per-clinic and per-patient costs.

DISCUSSION

To our knowledge, this is the first study to estimate costs required to implement and sustain the MA Model to treat OUD in primary care settings. We found that fixed onetime start-up costs were minimal and that variable costs (ie, those incurred with each patient treated) accounted for 66.7% of total year 1 costs and 69.3% of total costs in subsequent years. This finding offers distinct advantages from a budgeting standpoint. The larger the proportion of variable costs, the more linear and predictable the marginal cost of expansion, thereby simplifying the budget planning process and making it easier to identify areas where expenses can be optimized to achieve financial objectives. We found that the cost of the MA Model in steady state was on average $3062 per patient per year, comparable with reimbursement for MOUD in opioid treatment program settings, which ranges from $2745 to $13,510 per patient per year.^34-38

We found that primary cost drivers for the MA Model, across both the implementation and sustainment phases, were labor related. We also varied what was assumed to be the drivers of cost for the MA Model. As shown in Figure 3, labor-related costs differed greatly across phases. The NCM spent 20 hours on implementation-phase activities and 2063.6 hours during the sustainment phase. The majority of labor-related costs were due to NCM time spent with patients at 1967.6 hours per year, in contrast to the 96 hours per year the NCM spent on TA. Our findings show that the NCM spent more than 7 times the amount of time as the PCP on MA Model–related work, suggesting that the MA Model may allow PCPs to allocate effort to other clinical responsibilities. Space was the primary driver (64%) of time-dependent costs due to estimated monthly rent for the NCM’s office. The cost of supplies required for implementation was minimal. In our sensitivity analyses, costs varied drastically with patient caseload, provider mix, and new patient case mix, supporting the claim that the cost of the intervention was primarily driven by labor-related costs.

A customizable budget impact tool was created using the information garnered from the microcosting analysis and from previous tools.^13,14 Users can tailor the tool by selecting resources specific to their MOUD delivery model, adjusting unit costs to reflect local conditions and estimating service demand. The tool will be publicly available on the resources page of the Center for Health Economics of Treatment Interventions for Substance Use Disorder, HCV, and HIV website.¹⁷ The mean per-patient sustainment costs from this study will also be incorporated into a comprehensive economic evaluation of the MA Model, in which the program cost will be considered in the context of potential cost offsets associated with effective OUD treatment that are relevant to decision makers, such as insurers and policy makers.

Strengths and Limitations

As mentioned, this is the first study to provide estimates of the implementation and sustainment costs for an evidence-based intervention to increase treatment of OUD in primary care settings. We directly solicited estimates of resources needed to deliver the MA Model from the PROUD lead team, implementation monitoring team, TA team, and most importantly, the NCMs. Given our direct access to these different viewpoints of individuals with the most intimate knowledge of the MA Model, we were able to derive a comprehensive estimate of the average cost to implement and maintain the model. Additionally, our microcosting was conducted via activity-based costing^12,19 and was guided by an established structured instrument³³ designed to estimate costs of substance use treatment programs.

Our analyses did not account for loss to follow-up and reestablishment of care, as these were not a significant cost driver in the microcosting interviews. The focus during our interviews was to estimate what the maximum patient caseload would be, both new and existing patients, before another NCM would be needed. However, potential sites will be able to incorporate patient case-mix considerations into the budget impact tool.

Despite the sites being well distributed by geographic location, the small number of sites and their residing in urban areas limit our ability to generalize. Additionally, the sites differed noticeably in their implementation and maintenance of the MA Model (eg, engagement of backup nurses) and similarly had variation in patient population characteristics (eg, patient census), which may cause concerns regarding generalizability of the model. Given the variability across the health care systems, we averaged the rigorously obtained, site-specific values across all 6 sites to derive nationally generalizable results.

As per the PROUD protocol, we assumed the NCM was fully dedicated to the MA Model, with a 75-patient caseload, which could be unrealistic in real-world settings. However, the budget impact tool we designed allows stakeholders to modify the NCM’s effort dedicated to the MA Model. This customizability can provide alternative implementation and sustainment models based on existing resources and requirements, including funding the NCM role via other potential cost centers.

In terms of establishing the program, our analysis did not include the time and resources associated with hiring a new employee because hiring practices and associated costs vary widely across organizations and settings and tend to be well understood by the organization itself. We also did not account for potential differences in the number of patients treated in the implementation phase, but given that we tracked the per-patient resource/cost requirements, it would only affect the average fixed and time-dependent costs, biasing them downward.

This analysis also does not measure the benefits of the MA Model or OUD treatment, including downstream improvements in health and reductions in health care costs. A forthcoming cost-effectiveness analysis, using the results from this study, will provide further key information to stakeholders regarding the net value of the MA Model from health system and societal perspectives.

CONCLUSIONS

Primary care settings offer several unique advantages to help address the ongoing OUD crisis, such as a relatively accessible and less stigmatizing environment for receiving OUD treatment and providers capable of simultaneously caring for the comorbid conditions that often accompany OUD. The MA Model has demonstrated the ability to increase the capacity of primary care clinics to offer MOUD; however, uncertainty regarding resource/cost requirements can serve as a barrier to MA Model uptake. This article and the accompanying budget impact tool can provide interested clinics with the information they need to overcome this barrier.

Acknowledgments

The PROUD Trial Collaborators are Amy M. Loree, PhD, Center for Health Policy and Health Services Research, Henry Ford Health; Leah K. Hamilton, PhD, Kaiser Permanente Washington Health Research Institute; Julia H. Arnsten, MD, Albert Einstein College of Medicine, Montefiore Medical Center; Mark T. Murphy, MD, MultiCare Health System; José Szapocznik, PhD, Department of Public Health Sciences, University of Miami Miller School of Medicine; Mohammad Zare-Mehrjerdi, MD, Department of Family and Community Medicine, UTHealth Houston McGovern Medical School; Mary Shea, MA, Kaiser Permanente Washington Health Research Institute; and Rebecca Phillips, MA, Kaiser Permanente Washington Health Research Institute.

Author Affiliations: Department of Population Health Sciences, Weill Cornell Medicine (PJJ, AJ, SMM), New York, NY; Comparative Health Outcomes, Policy, and Economics Institute (KY), University of Washington (AKL), Seattle, WA; Kaiser Permanente Washington Health Research Institute (CL, KAB), Seattle, WA; RTI International (TL), Durham, NC; Clinical Addiction Research & Education Unit, Boston University Chobanian & Avedisian School of Medicine (CTL), Boston, MA.

Source of Funding: This study was sponsored by the National Institute on Drug Abuse (NIDA) through grants P30DA040500 and UM1DA049412 and through the following nodes and awards of the NIDA Clinical Trials Network (CTN): Health Systems Node (UG1DA040314), Pacific Northwest Node (U10DA013714), Florida Node Alliance (UG1DA013720), New England Consortium (UG1DA015831), Big South-West Node (UG1DA020024), and New York Node (UG1DA013035). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Author Disclosures: Dr Bradley receives royalties from UpToDate. The remaining authors report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.

Authorship Information: Concept and design (PJJ, AJ, KY, TL, KAB, SMM); acquisition of data (PJJ, KY, AKL, KAB, SMM); analysis and interpretation of data (PJJ, AJ, KY, TL, SMM); drafting of the manuscript (PJJ, SMM); critical revision of the manuscript for important intellectual content (PJJ, AJ, KY, CL, TL, AKL, CTL, KAB, SMM); statistical analysis (PJJ, SMM); provision of patients or study materials (CL, AKL, CTL, KAB); obtaining funding (KAB, SMM); administrative, technical, or logistic support (PJJ, CL, CTL); and supervision (KY, KAB, SMM).

Address Correspondence to: Philip J. Jeng, MS, Weill Cornell Medicine, 575 Lexington Ave, 6th Floor, New York, NY 10022. Email: phj2003@med.cornell.edu.

REFERENCES

1. Substance Abuse and Mental Health Services Administration. Key Substance Use and Mental Health Indicators in the United States: Results From the 2023 National Survey on Drug Use and Health. Center for Behavioral Health Statistics and Quality, Substance Abuse and Mental Health Services Administration; 2024. Accessed June 13, 2025. https://www.samhsa.gov/data/sites/default/files/reports/rpt47095/National%20Report/National%20Report/2023-nsduh-annual-national.pdf

2. Drug overdose deaths: facts and figures. National Institute on Drug Abuse. Updated August 2024. Accessed June 13, 2025. https://nida.nih.gov/research-topics/trends-statistics/overdose-death-rates

3. Murphy SM. The cost of opioid use disorder and the value of aversion. Drug Alcohol Depend. 2020;217:108382. doi:10.1016/j.drugalcdep.2020.108382

4. Medications for opioid use disorder. National Institute on Drug Abuse. Accessed April 24, 2024. https://nida.nih.gov/research-topics/medications-opioid-use-disorder

5. Chatterjee A, Weitz M, Savinkina A, et al. Estimated costs and outcomes associated with use and nonuse of medications for opioid use disorder during incarceration and at release in Massachusetts. JAMA Netw Open. 2023;6(4):e237036. doi:10.1001/jamanetworkopen.2023.7036

6. Onuoha EN, Leff JA, Schackman BR, McCollister KE, Polsky D, Murphy SM. Economic evaluations of pharmacologic treatment for opioid use disorder: a systematic literature review. Value Health. 2021;24(7):1068-1083. doi:10.1016/j.jval.2020.12.023

7. Fardone E, Montoya ID, Schackman BR, McCollister KE. Economic benefits of substance use disorder treatment: a systematic literature review of economic evaluation studies from 2003 to 2021. J Subst Use Addict Treat. 2023;152:209084. doi:10.1016/j.josat.2023.209084

8. Medications for Opioid Use Disorder: For Healthcare and Addiction Professionals, Policymakers, Patients, and Families. Treatment Improvement Protocol (TIP) Series 63 Publication No. PEP21-02-01-002. Substance Abuse and Mental Health Services Administration; 2021. Accessed March 12, 2026. https://www.ncbi.nlm.nih.gov/books/NBK574910/

9. Wartko PD, Bobb JF, Boudreau DM, et al; PROUD Trial Collaborators. Nurse care management for opioid use disorder treatment: the PROUD cluster randomized clinical trial. JAMA Intern Med. 2023;183(12):1343-1354. doi:10.1001/jamainternmed.2023.5701

10. Alford DP, LaBelle CT, Kretsch N, et al. Collaborative care of opioid-addicted patients in primary care using buprenorphine: five-year experience. Arch Intern Med. 2011;171(5):425-431. doi:10.1001/archinternmed.2010.541

11. LaBelle CT, Han SC, Bergeron A, Samet JH. Office-based opioid treatment with buprenorphine (OBOT-B): statewide implementation of the Massachusetts collaborative care model in community health centers. J Subst Abuse Treat. 2016;60:6-13. doi:10.1016/j.jsat.2015.06.010

12. Neumann PJ, Sanders GD, Russell LB, Siegel JE, Ganiats TG, eds. Cost-Effectiveness in Health and Medicine. 2nd ed. Oxford University Press; 2016.

13. Ryan DA, Montoya ID, Koutoujian PJ, et al. Budget impact tool for the incorporation of medications for opioid use disorder into jail/prison facilities. J Subst Use Addict Treat. 2023;146:208943. doi:10.1016/j.josat.2022.208943

14. Lu TT, Parent SC, Chaytor N, et al. Budget impact tool for implementing contingency management for co-occurring alcohol use disorders and serious mental illness. Psychiatr Serv. 2024;75(4):326-332. doi:10.1176/appi.ps.20220547

15. Starbird LE, Gutkind S, Teixeira P, et al. Economic cost of the HealthCall smartphone intervention to reduce heavy alcohol drinking in adults with HIV. J Stud Alcohol Drugs. 2023;84(6):814-822. doi:10.15288/jsad.22-00377

16. Gutkind S, Starbird LE, Murphy SM, et al. Cost of hepatitis C care facilitation for HIV/hepatitis C co-infected people who use drugs. Drug Alcohol Depend. 2022;232:109265. doi:10.1016/j.drugalcdep.2022.109265

17. The Center for Health Economics of Treatment Interventions for Substance Use Disorder, HCV, and HIV (CHERISH). Accessed October 4, 2024. https://cherishresearch.org/

18. Campbell CI, Saxon AJ, Boudreau DM, 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. doi:10.1186/s13722-021-00218-w

19. Polsky D, ed. Economic Evaluation in Clinical Trials. 2nd ed. Oxford University Press; 2014.

20. Sullivan SD, Mauskopf JA, Augustovski F, et al. Budget impact analysis—principles of good practice: report of the ISPOR 2012 Budget Impact Analysis Good Practice II Task Force. Value Health. 2014;17(1):5-14. doi:10.1016/j.jval.2013.08.2291

21. Polsky D, Glick H. Costing and cost analysis in randomized controlled trials: caveat emptor. Pharmacoeconomics. 2009;27(3):179-188. doi:10.2165/00019053-200927030-00001

22. Compensation percentiles: a tool for assessing employee compensation. US Bureau of Labor Statistics. Updated June 13, 2025. Accessed June 13, 2025. https://www.bls.gov/ecec/factsheets/compensation-percentile-estimates.htm

23. Occupational Outlook Handbook. US Bureau of Labor Statistics. Accessed June 13, 2025. https://www.bls.gov/ooh/home.htm

24. Employment cost index summary. News release. US Bureau of Labor Statistics. January 31, 2024. Accessed June 13, 2025. https://web.archive.org/web/20240418232533/https:/www.bls.gov/news.release/eci.nr0.htm

25. Occupational employment and wage statistics (OEWS) tables. US Bureau of Labor Statistics. Updated April 3, 2024. Accessed June 13, 2025. https://www.bls.gov/oes/current/oes_nat.htm

26. GSA Advantage! US General Services Administration. Accessed June 13, 2025. https://www.gsaadvantage.gov/advantage/ws/main/start_page?store=ADVANTAGE

27. Public contracts and property management, space standards, criteria, and guidelines. 41 CFR §101-17.3 (1997). Accessed June 13, 2025. https://www.govinfo.gov/content/pkg/CFR-1997-title41-vol2/xml/CFR-1997-title41-vol2-part101-id360-subpart101-id430.xml

28. Jozsa E. National office report: U.S. office sales total $14.8B halfway through 2023. CommercialEdge. July 19, 2023. Accessed June 13, 2025. https://web.archive.org/web/20230810192851/https:/www.commercialedge.com/blog/national-office-report/

29. City Pair Program (CPP). US General Services Administration. Accessed June 13, 2025. https://www.gsa.gov/travel/plan-a-trip/transportation-airfare-rates-pov-rates-etc/airfare-rates-city-pair-program

30. Per diem rates. US General Services Administration. Accessed June 13, 2025. https://www.gsa.gov/travel/plan-book/per-diem-rates

31. Steiner GA. Strategic Planning. Simon and Schuster; 1979.

32. Niñerola A, Hernández-Lara AB, Sánchez-Rebull MV. Improving healthcare performance through activity-based costing and time-driven activity-based costing. Int J Health Plann Manage. 2021;36(6):2079-2093. doi:10.1002/hpm.3304

33. French MT. Drug Abuse Treatment Cost Analysis Program (DATCAP): User’s Manual. University of Miami; 2003. Accessed June 13, 2025. https://cherishresearch.org/wp-content/uploads/2024/08/DATCAPUser_Manual.pdf

34. Clemans-Cope L, Lynch V, Payton M, Aarons J. Medicaid professional fees for treatment of opioid use disorder varied widely across states and were substantially below fees paid by Medicare in 2021. Subst Abuse Treat Prev Policy. 2022;17(1):49. doi:10.1186/s13011-022-00478-y

35. How much does opioid treatment cost? National Institute on Drug Abuse. Updated April 13, 2021. Accessed August 19, 2024. https://web.archive.org/web/20240717044736/https://nida.nih.gov/publications/research-reports/medications-to-treat-opioid-addiction/how-much-does-opioid-treatment-cost

36. Barbosa C, Dowd WN, Buell N, Allaire B, Bobashev G. Simulated impact of Medicaid expansion on the economic burden of opioid use disorder in North Carolina. Int J Drug Policy. 2024;128:104449. doi:10.1016/j.drugpo.2024.104449

37. Toseef MU, Durfee J, Podewils LJ, et al. Total cost of care associated with opioid use disorder treatment. Prev Med. 2023;166:107345. doi:10.1016/j.ypmed.2022.107345

38. Final rule payment rates for opioid treatment programs. CMS. Updated January 8, 2024. Accessed August 19, 2024. https://www.cms.gov/medicare/payment/opioid-treatment-programs-otp/billing-payment/otp-payment-rates