Introduction

Opioids are commonly prescribed medications used in treatment modalities for all spectrums of pain. Although adequate for their use in treating pain, there remains a problem with potential opioid misuse. Data suggests that in 2015 alone, approximately 2.4 million people in America suffered from an opioid use disorder (OUD).1 The average prevalence of OUD in Pennsylvania 2017-2019 was 21 000 individuals (1.6%) in comparison to the national average of 1.0%.2 In 2016 alone, 4642 individuals died from an opioid-related overdose in Pennsylvania.3 The consequential rise in opioid misuse and addiction led to a declaration of a national public health emergency in 2017.4

Buprenorphine is a partial opioid agonist used to treat acute and chronic pain but is also indicated for treating opioid dependence. It is a Schedule III narcotic, meaning it has moderate or low physical dependence or high psychological dependence in pain management.5,6 Buprenorphine works by partially agonizing the mu opioid receptor, leading to a less intense activation than is seen with full agonists, such as morphine.5,6 Buprenorphine also acts as a weak kappa receptor antagonist and delta receptor agonist.5,6 It can be given independently or concurrently with naloxone. The latter combination allows for use of a partial opioid, without the achievement of full opioid effects. The use of buprenorphine was associated with a 32% relative rate of reduction in serious opioid-related acute-care use at 3 months and a 26% relative rate of reduction at 12 months compared with no treatment.7

Buprenorphine was developed in the 1960s as an alternative to stronger, full opioid agonist pain medications like morphine.6 In 2002, the Food and Drug Administration approved the drug for use in treating opiate disorders.1 This led to a larger distribution of buprenorphine as a treatment for OUD. In 2003, 11% of opiate treatment programs (OTPs) and about 5% of non-OTPs in the US offered buprenorphine as a treatment method. By 2015, those numbers rose to 58% and 21%, respectively.8 During the COVID-19 pandemic, OUD patients could be initiated with buprenorphine, but not methadone, pharmacotherapy.9 Pennsylvania ranked tenth in the US in 2017 for per capita buprenorphine distribution.10

The purpose of this study was to analyze the trends in buprenorphine distribution, overall and by 3-digit zip codes, in Pennsylvania from 2010 to 2020.

Methods

Data was obtained from the Drug Enforcement Administration’s Automated Reports and Consolidated Orders System (ARCOS) database. The ARCOS database reports on the amount of each drug distributed by different issuance modalities (ie, pharmacies and hospitals) and has been used in prior research.10 The ARCOS database reports their distribution data in 3-digit zip codes. A 3-digit zip code is representative of a regional area, while the more familiar 5-digit zip code represents a group of delivery addresses within said 3-digit region. Each 3-digit zip code in Pennsylvania (eg, 185 = Scranton) is broken down into quarterly and yearly reports, reported in total grams of drug distributed. Procedures were approved by the institutional review boards of Geisinger and the University of New England.

The programs GraphPad Prism and Microsoft Excel were used to graph and analyze data. In order to normalize the distribution and adjust for the various population densities, number of pharmacies, and differing distributions between each 3-digit zip code, a percent change for each 3-digit zip code was calculated. To account for less populated areas having fewer total grams of distribution, the ratios were normalized by comparing the zip codes to themselves. The percent change was calculated by comparing each individual zip code’s total grams of buprenorphine distributed in 2010 to their totals in 2020. A 95% confidence interval was then calculated to determine if there was a statistically significant change in buprenorphine distribution as expressed as mean ± (1.96 x SD).

To display the regional variation, the percentage changes for 3-digit zip codes were displayed on a heat map. Someka, an add-on to Excel, was used to plot data points onto geographical locations. The percent change of each 3-digit zip code was applied to each 5-digit zip code. For example, the percent change of zip code 150 was 291%. Therefore, 291% was input for every 5-digit zip code starting with 150.

Results

The year with the largest total buprenorphine distribution between 2010 and 2020 was 2020, with 369 000.54 grams (369.0 kg) distributed. Comparing this to the year in which buprenorphine distribution was the lowest, 2010, which had a total of 116 301.74 grams (116.3 kg), there was a 217.3% increase (Figure 1).

Figure 1
Figure 1.Buprenorphine distribution in total grams from 2010 to 2020 in Pennsylvania as reported by ARCOS.

The percent change in buprenorphine distribution in Pennsylvania from 2010 to 2020 was 217%.

Total buprenorphine distribution was further broken down by 3-digit zip codes. There were 47 3-digit zip codes. In 2010, the 190 (Philadelphia) zip code had 15.2 kg of buprenorphine distributed. In 2020, the same zip code had 27.1 kg. The percent change in buprenorphine distributed was 79%, which was the lowest percent change of all 3-digit zip codes in the state (Figure 2). In 2010, zip code 155 (Somerset) had 0.43 kg of buprenorphine distributed. In 2020, the same zip code had 4.2 kg. The percent change in buprenorphine distributed was 885%, which was the highest of all 3-digit zip codes (Figure 2).

Figure 2
Figure 2.Percent change in buprenorphine distribution as reported by ARCOS across 3-digit zip codes in Pennsylvania from 2010 to 2020.

The vertical line demarcates the average percent increase of 217% calculated from the total grams of buprenorphine distributed in PA from 2010 to 2020. Zip code values outside of a 95% confidence interval, calculated as mean ± (1.96 x SD), were marked statistically significant with an asterisk. Zip codes ± 1.50 x SD were marked in red.

No zip code showed a statistically significant decrease (ie, outside of a 95% confidence interval) in the percent change in buprenorphine distribution. Zip codes 155 (Somerset), 169 (Wellsboro), and 177 (Williamsport) yielded statistically significant increases of 885%, 739%, and 633%, respectively (Figure 2). Figure 3 visualizes these regionally disparate patterns across the state. Refer to 2010 Census: Pennsylvania Profile for population densities of Pennsylvania by zip code and county.11

Figure 3
Figure 3.Percent change in buprenorphine distribution by zip codes as reported to ARCOS from 2010 to 2020.

Discussion

The key finding of this report was that buprenorphine distribution in Pennsylvania increased from 2010 to 2020 by 217% (Figure 1). From 2009 to 2018, buprenorphine distribution across the United States rose by approximately 130%.8 From 2010 to 2018, buprenorphine distribution in Pennsylvania rose 149% (Figure 1). The rise in buprenorphine distribution in Pennsylvania is comparable to the rest of the United States.8 However, studies show that buprenorphine use has been disproportionally distributed to rural or suburban areas in comparison to urban areas.3,10,12 This study and previous literature show a similar consensus.10

Buprenorphine distribution in more densely populated areas of Pennsylvania were near or below the average of 217%: 150-152 (Pittsburgh), 228%; 190 (Philadelphia), 79%; 170-171 (Harrisburg), 202% (Figures 1 and 2). At the same time, the 3-digit zip codes that had the highest percent increase in distribution belong to the 3 of the least densely populated areas: 155 (Somerset), 885%; 169 (Wellsboro), 739%; 177 (Williamsport), 633% (Figures 1 and 2). It is important to note that the 3-digit zip code 191 (Philadelphia) had an average increase in distribution of 316%, which was above the state average. The 3-digit zip codes 155, 169 and 177 had statistically significant increases relative to the rest of Pennsylvania. These statistics, showing a lack of buprenorphine distribution to densely populated areas of Pennsylvania, are troublesome given that Philadelphia and Allegheny County rank highest in rates of opioid-related deaths among US counties with a population over 1 million people.13

It was not an objective of this study to explore why the difference in percent change in buprenorphine distributions exists for each population density. However, some factors that may account for the data displayed are explored here. Although buprenorphine is effective in treating OUD,14 the lack of homogenous distribution is apparent.3,10,12 Studies show that roughly 56% of US counties that have the greatest need for buprenorphine treatment likely demonstrate inadequate measures to be able to use buprenorphine as an effective treatment.3 In Philadelphia, buprenorphine access disparities for minorities, in particular Hispanic populations and non–US-born, non-citizen populations, may exist because these populations have the highest documented uninsurance rates.15 The lack of resources in higher-populated areas may play a role in the ability to prescribe buprenorphine.

Methadone is another evidence-based treatment for OUD. It has been demonstrated that buprenorphine and methadone are equally effective in treating OUD.16 However, buprenorphine is inferior to methadone in retaining patients in treatment.14 Studies show that in comparison to methadone, buprenorphine was more likely to be prescribed to patients who are white (92% vs 53% of methadone patients), are employed (56% vs 29% of methadone patients), and who have had some college education (56% vs 19% methadone).12 White individuals make up approximately 80% of the rural population and 56% of the urban population.17 An explanation for this phenomenon may be the fact that methadone was approved for use by the FDA in 1972, and urban populations were the first to prescribe methadone.18 The observed geographical disparities could reflect greater availability of methadone from narcotic treatment programs that are typically located in more urban areas,19,20 whereas buprenorphine is available from primary care providers. In Philadelphia, it was demonstrated that areas with the lowest access to primary care had higher concentrations of non-Hispanic Black individuals and lower median household incomes compared to areas of higher access to primary care.21

It has been demonstrated that the prevalence of obesity in rural populations is significantly higher than urban populations.22 Patients with obesity are predisposed to mechanical injuries commonly treated with opiate analgesics, such as osteoarthritis, other joint disorders, and chronic back pain. This has resulted in a significantly higher risk of opioid use in patients with obesity.23 Studies show that an overweight and obese BMI contribute to approximately 16.2% of opioid prescriptions.23 The relationship between rural obesity and opioid use may be another contributing factor to increased use of opioids, leading to an increase in buprenorphine distribution in less densely populated areas.

Limitations

This novel pharmacoepidemiological study has some potential limitations. ARCOS does not distinguish between the buprenorphine that is distributed to treat OUD versus the presumably much smaller subset, including by veterinarians,24 that is subsequently used to treat pain. Similarly, distribution information does not illuminate how much reached the intended patients versus how much of this Schedule III drug was diverted to others.25 Although the ARCOS data was reported at the level of 3-digit zip codes, some patients, particularly those in rural areas, may not reside in the same zip code as the pharmacy where the buprenorphine was distributed. In addition, the consolidation of 5-digit zip codes in order to represent one 3-digit zip code decreased the granularity of regional percentage changes. It is also unclear whether the areas with the largest percent changes in buprenorphine distribution had particularly low distribution rates in 2010, resulting in a relatively large percentage increase, or whether there was a truly a disproportionate increase in distribution in these areas compared to other areas. Future research could be completed at the county or even patient level using electronic medical records to further characterize the regional, rural/urban, racial/ethnic, and socioeconomic disparities in OUD treatment in the US.

Conclusion

This analysis uncovered that from 2010 to 2020 the percent increase in buprenorphine prescriptions in the state of Pennsylvania was 217%. With the increasing awareness of opioid addiction and the over-prescription of opioids in the US, along with additional physician, nurse practitioner, and physician assistant training in buprenorphine treatment delivery, this percent increase was expected. The zip codes of 155 (Somerset), 169 (Wellsboro), and 177 (Williamsport) showed a statistically significant increase in buprenorphine distribution. Interestingly, these zip codes are in some of the least densely populated areas in Pennsylvania. No zip codes displayed a statistically significant decrease in buprenorphine distribution. However, some of the more densely populated areas of Pennsylvania were near or below the average of 217%: 150-152 (Pittsburgh), 228%; 190 (Philadelphia), 79%; 170-171 (Harrisburg), 202%. This pattern warrants further investigation into the gaps of care in buprenorphine distribution in higher population density areas. Overall, this study provides a foundational basis for investigation of additional opioid and opioid treatment patterns in Pennsylvania and other states that continue to be adversely impacted by the iatrogenic US opioid epidemic.


Contributor Roles

AJ Mileto preformed literature search and review, analyzed the data, prepared figures, authored drafts of the paper, and approved the final manuscript.

RJ Rinaldi preformed literature search and review, analyzed the data, prepared figures, authored drafts of the paper, and approved the final manuscript.

SJ Grampp preformed literature search and review, analyzed the data, authored drafts of the paper, and approved the final manuscript.

KL McCall provided feedback on data analysis and interpretation and approved the final manuscript.

BJ Piper authored drafts of the paper, analyzed data, and approved the final manuscript.

Disclosure Statement

BJP was part of an osteoarthritis research team from 2019 to 2021 supported by Pfizer and Eli Lilly. BJP is a member of the Editorial Board of The Guthrie Journal, for which he receives no financial compensation. The other authors do not report any conflicts of interest.

Institutional Review Board Statement

This study was deemed exempt by the IRB of Geisinger 2020-0223 and by the IRB of the University of New England.

Acknowledgments

We would like to thank Poul Chinga, MBS, for his guidance during the data analysis process. We would like to thank Olivia Lattanzi for her guidance and feedback during the editorial process. This study was supported by the Health Resources Services Administration (D34HP31025).