Analysis of buprenorphine distribution patterns among pharmacies and hospitals in the USA from 2019 to 2023

Abstract

Background

Opioid use disorder (OUD) is a debilitating condition characterised by the overuse of opioid medications and the development of physical and/or psychological dependence. Consequences of this condition include chronic impairment, distress and later life-altering health conditions such as overdose, all of which have been highlighted by the prominence of OUD in the USA in recent years. Buprenorphine is a standard OUD treatment and commonly used for pain management. Understanding changes in distribution patterns across the USA is vital for continuing to improve outcomes for OUD patients.

Methods

This study used the Drug Enforcement Administration’s Automated Reports and Consolidated Ordering System (ARCOS) and the US Census Bureau Population Estimates databases to analyse changes in buprenorphine distribution among pharmacies and hospitals from 2019 to 2023, to determine temporal patterns and to identify state-level disparities using the data. The data were corrected for population to identify patterns of buprenorphine distribution in the USA from 2021 to 2022 and 2022 to 2023 through examining percent changes in milligrams per 100 population at the national and state levels.

Results

The year-to-year percent change of national buprenorphine distribution from pharmacies has remained positive but changed from a 12.2% increase from 2019 to 2020 (figure 4) to a four per cent increase every year from 2020 to 2023. From 2021 to 2022, there was a +4.9% increase in total grams of buprenorphine distributed to pharmacies and a 95% CI [−5.1, 14.9], with the District of Columbia, South Dakota and Nebraska outside of the 95% CI. Distribution to hospitals increased by 10.2% [-32.3, 52.7] during 2021–2022, with Hawaii, New Hampshire and Delaware being outside of 95% CI. From 2022 to 2023, there was an increase of +5.7% and 95% CI [−3.5, 14.9] in pharmacy distribution, with states including Washington, Rhode Island and Kansas remain outside of the 95% CI. Hospital distribution has decreased from twenty per cent between 2019 and 2020 (figure 4) to eighteen per cent between 2022 and 2023.

Conclusion

Following increases in buprenorphine distribution during the COVID pandemic, a consistent increase has continued year-over-year in most states and the country overall by both pharmacies and hospitals. Some states (eg, Rhode Island, Georgia, District of Columbia) have not followed this pattern. Notably, Hawaii went from the most negative percent change in hospital distribution to the most positive change in the timeframe analysed. This may offer opportunities to analyse more specific impacts of the increased buprenorphine distribution on populations and their outcomes associated with OUD.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• National-level data: This study used Automated Reports and Consolidated Ordering System (ARCOS) data, a comprehensive national database maintained by the Drug Enforcement Administration, enabling a broad analysis of buprenorphine distribution patterns across the USA.

• Trend analysis over time: This study analysed data from 2019 to 2023, allowing for the observation of distribution trends before, during and after the COVID-19 pandemic.

• State-level focus: Inclusion of individual state-level analysis highlighted regional disparities in buprenorphine distribution, supporting targeted policy and public health interventions.

• Lack of patient-level data: The ARCOS database does not include patient demographics or clinical outcomes, which limits the ability to assess the direct impact of distribution patterns on treatment access or effectiveness.

• Cannot infer causality: As an observational study using aggregate data, it cannot determine relationships between policy changes, prescribing patterns and buprenorphine distribution.

Introduction

Opioid use disorder (OUD), defined in the Diagnostic and Statistical Manual of Mental Disorders, Fifth edition (DSM-5) as a ‘problematic pattern of opioid use leading to clinically significant impairment or distress’, is a chronic and life-altering health condition that can lead to comorbidities and mortality, often due to overdose.¹ From 2018 to 2023, opioid-related overdose deaths had continued to increase in the USA, although in 2023, deaths decreased to an estimated 81 083 from an estimated 84 181 in 2022.² While this trend appears encouraging for existing treatments and programmes, approximately 5.7 million people (2.0%) were estimated to have OUD in 2023.³ The overall economic cost of OUD was estimated at $1.5 trillion in 2020 and has continued to increase.⁴ The continued prevalence of OUD, the costs associated with it, and most importantly, the impact on and loss of lives make insight into treatments for OUD vital.

Medications such as buprenorphine and methadone are currently the standard for OUD treatment and pain management.⁵ Buprenorphine is categorised as a schedule III drug that affects all four major opioid receptors, functioning as a mu receptor partial agonist, a kappa receptor agonist and an antagonist to both delta receptors and nociceptin receptors.⁶ Its high affinity and low intrinsic activity on mu receptors cause a weaker opioid effect than alternative options such as methadone. Compounded with a long half-life, high potency and lower relative potential for adverse effects, buprenorphine has been identified as an optimal treatment alternative for those presenting with withdrawal.⁷ Buprenorphine has been demonstrated to reduce opioid-related deaths by up to 50% through increased treatment retention and subsequent risk reduction. When prescribed at fixed, higher dosages (greater than 7 mg daily), it has been shown to effectively decrease illicit opioid use.⁸ Deaths involving buprenorphine occur less frequently than with methadone.⁹ Three-quarters of doctors working in emergency departments preferred buprenorphine over methadone as an OUD treatment.¹⁰ Considering its status as a likely safer alternative treatment option and its being more widely available than methadone in the USA, buprenorphine is garnering continued attention in treating OUD.^{5 11}

Buprenorphine’s use as a treatment has increased across the USA in a non-homogeneous manner since its approval in 2002.¹¹ Policy changes at both federal and state levels have contributed to pronounced differences across states in both access and utilisation.¹² A national-level retrospective cohort study from 2012 to 2015 demonstrated that buprenorphine treatment was disproportionately concentrated among White patients and those with private insurance or self-pay coverage, highlighting racial and socioeconomic gaps in access.¹³ These disparities were further influenced by the federal X-waiver requirement, which, until recently, mandated that healthcare providers complete specialised training and obtain Drug Enforcement Administration (DEA) certification to prescribe buprenorphine for OUD. The waiver was introduced under the Drug Addiction Treatment Act of 2000 (DATA 2000), implemented shortly after buprenorphine’s approval and remained in effect until its removal in January 2023 under the Consolidated Appropriations Act of 2023.

Following the emergence of SARS-CoV-2, responses to the ensuing COVID-19 pandemic by governments at the federal and state levels prioritised increasing and maintaining OUD treatment access by introducing temporal flexibilities, such as telemedicine-based inductions and loosened in-person visit requirements, expanding access to buprenorphine. These changes were associated with increased initiation rates, particularly in rural and underserved populations.¹²

Healthcare impacts from both the pandemic and the years following are still unfolding and yet to be completely understood. This study’s objective was to compile and analyse data for buprenorphine usage by state across the USA to quantify and characterise these disparities in access and distribution of buprenorphine for 2019–2023 overall as well as more focused analysis to compare patterns between 2021–2022 and 2022–2023.

Methods

Procedures

The annual distributions of buprenorphine were extracted from Automated Reports and Consolidated Ordering System (ARCOS) Retail Drug Summary Reports generated by the DEA.¹⁴ ARCOS is a national reporting system that monitors the flow of controlled substances from manufacturers and distributors to retail-level registrants, including pharmacies, hospitals, practitioners, teaching institutions and mid-level providers.¹⁴ The ARCOS Report 5 was used for this analysis, as it provides annualised data by business activity and state, including total grams of buprenorphine distributed to pharmacies and hospitals.

It is important to note that ARCOS differentiates between standalone retail pharmacies and hospitals as separate business activity types. However, hospital-based outpatient pharmacies may be categorised under either group, depending on how they are registered with the DEA. This presents a limitation, as it is not always possible to distinguish outpatient hospital-affiliated dispensing from inpatient hospital use solely from ARCOS data. While ARCOS captures distribution to registered facilities, it does not confirm actual dispensing to patients. Therefore, the reported quantities may include buprenorphine intended for both inpatient treatment and outpatient dispensing but do not directly reflect prescription fills or individual patient utilisation.

Annual estimates of the resident population were obtained from the US Census Bureau for 2019–2023 to normalise buprenorphine distribution per 100 persons.¹⁵ The total quantities of buprenorphine (in grams) distributed to both pharmacies and hospitals across all 50 US states and the District of Columbia were extracted from the ARCOS Report 5 statistical summary for retail drug purchases for 2019 to 2023. For this analysis, our focus was to identify patterns of buprenorphine distribution (total grams) to pharmacies and hospitals post-COVID-19 pandemic and compare 2021–2022 and 2022–2023. In addition, distribution from 2019 to 2023 was used to identify an overall change in buprenorphine.

Data analysis

The distribution rates of buprenorphine in milligrams per 100 population were calculated for each state overall from 2019 to 2023 and annually between 2021–2022 and 2022–2023. Percentage changes in buprenorphine distribution were calculated using the formulas:

$${\displaystyle \frac{\left(Gramsofbuprenorphine2022-Gramsofbuprenorphine2021\right)}{Gramsofbuprenorphine2021}x100}$$

$${\displaystyle \frac{\left(Gramsofbuprenorphine2023 - Gramsofbuprenorphine2022\right)}{Gramsofbuprenorphine2022}x 100}$$

for 2021–2022 and 2022–2023, respectively. The absolute changes in buprenorphine distribution were determined by calculating the differences in grams of buprenorphine per 100K population¹⁴ reported from 2021 to 2022 and 2022–2023 and reported in units of mg/100 people. For the calculation of 95% confidence intervals (CIs), SD was determined by the STDEV.P function in Excel. The margin of error (MoE) was calculated using a=0.05 by the equation:

$${\displaystyle MoE=x\pm 1.96\bullet SD}$$

Data were plotted using GraphPad Prism and statewide geographic maps using the external heat map application, Datawrapper.^{16 17}

Z-scores were standardised in relation to each data set’s calculated mean (μ) and standard deviation (σ) with the following formula:

$$Z=\frac{x-\mu }{\sigma }$$

P-values were likewise calculated in Excel, with significance determined via a two-tailed z-test when α<0.05.

Correlation is determined in GraphPad Prism by calculating the r2 value using sum of squares with the following formula, where SS_reg is the sum-of-squares of the regression and SS_tot is the sum-of-squares of the null hypothesis:

$$r^2=1-\frac{S{S}_{reg}}{S{S}_{tot}}$$

Results

ARCOS

At the national level, buprenorphine distribution increased across both pharmacies and hospitals throughout the study period, though the rate of increase varied by year and distribution site. Pharmacy distribution consistently exhibited positive year-to-year percent increases, though the magnitude of change declined over time. From 2019 to 2020, distribution through pharmacies rose by 12.2%, followed by smaller increases of 4.7% (2020–2021) (figure 1), 4.9% (2021–2022), and 5.7% (2022–2023). In contrast, hospitals demonstrated greater variability. A 20.3% increase was observed from 2019 to 2020 (figure 2), followed by more modest changes of 10.2% from 2021 to 2022 and a subsequent 18.7% rise from 2022 to 2023. These patterns are visually represented in figure 2, which displays percent change in buprenorphine distribution per 100 persons across states from 2019 to 2023—panel A for pharmacies and panel B for hospitals. In pharmacy-based distribution (figure 2A), Nebraska (78.5%), Arkansas (75.6%), South Dakota (73.8%) and Kansas (73.4%) ranked highest, with values significantly above the national average (marked with asterisks). The District of Columbia recorded the lowest percent change (−10.5%), alongside minimal changes in Rhode Island, Connecticut and Massachusetts. For hospitals (figure 2B), California was an extreme outlier with a 1043.6% increase, followed by New Hampshire (274.4%), South Dakota (244.9%) and Alaska (167.6%). In contrast, Georgia experienced the greatest decline (−36.7%) (figure 1A,B).

Figure 1. Relative percent change 2020–2021 pharmacies (A) and hospitals (B) in buprenorphine distribution as reported to the Drug Enforcement Administration’s Automated Reports and Consolidated Orders System. States outside a 95% CI are designated with asterisk.

Figure 2. Relative percent change from 2019 to 2023 in pharmacies (A) and hospitals (B) in buprenorphine distribution as reported to the Drug Enforcement Administration’s Automated Reports and Consolidated Orders System. States outside a 95% CI are designated with asterisk.

To better understand the changes behind national shifts, a closer examination of temporal trends by distribution type—pharmacy versus hospital—helps clarify how each setting contributed to the overall increase in buprenorphine distribution. While pharmacies exhibited more consistent growth, hospital distribution fluctuated more dramatically, likely due to differing roles in long-term versus acute care during and after the COVID-19 pandemic.

Pharmacy distribution demonstrated relative stability between 2020 and 2023, with year-over-year percent changes consistently falling within a narrower CI. Between 2021 and 2022, pharmacy distribution increased by 4.9% nationally, with a 95% CI [−5.1, 14.9], with the District of Columbia, South Dakota and Nebraska outside of the 95% CI (figure 3A). From 2022 to 2023, the increase was +5.7% (95% CI: [−3.5, 14.9]), with states including Washington, Rhode Island and Kansas remaining outside of the 95% CI (figure 4A). In both periods, a minority of states fell outside the 95% CI range, indicating regional variation (figures3A 4A).

Figure 3. Relative percent change from 2021 to 2022 in pharmacies (A) and hospitals (B) in buprenorphine distribution as reported to the US Drug Enforcement Administration’s Automated Reports and Consolidated Orders System. States outside a 95% CI are designated with asterisk.

Figure 4. Relative percent change 2022–2023 in pharmacies (A) and hospitals (B) in buprenorphine distribution as reported to the US Drug Enforcement Administration’s Automated Reports and Consolidated Orders System. States outside a 95% CI are designated with asterisk.

Hospital distribution, on the other hand, showed more dramatic shifts. From 2021 to 2022, hospital-based distribution rose by 10.2% (95% CI: [−32.3, 52.7]), and from 2022 to 2023, it rose by 18.7% (95% CI [−40.8, 78.3]), with several states displaying outlier behaviour (figures3B 4B). This greater variability in hospital distribution likely reflects the acute care focus of hospitals and shifts in pandemic-related demand.

To further contextualise these temporal trends, a state-level analysis reveals where deviations from national patterns were most pronounced. Between 2021 and 2022, the largest changes in pharmacy distribution occurred in the District of Columbia (−16.62%) and Nebraska (17.1%), with only five states showing a decrease overall. For hospital distributions, the largest shifts were observed in Hawaii (−55.6%) and Delaware (85.7%), with ten states exhibiting a decrease. States outside the 95% CI included the District of Columbia, South Dakota and Nebraska for pharmacy distribution and Hawaii, New Hampshire and Delaware for hospital distribution (figure 2A,B).

Between 2022 and 2023, the District of Columbia (−7.5%) and Kansas (18.3%) showed the most extreme changes in pharmacy distribution, with five states overall experiencing a decrease. Hospital distribution was most variable in Maine (−16.8%) and Hawaii (188.3%), and eight states experienced a year-over-year decrease. Once again, the District of Columbia remained the state with the smallest year-over-year increase in pharmacy distribution across both time periods (−16.6% from 2021 to 2022 and −4.7% from 2022 to 2023) (figure 3A,B).

Discussion

This report identified a notable increase in buprenorphine distribution during the COVID-19 pandemic, which has consistently risen year-to-year in most states and the country overall by both pharmacies and hospitals. Analysis of ARCOS usage data revealed that in comparison to increased changes observed during the COVID-19 pandemic, the percent change of buprenorphine distribution from hospitals decreased in growth to under ten percent from 2022 to 2023.¹⁴ Additionally, although it remains positive, distribution from pharmacies changed from over a ten per cent increase from 2019 to 2020 (figure 5) to an increase of four per cent every year from 2020 to 2023.¹⁴

Figure 5. Relative percent change 2019-2020 pharmacies (A) and hospitals (B) in buprenorphine distribution as reported to the Drug Enforcement Administration’s Automated Reports and Consolidated Orders System. States outside a 95% CI are designated with asterisk.

Since the onset of the COVID-19 pandemic, buprenorphine distribution to US pharmacies has exhibited a notable upward trend. Between 2019 and 2020, distribution rates surged by over ten per cent, driven by heightened demand for OUD treatment amid pandemic-related stressors and disruptions to traditional care settings.¹³ Regulatory adaptations, including the expansion of telehealth services, further facilitated access to medications like buprenorphine.^{18 19} However, as the pandemic progressed, growth rates stabilised at approximately four per cent annually, indicating a transition from emergency-driven increases to more sustained expansion. This moderation may reflect factors such as a plateau in demand, systemic adaptations within treatment infrastructures and potential supply chain constraints.^{20 21}

A pivotal policy influencing buprenorphine prescribing was the federal ‘X-waiver’, instituted under DATA 2000 and later lifted through the Mainstreaming Addiction Treatment (MAT) Act, aiming to broaden prescriber participation and enhance treatment accessibility. Despite these intentions, early analyses suggest that the waiver’s elimination did not immediately lead to significant increases in buprenorphine dispensing. For instance, a study in Pennsylvania found no substantial change in prescribing patterns post-waiver removal, indicating that additional systemic barriers, such as stigma and provider readiness, continue to impede access.

In contrast, Canada’s approach to buprenorphine prescribing has been more permissive than that of the USA. In 2018, Health Canada rescinded the federal exemption requirement for prescribing methadone, effectively decentralising control to provincial authorities and simplifying access to opioid agonist therapies (OAT). This policy shift led to increased OAT utilisation across multiple provinces, demonstrating the potential impact of regulatory relaxation on treatment uptake. Unlike the U.S., Canada did not implement an X-waiver equivalent for buprenorphine, allowing for a more streamlined integration of OAT into general medical practice.

Meanwhile, in the USA, the distribution to hospitals initially increased during the COVID-19 pandemic, driven by heightened demand for emergency care and expanded efforts to ensure treatment continuity.¹⁹ However, from 2022 to 2023, hospital-based distribution declined, likely due to several converging factors, including a waning pandemic-related surge in demand and medication supply limitations in inpatient settings.²² Furthermore, hospitals typically focus on acute care and stabilisation rather than long-term maintenance therapy, which is more commonly managed in outpatient settings. This shift helps explain why buprenorphine distribution to pharmacies continued to grow more steadily, reflecting the medication’s integration into ongoing outpatient OUD management.^{13 23}

Some states, such as Rhode Island, Georgia and the District of Columbia, have not followed this pattern. Strikingly, in the timeframe analysed, Hawaii began with the most negative percent change in hospital buprenorphine distribution but finished with the most positive. These abnormalities interestingly coincided with a few notable legislative and administrative actions during the COVID-19 pandemic to address OUD. In the District of Columbia, expansion of telemedicine services allowed for remote prescribing and increased distribution of buprenorphine during the COVID-19 lockdown.²⁴ Similar policy changes reduced barriers to accessing treatment (ie, clinic closure and infective precautionary measures), and could explain increased buprenorphine prescriptions during the pandemic.^{18 19} In contrast, Hawaii experienced more restricted access to buprenorphine due to limited expansion of telemedicine services and distribution of the quantity of medication prescribed without sufficient visits, resulting in more gradual changes over time.^{21 22} This could have led to reduced buprenorphine treatment during the pandemic.²⁵

The increase in buprenorphine distribution is a positive indicator of increased demand for the medication. Despite the rise in buprenorphine availability, overdose rates remain a concern.²⁶ States have shown a general decrease in mortality rates from 2022 to 2023, which could indicate widespread progress with regard to reducing the impacts of OUD in the USA. The effectiveness of buprenorphine for OUD treatment and reducing opioid overdose deaths has been well investigated, highlighting the need for strategies that include medication-assisted treatment and public health interventions.²⁷ Future studies may be indicated to further analyse associations between buprenorphine distributions and state-level policy changes, specifically regarding proactive versus restrictive approaches, with the aim to expand access to OUD treatments. Additionally, research should explore the integration of buprenorphine with other interventions such as mental health services, facilitating a more holistic approach to combating opioid overdoses. Identifying and addressing factors impacting buprenorphine availability and effectiveness in diverse populations will be crucial for improving its role in response to the opioid crisis.

The method of buprenorphine distribution reporting in ARCOS introduces several limitations to this study. First, ARCOS reports the weight of buprenorphine distributed rather than the number of prescriptions or patients treated, which can introduce variability if dosages differ significantly across prescriptions. Additionally, the database does not distinguish between buprenorphine mono- and combination formulations (eg, with naloxone), making it difficult to assess which products are more prevalent, preferred or clinically effective for OUD. Another important limitation is that ARCOS does not capture patient-level data such as medical comorbidities or social determinants of health, thereby restricting the ability to assess how these factors influence access to or outcomes of treatment. Moreover, ARCOS includes buprenorphine quantities used not only for OUD treatment but also for chronic pain management, which may inflate distribution figures and obscure true trends in OUD-specific prescribing. This overlap complicates the interpretation of buprenorphine distribution as a direct proxy for OUD treatment uptake. Lastly, the study’s reliance on percent change in distribution can disproportionately amplify trends in smaller states, where even modest changes in distribution volume can appear as significant percentage shifts due to their lower baseline population. Future studies should explore nuances and investigate how social determinants of health correlate with treatment availability and access to medications like buprenorphine for OUD. Note that this study analysed state-level patterns, and we recognise the vast diversity and variation within each state cannot be fully appreciated within this study’s limited scope. Future studies regarding economic or cultural variations within a smaller region may be warranted to better understand socioeconomic contributors to opioid treatment. Further expansion of research may also consider comparing drug rehabilitation admissions and recovery.

Data availability statement

Data are available upon reasonable request. All data relevant to the study are included in the article or uploaded as online supplemental information.