In 2021, over 80,000 people in the U.S. died of an overdose involving an opioid, nearly one fifth of which were prescribed.¹ Although opioid dispensing rates have declined since 2012, differences persist among states and counties.^2,3 Urban–rural differences in opioid dispensing were assessed between 2019 and 2021.

IQVIA Xponent data, representing 93% of U.S. retail prescriptions projected to national estimates, were used to calculate five annual opioid dispensing measures between 2019 and 2021 at the national and county levels: morphine milligram equivalents (MMEs) per person, average daily MME per prescription, overall opioid dispensing rate, dispensing rate of opioid prescriptions filled in high dosage (≥50 MME/day), and average days supply per prescription. To compute MME per person, total MME was divided by population estimates. Dispensing rates were calculated per 100 persons. Annual population estimates were obtained from the U.S. Census Bureau and included individuals of all ages.^4,5 A positive or negative change of any metric by ≥10% between 2019 and 2021 was considered an increase or decrease, respectively.

Metropolitan classification of counties in the U.S. was determined by population size using county location of the provider. Three levels examined were metropolitan (≥50,000 people), micropolitan (10,000–49,999 people), and noncore (<10,000 people).⁶ Differences in dispensing among metropolitan levels were computed using absolute difference between the 90^th and 10^th percentiles and their relative ratio.³ Because records contained no identifying information, this study was determined to be exempt from review by the Centers for Disease Control and Prevention IRB. Analyses were conducted using SAS, Version 9.4 (SAS Institute Inc.).

Nationally, all five measures of opioid dispensing decreased, with high-dosage opioid dispensing decreasing the most (−17.4%) and average days of supply decreasing the least (−0.7%). Changes varied substantially over time according to metropolitan level (Table 1 and Appendix Figure 1, available online). For example, the ratio of counties with decreased versus increased MME per person across this period was only 1.6 among noncore counties compared with 4.3 among metropolitan and 2.6 among micropolitan counties (Table 1). Noncore counties experienced the lowest MME per person and opioid dispensing rates. All 5 measures within each metropolitan level had large variability among counties. Of note, among noncore counties, a nearly 70-fold relative differences over time occurred between the county-based 90^th and 10^th percentiles of high-dosage dispensing rates (Table 1).

Nearly 14% (46 million) of U.S. residents live in rural/ metropolitan counties where they may face greater public health challenges than those living in denser population areas.⁷ Rural residents, particularly in noncore counties, typically have a higher prevalence of chronic pain, lower income, less access to health care, and higher uninsured rates, and interventions may not be practical without equitable outreach and engagement of a broader group of stakeholders.^7,8 Each factor can influence the need for prescription pain medication and its ease of access. Within each metropolitan level, the wide variation in opioid prescription dispensing may be related to a shortage of providers, high patient load, and/or greater need for pain relief.^7,8

The continued overall downward national trend in opioid dispensing could be related to state legislation, pharmacy regulations, and prescription drug monitoring programs that track and improve prescribing practices.^2,3 Even so, overdoses and deaths from illicitly manufactured fentanyl, methamphetamine, and cocaine increased during this period.¹ Multiple simultaneous approaches (e.g., primary prevention, harm reduction, evidence-based treatment, and recovery support) could help reduce the nation’s overdose epidemic while improving care for patients with pain, mental health conditions, or substance use disorder.⁹

This study supports that opioid prescribing practices may differ by community on the basis of metropolitan level. Evidence-based interventions suggested by clinical practice guidelines could modify this, even if they are unevenly effective across communities of different metropolitan levels.¹⁰ Suggested evidence-based interventions could include (1) more pain management education of local clinicians concerning the full range of pain treatments; (2) improvements in electronic health records, particularly by embedded clinical decision-making support; (3) enhanced shared decision making with patients where applicable; and (4) team-based care.