Journal Logo

Research Paper

Declines and regional variation in opioid distribution by U.S. hospitals

Eidbo, Sarah A.a; Kropp Lopez, Amalie K.a; Hagedorn, Joseph D.a; Mathew, Varkeya; Kaufman, Daniel E.a; Nichols, Stephanie D.b; McCall, Kenneth L.b; Piper, Brian J.a,c,*

Author Information
doi: 10.1097/j.pain.0000000000002473

1. Introduction

The United States is enduring a persistent opioid crisis with 128 people dying each day from opioid overdoses.26 Forty percent of opioid overdose deaths involved a prescription opioid.31 Despite the United States constituting 4.4% of the global population, it constituted 30.2% of global opioid consumption.19 There are currently 6146 U.S. hospitals with 924,107 staffed beds.4 In an observational study of opioid exposure and adverse effects, opioids were used for over half of hospital admissions of nonsurgical patients in 286 U.S. hospitals.21 Hospital inpatients and emergency department patients are very frequently administered opioids. Approximately 50 million surgical procedures are performed in the U.S. yearly. Postoperative and predischarge opioid requirements highly correlate with postdischarge opioid requirements and can be used to determine the need for acute outpatient opioid prescriptions.5 In addition, 6% of patients continued opioid use 90 days after surgery, well after the normal scope of postprocedure pain treatment.5 These findings align with the 5% incidence of new long-term opioid use after the first opioid exposure identified by the Oregon Prescription Drug Monitoring Program, which included nonoperative opioid prescriptions.15 Within the inpatient hospital setting, opioids are among the drugs most frequently associated with potentially life-threatening events. Up to 14% of postoperative patients who were prescribed opioids experienced an adverse event. Furthermore, opioid-receiving patients were 3 times as likely to experience an adverse drug event as someone who was not prescribed opioids.11 Among 300,000 surgeries, 12% of patients experienced an opioid-related adverse event, leading to greater overall costs, longer hospitalizations, and increased likelihood of readmission.27

Although the U.S. is still witnessing the consequences of the opioid crisis,20 hospitals have noted these trends.18 As prescription opioids are integral for inpatients' acute pain, recent hospital initiatives include stewardship programs for better education of prescribers as well as increased efforts to reduce negative outcomes—both aimed at decreasing the risk of opioid-related adverse events.18 Postsurgical opioids are often not stored appropriately or disposed of safely. Among patients prescribed outpatient opioids for thoracic surgery pain, only 10% of opioids were stored in a locked location and only 7% were disposed of over the course of the 4 weeks after discharge. Over one-fourth (28.6%) of the 7% who did dispose of the opioid did so unsafely by placing them directly in the garbage.2

This observational study aimed to determine trends in hospital distribution of opioids in the United States from 2000 to 2019, using the Automation of Reports and Consolidated Orders System (ARCOS). The Automation of Reports and Consolidated Orders System is a comprehensive drug reporting system for substances controlled by the Drug Enforcement Administration, which reports on an array of opioids and opioid derivatives in addition to many other substances and has been used in previous pharmacoepidemiological reports.6,10,13,28,29 Several Schedule II/III opioids used for pain control (codeine, fentanyl, hydromorphone, hydrocodone, meperidine, morphine, powdered opium, oxycodone, oxymorphone, remifentanil, and tapentadol) and 2 often used for opioid use disorder (OUD), buprenorphine and methadone, were studied.

2. Methods

2.1. Procedures

All data were collected from ARCOS16,17 for this observational study. The database is an ongoing reporting system that monitors Drug Enforcement Administration–controlled substances by weight with the total of multiple formulations from their manufacture through their retail distribution to hospitals, pharmacies, practitioners, and teaching institutions, which are collectively grouped under business activities in several reports. All Drug Enforcement Administration registrants that manufacture or distribute specific controlled substances according to 21 C.F.R. § 1304.33 are required to report to ARCOS17 which includes when private or public (Veterans Affairs or military) hospitals are the recipients of prescription opioids. To the best of our knowledge, this is the first ARCOS study that has focused exclusively on hospitals. Data from reports 5 and 7 were collected from 2000 to 2019. Report 5 details the purchased amount of each drug, for each business activity, in each state. Report 7 gives the distributed amount of drug for each business activity within the entire U.S. This study was deemed exempt by the IRBs of the University of New England and Geisinger.

2.2. Statistics

The programs GraphPad Prism, Microsoft Excel, and JMP were used to graph and analyze these data. The American Community Survey estimates of yearly state populations30 were used with opioids reported per 100 population. To compare opioids of different formulations, established morphine milligram equivalent (MME) oral conversion factors were used to calculate the MME of each opioid listed with their corresponding multipliers. Oral conversion factors used were morphine: 1, oxycodone: 1.5, fentanyl: 75, hydrocodone: 1, hydromorphone: 4, oxymorphone: 3, tapentadol: 0.4, codeine: 0.15, meperidine: 0.1, methadone: 8, and buprenorphine: 10.19 The percent change was calculated using the MME per person per state to compare 2012, the year with the highest U.S. MME total, with 2019. State values that were outside a 95% confidence interval, calculated as mean ± 1.96 × SD, were considered statistically significant.

3. Results

Between 2012, the peak year, and 2019, hospital opioid use has seen a 46.6% reduction (Fig. 1). When hospital opioid totals were stratified by their individual opioids, patterns became more apparent (Fig. 2). A wide range in distribution weights among opioids required that they be separated into 3 groups. The opioids with the highest hospital distribution were oxycodone, morphine, and hydrocodone, all 3 of which displayed declining usage over the past several years. The middle group included codeine, methadone, hydromorphone, meperidine, buprenorphine, and tapentadol. These drugs also displayed a steady decline, except for buprenorphine, which had distribution increases, and tapentadol, which rose after coming onto the market and then began to fall in 2017. The group of opioids with the lowest distribution included fentanyl, powdered opium, oxymorphone, and remifentanil which all remained steady.

Figure 1.:
Total opioid morphine milligram equivalents (MMEs) in kilograms from 2001 to 2019 as reported by the Drug Enforcement Administration's Automated Reports and Consolidated Ordering System for all opioids and with buprenorphine removed.
Figure 2.:
Individual opioid weights (kg) distributed to U.S. hospitals as reported by the Drug Enforcement Administration's Automated Reports and Consolidated Ordering System (ARCOS) from 2001 to 2018. The percent change relative to the peak year is shown in parentheses.

Opioids were then organized by their percent change between 2012, the peak year, and 2019 (Fig. 3). All opioids displayed a percent decrease except for buprenorphine, which had a +122.5% increase. Six opioids had decreases of 50% or over: hydromorphone (−49.9%), oxymorphone (−57.7%), methadone (−58.7%), morphine (−66.9%), codeine (−67.5%), and meperidine (−77.6%).

Figure 3.:
Individual opioid percent change between 2012 and 2019 as reported by the Drug Enforcement Administration's Automated Reports and Consolidated Ordering System. Changes of greater than +50% shown in green, and changes of greater than −50% shown in red.

Hospital opioid distribution was then broken down by state, by the opioid weight distributed per person by each state's population (Figs. 4 and 5). The average opioid weight distributed per person was 17.91 (SD= 6.92) mg. Compared with 2012, several changes are worth noting: First, half of the United States experienced a 50% or greater percent decrease in hospital-distributed opioid weights per person (Fig. 6). Second, the state with the highest hospital-distributed opioid weight per person in 2012 was South Carolina, which has seen a significant −94% decrease in 2019. It is also worth mentioning that in 2019, the states with the highest hospital opioid distribution per person by weight were Alaska, Montana, Colorado, and North Dakota (Fig. 5). These have partially changed from 2012, when the states with the highest distribution were South Carolina, Alaska, North Dakota, and South Dakota (Fig. 5). Alaska and North Dakota have remained among the states with the highest hospital opioid distributions by weight per person (Fig. 5).

Figure 4.:
Hospital-distributed opioid weights per person per state in milligrams in 2019 as reported by the Drug Enforcement Administration's Automated Reports and Consolidated Ordering System. Average opioid weight per person was 17.91 mg, and SD was 6.92 mg. States with opioid weights per person of over 24.84 mg were 1 SD above the average weight and are colored green. States with opioid weights per person of under 10.98 mg were 1 SD below the average weight and are colored red. States with an opioid weight per person that were significantly different than the national average (*P < 0.05).
Figure 5.:
Hospital-distributed opioid weights in milligrams per person per state in 2012 and 2019 as reported by the Drug Enforcement Administration's Automated Reports and Consolidated Ordering System.
Figure 6.:
Percent decrease in hospital-distributed opioid weights per person per state in milligrams from 2012 to 2019 as reported by the Drug Enforcement Administration's Automated Reports and Consolidated Ordering System. Average decrease in opioid weight per person was −51.38%, and SD was 9.34%. States with decreases in opioid distributions per person of more than −69.68% were 1 SD beyond the average decrease (green). States with decreases in opioid distributions per person of under −33.08% were 1 SD below the average decrease (red). States with a decrease in opioid distribution per person that were significantly different than the national average (*P < 0.05).

4. Discussion

This study identified dynamic and pronounced changes in opioid distribution within U.S. hospitals from 2000 to 2019. Many opioid medications have been declining in use since peaking in 2012. Hospital opioid distribution overall has seen a decrease of 46.6% between 2012 and 2019. This comes with the exception of buprenorphine, which has seen a 122.5% increase in use since 2012 because it is being used to treat OUD. Other opioid medications observed in this investigation reached their peak distributions in 2011, followed by declining distributions. The hospital peak of 19.2 MME metric tons is less than 5% of the national total (389.5) from all sources (ie, pharmacies and hospitals) which peaked in 201129 but otherwise demonstrated comparable reductions.3

The disproportionate amounts of certain opioid medications distributed in some states were another interesting finding. In 2012, South Carolina had the largest hospital-distributed opioid weight per person in the country—it was over twice the weight per person of the state with the second highest distributed opioid weight. Further investigation revealed this high weight to be mostly due to codeine and hydrocodone distribution. Other opioid medications seemed to be within a similar range of the distributed weights of other states. Although the reasoning behind the distribution of these 2 opioid drugs in South Carolina is unclear, it had receded by 2019. South Carolina had a remarkable 93.9% decrease in hospital opioid distribution per person, making it 44th in rank for the highest opioid distribution weight. Alaska had the highest hospital distribution per person in 2019, followed by Montana, Colorado, and North Dakota. The rates of opioid-related inpatient hospitalizations in Alaska were 28.5 per 100,000 persons in 2016 and 26.0 per 100,000 persons in 2017.3 The state is beginning to implement changes to reduce the escalating overdose deaths and high hospitalization rates.

The lowest hospital opioid distribution per person in 2019 was seen in New Jersey, followed by Illinois, New York, and Connecticut. In several of these states, guidelines have been implemented that include a multifaceted approach to reducing hospital opioid distribution. These approaches include performing a patient evaluation before prescribing and periodic review of the complete medical record. This comes from state advisory guidelines on the use of controlled substances in treatment of intractable pain.12 There was a 6-fold difference in population-corrected opioid use between the highest and lowest states. The opioid crisis affects people of all metrics across the nation, but these state-dependent differences indicate that further investigation is necessary for better understanding of this complex crisis.

The U.S. is not alone in the observation of this trend over the past 2 decades. In Canada, 1 of 5 adults reports living with chronic pain.7 Codeine, oxycodone, and hydromorphone were also the most commonly prescribed opioid medications for pain in Canada.7 However in a study from 2013 to 2016 in Ontario, Saskatchewan, and British Columbia, some slight decreases in opioid prescriptions were observed. Fewer Canadians were being prescribed opioids (8.0% fewer), and fewer citizens were starting opioid therapy as treatment for pain in all age groups (9.6% fewer). Changes were also made to prescribed opioid dosages and durations. In summary, the frequency in prescribing opioids for under a week increased and frequency of opioids prescribed for longer than a week declined.7

Methods to decrease opioid medications prescribed to patients are multifaceted and must be addressed according to multiple variables. Funding has been allocated to several areas thus far: addiction prevention strategies, programs aimed at identifying individuals at high risk of opioid use and misuse, increasing treatment facility number, and increased efforts of law enforcement agencies.1 However, despite new safety nets put in place, healthcare providers still balance on a fine line between treating the pain of a patient and potentially overprescribing these complicated medications.

Hospitals in the United States are integrating improved guidelines to reduce amounts of potentially unnecessary opioid prescriptions. A recent report from 81 hospitals found that 98% implemented changes to their opioid delivery practices between July 2017 and July 2018. The changes made included improved prescriber education and new technologies to monitor prescribing practices, which limited the dosage and quantities of given opioids.18

Postsurgical opioid use could be a potential gateway to chronic opioid use if opioid medications are used incorrectly. With the low number of patients taking their prescribed doses of opioid medications after surgery,8 this suggests that perhaps their pain could be managed using other modalities. Patients' pain can often be controlled with nonsteroidal anti-inflammatory drugs and acetaminophen, removing the need for opioids and other stronger pain medications.25 Intravenous ketamine9 or regional anesthesia may also serve as an alternative to opioids in some circumstances. However, the decision to prescribe opioids or their alternatives is the responsibility of the healthcare team. The ability to educate patients and create appropriate expectations for home pain management is vital in these situations.24

Certain strengths and limitations of this novel data set should be taken into consideration. ARCOS only reports in the hospital business activity on agents that are sent by distributors to hospitals. Prescriptions that are written within the hospital—such as those after a surgery—but filled in a nonhospital pharmacy are then considered pharmacy-distributed rather than hospital-distributed, adding a key caveat to this data set. ARCOS does not publicly report on prescription opioids that are purchased but subsequently wasted or lost to theft. There were shortages of parental formulations of some opioids (morphine, hydromorphone, and fentanyl).14 This observational study cannot distinguish the extent that reductions in distribution were due to decreases in demand or supply for these controlled substances. Furthermore, this descriptive report did not characterize the extent that an increase in outpatient procedures being performed at ambulatory surgical centers23 contributed to the observed results. However, the comprehensive nature of ARCOS makes it an important source for follow-up reports. Future research should consider characterizing whether a subset of prescribers were responsible for a disproportionate amount of opioids in the inpatient setting as has been identified for outpatients22 and the impact of the COVID-19 pandemic on opioid use and misuse.

In conclusion, this study observed the differences and trends of 13 different opioids within hospitals from 2000 to 2019 at a national and state level. These changes support those observed at a national level for all opioid distribution—not just within hospitals.10,29 Since 2011, there has been a steady decrease in the distribution of most medications used for pain10,29 and increases for medications used to treat OUD.28 Hospitals were prescribing fewer opioids while also taking steps to make prescribing opioids safer for patients. This is without a doubt a positive step toward combating the persistent effects of the U.S. opioid crisis.

Conflict of interest statement

B.J. Piper is part of an osteoarthritis research team supported by Pfizer and Eli Lilly. The remaining authors have no conflicts of interest to declare.

Supplemental video content

A video abstract associated with this article can be found at


[1]. 340B Health. 340B hospitals on the front lines of addressing the US opioid epidemic: Case examples from the field. (2019). Available at: Accessed June 19, 2021.
[2]. Abrams BA, Murray KA, Mahoney K, Raymond KM, Shannon K, McWilliams SK, Nichols S, Mahmoudi E, Mayes LM, Fernandez-Bustamante A, Mitchell JD, Meguid RA, Zanotti G, Bartels K. Postdischarge pain management after thoracic surgery: a patient-centered approach. Ann Thorac Surg 2020;110:1714–21.
[3]. Alaska Department of Health and Social Services. Opioid epidemic in Alaska. Available at: Accessed June 11, 2021.
[4]. American Hospital Association. (2020). Fast facts on U.S. hospitals, 2020. Available at: Accessed June 11, 2021.
[5]. Brummett CM, Waljee JF, Goesling J, Moser S, Lin P, Englesbe MJ, Bohnert ASB, Kheterpal S, Nallamothu BK. New persistent opioid use after minor and major surgical procedures in us adults. JAMA Surg 2017;152:e170504. Epub 2017 Jun 21. Erratum in: JAMA Surg. 2019;154:272.
[6]. Cabrera FF, Gamarra ER, Garcia TE, Littlejohn AD, Chinga PA, Pinentel-Morillo LD, Tirado JR, Chung DY, Pande LJ, McCall KL, Nichols SD, Piper BJ. Opioid distribution trends (2006-2017) in the US Territories. PeerJ 2019;7:e6272.
[7]. Canadian Institute for Health Information. Opioid prescribing in Canada: How are practices changing? Ottawa: CIHI, 2019.
[8]. Chang AK, Pijur BE, Esses D, Barnaby DP, Baer J. Effect of a single dose of oral opioid and nonopioid analgesics on acute extremity pain in the emergency department. JAMA 2017;318:1661–7.
[9]. Cohen SP, Bhatia A, Buvanendran A, Schwenk ES, Wasan AD, Hurley RW, Viscusi ER, Narouze S, Davis FN, Ritchie EC, Lubenow TR, Hooten WM. Consensus guidelines on the use of intravenous ketamine infusions for chronic pain from the American Society of Regional Anesthesia and Pain Medicine, the American Academy of Pain Medicine, and the American Society of Anesthesiologists. Reg Anesth Pain Med 2018;43:521–46.
[10]. Collins LK, Pande LJ, Chung DY, Nichols SD, McCall KL, Piper BJ. Trends in the medical supply of fentanyl and fentanyl analogues: United States, 2006 to 2017. Prev Med 2019;123:95–100.
[11]. Davies EC, Green CF, Taylor S, Williamson PR, Mottram DR, Pirmohamed M. Adverse drug reactions in hospital in-patients: a prospective analysis of 3695 patient-episodes. PLoS One 2009;4:e4439.
[12]. Davis C. Appendix B: state-by-state summary of opioid prescribing regulations and guidelines. Available at: Accessed June 11, 2021.
[13]. Davis CS, Piper BJ, Gertner AK, Rotter JS. Opioid prescribing laws are not associated with short-term declines in prescription opioid distribution. Pain Med 2019;21:532–7.
[14]. Davis MP, McPherson ML, Mahta Z, Behm B, Fernandez C. What parenteral opioids to use in face of shortages of morphine, hydromorphone, and fentanyl. Am J Hosp Palliat Care 2018;35:1118–22.
[15]. Deyo RA, Hallvik SE, Hildebran C, Marino M, Dexter E, Irvine JM, O'Kane N, Van Otterloo J, Wright DA, Leichtling G, Millet LM. Association between initial opioid prescribing patterns and subsequent long-term use among opioid-naïve patients: a statewide retrospective cohort study. J Gen Intern Med 2017;32:21–7.
[16]. Drug Enforcement Administration, Diversion Control Division. ARCOS retail drug summary reports. Available at: Accessed June 11, 2021.
[17]. Drug Enforcement Administration, Diversion Control Division. Background: what is ARCOS and what does it do? Available at: Accessed June 11, 2021.
[18]. Gallegos A. Hospitals changing practices to curb opioid epidemic. 2019. Available at: Accessed June 11, 2021.
[19]. Garrison G. Claire McCaskill cites disproven figure on opioid use. 2017. Available at: Accessed June 11, 2021.
[20]. Goodnough A. Opioid deaths have surged during the pandemic, C.D.C. data shows. New York Times, 2021. Available at: Accessed September 22, 2021.
[21]. Herzig SJ, Rothberg MB, Cheung M, Ngo LH, Marcantonio ER. Opioid utilization and opioid-related adverse events in nonsurgical patients in US hospitals. J Hosp Med 2014;9:73–81.
[22]. Kiang MV, Humphreys K, Cullen MR, Basu S. Opioid prescribing patterns among medical providers in the United States, 2003-17: Retrospective, observational study. BMJ 2020;368:I6968.
[23]. Koenig L, Gu Q. Growth of ambulatory surgical centers, surgery volume, and savings to Medicare. Am J Gastroenterol 2013;108:10–5.
[24]. Mark J, Argentieri DM, Gutierrez CA, Morrell K, Eng K, Hutson AD, Mayor P, Szender JB, Starbuck K, Lynam S, Blum B, Akers S, Lele S, Paragh G, Odunsi K, de Leon-Casasola O, Frederick PJ, Zsiros E. Ultrarestrictive opioid prescription protocol for pain management after gynecologic and abdominal surgery. JAMA Netw Open 2018;1:e185452.
[25]. Moore PA, Hersh EV. Combining ibuprofen and acetaminophen for acute pain management after third-molar extractions: translating clinical research to dental practice. J Am Dent Assoc 2013;144:898–908.
[26]. National Institute on Drug Abuse. Opioid Overdose Crisis. e 10. Available at: Accessed June 11, 2021.
[27]. Oderda GM, Gan TJ, Johnson BH, Robinson SB. Effect of opioid-related adverse events on outcomes in selected surgical patients. J Pain Palliat Care Pharmacother 2013;27:62–70.
[28]. Pashmineh Azar AR, Cruz-Mullane A, Podd JC, Lam WS, Kaleem SH, Lockard LB, Mandel MR, Chung DY, Simoyan OM, Davis CS, Nichols SD, McCall KL, Piper BJ. Rise and regional disparities in buprenorphine utilization in the United States. Pharmacoepidemiol Drug Saf 2020;29:708–15.
[29]. Piper BJ, Shah DT, Simoyan OM, McCall KL, Nichols SD. Trends in medical use of opioids in the U.S., 2006-2016. Am J Prev Med 2018;54:652–60.
[30]. U.S. Census Bureau. (2020). American Community Survey (ACS). Available at: Accessed June 11, 2021.
[31]. U.S. Department of Health and Human Services. (2019). What is the U.S. opioid epidemic? Available at: Accessed June 11, 2021.

Opiate; Fentanyl; Hydrocodone; Oxycodone; Hydromorphone; Buprenorphine; Methadone; Codeine; Oxymorphone

Supplemental Digital Content

© 2021 International Association for the Study of Pain