This study investigated trends in oxycodone and hydrocodone weights prescribed in the US states of Delaware, Maryland, and Virginia between 2006-2014, and indicated a 53.59% increase in the weight of oxycodone distributed. Analysis for total weight and daily average dose (adjusted for population) showed a comparable increase in oxycodone usage with a relatively small change (+11.05%) in hydrocodone usage.

Reduction in use following peak years may be reflective of new data that has come to light involving the relationship between opioid dose and risk of overdose [17]. Regulatory intervention likely influenced practitioners to adjust prescribing habits across the US. The peak year for opioid prescriptions occurred 2010-2011 in adjusted MME form. Additional data from the CDC reported that between 2006 and 2015, MME peaked across all opioid classes at 782 mg per capita followed by a gradual decrease which still left per capita prescribing weights three times greater than in 1999 [18].

Differences in distribution patterns related to geography has been the subject of research since the institution of prescription monitoring programs. Overall, studies have shown that counties located in the Appalachia region, Southern, and Western states are more susceptible to high numbers of prescriptions per person. These statistics are correlated with the larger county populations and proportions of residents who identify as non-Hispanic/African American or are poor, uninsured, and living in urban areas [7]. There have also been studies indicating that the primary driver for large quantities of opioid prescriptions is unrelated to patient characteristics such as the number of office visits or insurance coverage [19]. Other correlates including the percentage of residents living in poverty in 2010 were examined and compared to oxycodone weight prescribed between 2010 and 2011. 9.7% of Greene County, VA residents reported living below the poverty line at this time. An appreciable portion of residents in Mecklenburg (20.2%) and Southampton (16.4%) were living below the poverty line. The Virginia state poverty rate was 11.1% of residents. With two out of three of the highest prescribing counties above average poverty rates, this seems to be a more accurate indicator for likelihood of high daily average oxycodone dosage during this spike. The percentage of Delaware residents living in poverty was estimated to be 11.9% in 2010. Kent (12.80%) and Sussex (14.12%) were the two counties with above average poverty rates and the only counties where the weight of oxycodone prescribed had increased. However, New Castle’s reported poverty rate was a 11.2% and demonstrated a comparatively moderate increase of 4.40% in usage. Thus, in Delaware the poverty status of residents does not necessarily predict oxycodone dosage. The same is true for Maryland. Frederick, MD (5.6%) and Harford, MD (6.9%) both had poverty rates lower than 2010 state average (9.9%) despite reporting large increases in oxycodone dosage. Although poverty may be a factor the dosage of a prescription, this is likely confounded by other factors such as availability of primary care physicians, average ages of residents, and employment status may play a more important role in this correlation.

Additionally, we sought to explore changes in oxycodone distribution methods by evaluating ARCOS buyer data. ARCOS maintains yearly records on the weight of controlled substances purchased by pharmacies, hospitals, practitioners, mid-level providers, and teaching institutions across the US. Understanding opioid distribution patterns is key in designing protocol aimed at limiting demand in areas at high risk of overprescribing. Many states have implemented laws in recent years designed to strictly regulate the methods for prescribing opioids and other scheduled substances. In Delaware, the “Uniform Controlled Substances Act” was enacted in year and places clear definitions on administration and dispensing while also laying out titles for prescribers for which the law applies including physicians, dentists, and nurse practitioners. As defined by this law, all prescriptions, must be verified by a licensed provider regardless of circumstance. For example, emergent prescriptions must be verified by a board-certified physician [20].

Similar laws in other Mid-Atlantic states apply to prescribers and dispensers. Virginia does not allow for a narcotic supply exceeding seven days without clearly documented and affirmed medical records. However, this can be extended to 14 days following a surgical procedure. The state law also indicates specific measures to be taken with high doses or recurrent use.

Prescriptions above 120 MME require reasonable justification or a pain management referral. Naloxone is also prescribed to patients with risk factors including a prior overdose, substance misuse disorder, or concomitant benzodiazepine use [21]. In Maryland, doses >50 MME require an increase in follow up frequency and for naloxone to be offered to address overdose concerns. Pain management is to be consulted with doses >90 MME [25]. With emphasis on provider data, similar interest was taken in entities buying controlled substances. It should also be noted that prior ARCOS research has found that prescribing laws had no appreciable impact in Maryland and Delaware [23].

Within the region, pharmacies were the largest buyers of both oxycodone (69.2%) and hydrocodone (75.3%). This represented 70.2% opioid weight bought between 2006 and 2014. Hospitals were the second largest buyer of oxycodone (26.7%) and hydrocodone (22.8%). Year to year trends showed that oxycodone purchases by weight substantially increased between 2006 and 2011 between pharmacies in the state. In the peak year for average oxycodone dosage, 2011, pharmacies increased purchasing by +248.3% from 2006, compared to a +9.64% increase in hospital purchasing. This substantial increase likely associated with increased demand for outpatient opioid treatment. These trends fall in line with national analysis by the FDA which found an average increase of 15 billion MME per year supplied to retail pharmacies [24]. ARCOS data provided limited information on the type of pharmacy in question. This complicates trend analysis when evaluating changes in corporate vs independent pharmacy distribution patters. Delaware saw a +95% increase in the number of independent pharmacies operating in the state between 2010 (19) and 2019 (37) which was the largest increase in the United States during this period. Smaller increases were seen in Maryland (+36.0%) and Virginia (+8.0%) [25].

There was additional interest in purchasing trends of practitioners and mid-level providers. In total, practitioners in Delaware, Maryland, and Virginia, increased oxycodone purchase weight (+3170.4%) and hydrocodone weight (+63.1%) between 2010 and 2011. This describes 4.11% and 1.47% of purchase weights between the two medications, respectively. Consistent data for mid-level providers was only available for hydrocodone and represented a decrease of -98.1% during peak years accounting for <1% of hydrocodone purchased. Mid-Level providers, specifically Nurse Practitioners (NP) and Physician’s Assistants (PA), are increasingly autonomous in key areas of clinical care. In 2006, they accounted for 12.7% of emergency department patients seen, which was an increase from 5.5% in 1997 [26]. In all three states of interest, NPs and PAs are permitted to prescribe and dispense samples of Schedule II-V substances. NPs are allowed to procure in all three states while Virginia is the only one that allows NPs to do so. Other differences in restriction include those placed on ambulance services, chiropractors, and optometrists [27].

National studies have suggested that NPs/PAs prescribe opioids in similar patterns compared to MD’s, however, they were more likely to give high-frequency high-dose regimens [28]. Though procurement by mid-level providers was evaluated in this study, the possibility that mid-level providers are writing for opioid regimens in hospital settings cannot be discounted. As the US ranked third among all countries for highest per capita consumption of prescription opioids [1], further data to guide the policies of opioid stewardship programs are needed.

Thorough investigation is also needed to evaluate the relationship between diagnosis and opioid prescriptions particularly in rural areas that saw the highest population adjusted usage in these states. It was noted that counties with lower population densities saw the most significant increases in dosage by percent in their respective states. This finding may indicate the need for a more specified analysis of the clinical use of opioids in rural health centers to confirm other studies. For example, pervious analysis has suggested that there was an undertreatment of cancer-related pain in rural areas of Virginia in 2015 [29]. ARCOS data does not catalogue ICD-10 codes and therefore conclusions about prescriber behavior could not be drawn on the case of diagnostic trends. However, this information is imperative towards identifying the prescribing habits of providers and improving these trends in the interest of public health and patient safety. Other variables such as access to specialized oncologic care, diagnostic rates of terminal cancer, and local attitudes towards terminal pain treatment should also be considered.

Therefore, future studies, drawn from smaller cohorts, could analyze the relationship between geography, conditions, and use of opioids.

Overall, this analysis demonstrates showed three contiguous states with significant variability in the prescription of oxycodone and hydrocodone between their counties. These findings imply the need for federal, state, and local resources and policies to limit the use of these medications to address the opioid epidemic. This variability suggests that the need for resources from state and local funding for monitoring or treatment be handled by county authorities and health systems as opposed to regional ones. It stands to reason the that the maximum benefit from these interventions may be obtained by these channels