Abstract
The opioid epidemic, driven in part by increased prescribing, is a public health emergency. This study examines dispensed prescription patterns and approvals of new opioid analgesic products to investigate whether the introduction of these new drugs increases prescribing.
Prescribing patterns based on dispensed prescription claims from the U.S. retail setting were assessed with new brand and generic opioid analgesic products approved in the United States from 1997 through 2015.
From 1997 through 2015, the U.S. Food and Drug Administration (Silver Spring, Maryland) approved 263 opioid analgesic products, including 33 brand products. Dispensed prescriptions initially increased 80% from 145 million prescriptions in 1997 to a peak of 260 million prescriptions in 2012 before decreasing by 12% to 228 million prescriptions in 2015. Morphine milligram equivalents dispensed per prescription increased from 486 in 1997 to a peak of 950 in 2010, before decreasing to 905 in 2015. In 2015, generic products accounted for 96% (218/228 million prescriptions) of all opioid analgesic prescriptions dispensed. The remaining prescriptions were dispensed for brand products, of which nearly half were dispensed for one brand product (OxyContin, Purdue, USA).
There has been a dramatic increase in prescriptions dispensed for opioid analgesics since 1997 and an increasing number of opioid analgesic approvals; however, the number of prescriptions dispensed has declined since 2012 despite an increasing number of approvals. Examination of dispensed prescriptions shows a shifting and complex market where multiple factors likely influence prescribing; the approval of new products alone may not be sufficient to be a primary driver of increased prescribing.
Since 1997 there has been an increasing number of opioid analgesics approved by the U.S. Food and Drug Administration, and an increase in prescriptions dispensed for opioid analgesics
There are concerns that new opioid products are driving prescribing, but the relationship between prescribing patterns and product approvals is not well understood
Data on new brand and generic opioid analgesic product approvals, and on retail dispensed prescription claims, were used to evaluate the opioid product space
Opioid prescriptions dispensed and amount per prescription nearly doubled, and total morphine milligram equivalents more than tripled, from 1997 to the peak in 2010, and partially declined thereafter
Generic products accounted for 68% of total opioid prescriptions in 1997 and 96% in 2015
Approval of new branded opioid products alone does not appear to be a primary driver of increased opioid prescribing
DURING the past two decades, there has been a marked increase in outpatient utilization of opioid analgesics in the United States, paralleled by increases in abuse, misuse, and adverse outcomes, including addiction, overdose, and death.1,2 In 2015, a reported 33,091 deaths were attributed to opioid overdose (prescription and illicit), an increase from 8,050 deaths in 1999.3 Addressing the prescription opioid abuse epidemic is a national priority.4,5
While the increased prescribing of opioids is widely considered to have contributed to this public health emergency, the specific factors leading to increased prescribing are the subject of debate.6 For example, with increasing acknowledgment that pain was often undertreated, Congress declared in 2000 that the upcoming decade would be the “Decade of Pain Control and Research.”7 Pain has also been considered the fifth vital sign since The Joint Commission, an organization that accredits and certifies healthcare organizations such as hospitals, included standards for pain assessment and treatment in 2001.8
During the last 20 yr, the U.S. Food and Drug Administration (Silver Spring, Maryland) has approved numerous new opioid analgesic products. However, most of these new products contain active moieties that have been used for decades in older products. In fact, only one new molecular entity opioid analgesic was approved during this period: tapentadol (2008). New product development of opioid analgesics has focused on new formulations of existing products, including, in some cases, the development of abuse-deterrent formulations.
In recent years, lawmakers, advocacy groups, and others have voiced concern that approvals of new opioid products are driving prescribing.9–11 In theory, new products could drive prescribing by expanding the market with new innovator products, thus increasing product choice, or by increasing accessibility to less expensive products (e.g., generic products). Using nationally estimated prescription claims data and U.S. Food and Drug Administration resources, we examine patterns in prescriptions dispensed in the outpatient setting relative to new opioid product approvals to shed light on the relationship between prescribing patterns and product approvals.
Materials and Methods
We obtained data on all new drug applications and abbreviated new drug applications approved for opioid analgesic products between 1997 and 2015.12 New drug application products include those with new active ingredients, new formulations, and new dosage forms.13 Abbreviated new drug application products are generic versions of products approved under new drug applications.14
National estimates of the number of prescriptions dispensed for opioid analgesics were obtained from a proprietary database available to the U.S. Food and Drug Administration under contract, IQVIA (USA), National Prescription Audit, for January 1997 through December 2015; data before 1997 were not available. The National Prescription Audit captures approximately 3.5 billion prescription claims (88% of U.S. retail prescriptions) annually from a sample of 46,000 out of the approximately 59,400 U.S. outpatient retail pharmacies, which are then projected to provide national-level estimates. Changes to the underlying source data and projection methodologies were conducted by IQVIA over time for greater accuracy. Prescriptions covered by commercial third-party payers, Medicaid, or Medicare, and cash payments are included. Prescription data were analyzed by active moiety (appendix 1), by formulation (e.g., immediate-release, extended-release), and by “brand” (brand or branded generic) or “generic.”
Prescriptions in the National Prescription Audit classified as “brand or branded generic” products include all trade name products. “Generic” products in the National Prescription Audit include products with no trade name. Of note, these classifications of “brand” and “generic” prescriptions do not precisely align with the U.S. Food and Drug Administration’s definitions of products based on new drug application or abbreviated new drug application approval. Prescriptions for products approved as new drug applications and abbreviated new drug applications were classified as “brand” prescriptions if they had a trade name,15,16 while prescriptions for products with no trade name, including some approved as new drug applications (appendix 2), were captured as “generic” prescriptions. Under the National Prescription Audit classification system, most abbreviated new drug application products are captured as “generic” prescriptions. Because authorized generics are products marketed under existing new drug applications without separate U.S. Food and Drug Administration approvals, they could not be counted as separate approvals,17 but were captured as “generic” prescriptions.
All formulations of opioid analgesics were included, except for injectable formulations, which are not commonly dispensed in the outpatient setting. Opioid-containing products used as part of medication-assisted treatment for opioid dependence and opioid-containing cough/cold products were not included in our analysis because of their different indications and patterns of use. Thus, methadone dispensed as medication-assisted treatment from methadone clinics was not captured in the database, but all other methadone prescriptions dispensed from retail pharmacies were included in the analysis.
We also quantified the total amount of morphine milligram equivalents dispensed based on standardization of total dispensed opioid prescriptions across the different active moieties, quantity, and strength of doses. To calculate overall morphine milligram equivalents dispensed per year, we calculated the milligrams of opioids dispensed annually by a morphine milligram equivalents conversion factor for each opioid analgesic using conversion factors outlined in a recent publication by the Centers for Disease Control and Prevention (Atlanta, Georgia).18
We first reported new opioid analgesic product approvals in the United States from 1997 through 2015 by year of approval for new drug applications and abbreviated new drug applications. Next, we show the total opioid analgesic market for brand and generic products by both the nationally estimated number of total prescriptions and morphine milligram equivalents dispensed and prescriptions dispensed adjusted for the total U.S. population calculated using U.S. Census data.19 We also calculated the market share of all new opioid analgesic brand products approved between 1997 and 2015. Our analyses also included all prescriptions dispensed for brand OxyContin (Purdue, USA). OxyContin was originally approved in 1995 before our study period; subsequently another new drug application was approved in 2010 as a reformulation of the original OxyContin, at which time the marketing of the original product ceased. We recorded the approval of the reformulated product as a new brand approval, although we did not distinguish between the original formulation and the new formulation in our calculations of total OxyContin prescriptions dispensed.
Additional supplemental analyses were performed to assess generic opioid analgesic approvals and prescription patterns for the years 1997 through 2015. First, we conducted an analysis of new and additional generic products approved after 1997 for the six most frequently dispensed opioid analgesics in 2015. Next, we analyzed opioid analgesics for which the first generic was initially dispensed between 2006 and 2015, as these represent the clearest cases for interpretation in the current opioid market. Descriptive statistics and calculation of morphine milligram equivalents were conducted using Excel (Microsoft, USA); no statistical hypothesis testing was performed.
Results
New Product Approvals
From 1997 through 2015, 263 new opioid analgesic applications were approved, including 222 abbreviated new drug applications and 41 new drug applications (fig. 1). Thirty-three (33) of the 41 new drug applications approvals were for brand products. The annual number of approvals, which generally increased over time, was higher in the second half of the study period (2007 through 2015; median, 20 approvals/yr; range, 10 to 27) than in the first half (1997 through 2006; median, 9 approvals/yr; range, 5 to 15).
National Prescription Trends
The nationally estimated number of prescriptions dispensed for opioid analgesics initially increased 80% from 145 million prescriptions in 1997 to a peak of 260 million prescriptions in 2012 before decreasing by 12% to 228 million prescriptions in 2015 (table 1). Adjusted for the total U.S. Census population, the overall trends were similar, but the peak in population-adjusted utilization during the study period was in years 2010 to 2012 (fig. 2). Total morphine milligram equivalents dispensed increased from 70 billion dispensed in 1997 to a peak of 244 billion in 2010 before decreasing by 16% to 206 billion morphine milligram equivalents dispensed in 2015. Morphine milligram equivalents per prescription dispensed nearly doubled from 486 in 1997 to a peak of 950 in 2010 before decreasing to 905 morphine milligram equivalents per prescription in 2015. Generic products accounted for 96% of opioid analgesic prescriptions dispensed in 2015, compared to 68% of total prescriptions in 1997.
New Brand Product Approvals
Of the 33 brand products approved as new drug application between 1997 and 2015, 23 brand products were dispensed in 2015. The annual combined market share of the 23 brand products accounted for 3% of all opioid analgesic prescriptions dispensed in 2015 (fig. 3 and appendix 3), largely dispensed for a single brand product (OxyContin); the combined market share of the other 22 brand products accounted for only 1% of the total opioid analgesic market in 2015. Tapentadol (Nucynta, 2008, and Nucynta ER, 2011; Ortho-McNeil, USA), the only new molecular entity approved during the study period, accounted for 0.3% of the total opioid analgesic market in 2015.
Figure 4 shows prescriptions dispensed for OxyContin and generic extended-release oxycodone increased from nearly 1 million prescriptions dispensed in 1997 to about 7 to 7.5 million prescriptions dispensed annually between 2001 and 2010 (the year reformulated OxyContin was introduced and original OxyContin marketing ceased), before decreasing by 39% from approximately 7 million prescriptions in 2010 to nearly 4.5 million prescriptions 2015. In terms of morphine milligram equivalents dispensed, approximately 33 billion morphine milligram equivalents were dispensed for extended-release oxycodone at its peak in 2010, accounting for 14% of total morphine milligram equivalents dispensed for the total opioid analgesic market.
Generic Product Approvals
The six most frequently dispensed opioid analgesic products in 2015 were oral immediate-release formulations of hydrocodone/acetaminophen (40% of total opioid analgesic prescriptions), single-ingredient tramadol (17%), oxycodone/acetaminophen (15%), single-ingredient oxycodone (8%), codeine/acetaminophen (7%), and extended-release morphine (3%), together accounting for nearly 90% of all opioid analgesic prescriptions dispensed in 2015. The majority of these prescriptions were dispensed as generics throughout the examined time (fig. 5). For the examined drugs, the number of prescriptions dispensed generally increased as the cumulative number of generic product approvals increased with the exceptions of hydrocodone/acetaminophen and codeine/acetaminophen. Prescriptions dispensed for hydrocodone/acetaminophen more than doubled from 53 million prescriptions in 1997 to a peak of 129 million prescriptions in 2011 to 2012 before decreasing 29% to 91 million prescriptions in 2015. Prescriptions generally declined for codeine/acetaminophen throughout the examined time before a 27% increase in prescriptions dispensed from 11 million prescriptions in 2013 to 15 million prescriptions in 2015.
Generic versions of six opioid analgesics, which were dispensed for the first time during years 2006 through 2015, are shown in figure 6. Four of the six opioid analgesic products with initial generic approvals between 2006 and 2015 (extended-release hydromorphone, extended-release tramadol, transmucosal immediate-release fentanyl products [excluding sprays], and oxycodone/ibuprofen) showed declines in total dispensing after generic introduction, while two (oxymorphone and extended-release oxymorphone) showed essentially stable total dispensing. In some cases, the decrease in prescriptions began before the introduction of generics.
Discussion
Our examination of brand and generic opioid analgesics product approvals and outpatient dispensing patterns shows that dispensing of opioid analgesics dramatically increased since 1997, both in the number of prescriptions dispensed and in the quantity of opioid per dispensed prescription, as measured by morphine milligram equivalents per dispensed prescription. This increase was accompanied by an increase in the market share of generic drugs, which accounted for 96% of dispensed prescription for opioid analgesics in 2015. The annual number of approvals, which generally increased over time, was higher in the second half of the study period (2007 through 2015) than in the first half (1997 through 2006).
Our data do not suggest a clear relationship between new product approvals and utilization; rather, several observations suggest a shifting and complex market in which multiple factors are at work. First, while the introduction of new products could increase prescribing if additional new molecular entities were brought to market, this had not occurred to any significant extent. Tapentadol, the only new molecular entity approved during the study period, was minimally prescribed, accounting for only 0.3% of the total opioid market in 2015.
Second, our finding that newly approved brand products, which in theory may provide more choice to practitioners and patients, accounted for only 3% of the dispensed prescription opioid analgesic market in 2015, suggesting that the availability of new medicines did not significantly drive prescribing decisions. The finding that OxyContin accounted for approximately 2% of the opioid analgesic market in 2015, while the other newly approved brand products collectively accounted for 1% of the opioid analgesic market, suggests that factors specific to OxyContin may have accounted for its relatively widespread usage among this group of products.
Third, among the six most commonly prescribed opioid analgesics, the relationship of cumulative generic drug approvals to the number of dispensed prescriptions was variable. For example, while the increase in the number of generic approvals was similar for three products (immediate-release hydrocodone/acetaminophen, immediate-release oxycodone/acetaminophen, and immediate-release oxycodone), the number of dispensed prescriptions differed markedly across the three products. In addition, the number of dispensed prescriptions for immediate-release codeine/acetaminophen decreased over a time period during which multiple generic versions were approved.
Fourth, the number of dispensed prescriptions for opioid analgesics whose initial generic version was introduced after 2006 was either stable or declined. The introduction of lower-priced generics has the potential to increase patient accessibility and utilization, with lower prices making medicines more accessible to patients,20 but the data suggest that the introduction of generics in recent years may not increase total dispensing. For example, we observed that the initial introduction of generic extended-release oxycodone did not appear to increase overall prescribing; rather, the increase in use of extended-release oxycodone occurred before the approval of generics. The pattern of stable or declining prescribing observed after introduction of new generic opioid products supports previous research that the approval of generic products may not drive increased utilization.21
Fifth, despite an increased number of approvals in the most recent quarter of the study period (2011 through 2015) relative to previous years, the number of prescriptions dispensed and morphine milligram equivalents decreased since 2012. This finding, along with the observed increase, a near doubling, in the quantity of opioids dispensed per prescription during the study period, further suggests that factors other than product approvals drove prescribing decisions.
The declines in total opioid analgesic dispensing observed in the later years of the study period may be driven in part by interventions implemented by federal, state, and local governments, regulatory agencies, medical associations, healthcare systems, and prescribers. Such efforts may include changes in prescribing guidelines, requirements for prescriber education, recommended limitations on opioid dosages, increased law enforcement activities, payer-based dispensing restrictions, prescription drug monitoring programs, and risk evaluation mitigation strategies.22–27 While the precise contribution of each of these interventions is difficult to ascertain, some patterns observed in this study may be due to actions such as the rescheduling of hydrocodone-combination products in 2014 from Controlled Substance Act Schedule III to Schedule II.28 Rescheduling to Schedule II precludes phone-in or refill prescriptions, potentially making Schedule III products (e.g., codeine/acetaminophen) and Schedule IV products (e.g., tramadol), which are not subject to the same restrictions, easier to prescribe. One study suggests this may have caused a shift from hydrocodone-combination products to other products, but this effect appears to be smaller than the reduction in hydrocodone-combination product dispensing.29
Our study also found that morphine milligram equivalents per prescription dispensed nearly doubled since 1997 to a peak of 950 morphine milligram equivalents per prescription in 2010, before decreasing to 905 morphine milligram equivalents per prescription in 2015. Changes in the annual total of average morphine milligram equivalent dispensed per prescription may be due to changes in drug product prescribed, dosage, strength, or quantity; further investigation is warranted to fully understand the complex factors contributing to the differences in prescribing pattern changes.
Although this study provides a broad overview of opioid analgesic utilization, we were not able to specifically address the impact of our findings on opioid abuse, misuse, addiction, and deaths. The opioid epidemic is a complex and multifactorial phenomenon associated with both legal and illicit sources (e.g., heroin and illicit fentanyl) of opioids.3 Several factors make attribution of overdoses and deaths particularly difficult. For example, the opioid data in our analyses relate to prescriptions obtained legally through retail pharmacies. However, drugs such as heroin may be laced with other drugs such as fentanyl; death certificate data do not capture whether the source of the drug attributed as the cause of death due to overdose is legal or illicit.
The observational and descriptive design of this study also limits our ability to infer causal relationships between regulatory approvals or other interventions and opioid analgesic dispensing patterns. However, observational trends in the data provide a robust, national-level understanding of prescribing patterns. More granular analyses for specific populations or locations may not be generalizable and may reflect local interventions. Although the definitions of “brand” and “generic” used in the analysis of prescription data do not exactly align with U.S. Food and Drug Administration definitions based on new drug application or abbreviated new drug application approval, the overall impact of this difference in the analyses of the opioid analgesic product market appears to be minimal. In addition, we did not examine the role that promotional efforts may have played in increased prescribing of opioids, particularly earlier in the study period. The marketing and promotion of prescription opioid analgesics and their effects on increased prescribing have been the subject of investigations and legal actions.30–32 Finally, we were unable to control for potential confounders such as drug pricing, and our results cannot be generalized to other drug classes.
There has been a dramatic increase in prescriptions dispensed for opioid analgesics since 1997. However, despite an increased number of opioid analgesic approvals in recent years, prescriptions dispensed in the outpatient setting declined since 2012. The current opioid analgesic market appears to be an ecosystem in which the introduction of new products, brand or generic, is more likely to lead to substitutions between products than increased dispensing. Our examination of dispensed prescription patterns shows a shifting and complex market where multiple factors likely influence prescribing, and the approval of new products alone may not be sufficient to be a primary driver of increased prescribing.
Acknowledgments
The authors would like to thank Rajdeep Gill, Pharm.D.; Corinne Woods, R.Ph., M.P.H.; at the U.S. Food and Drug Administration, Silver Spring, Maryland, for their excellent research assistance with IQVIA data. The authors would also like to thank Judy Staffa, Ph.D., R.Ph.; Sharon Hertz, M.D.; Judith A. Racoosin, M.D., M.P.H.; Douglas C. Throckmorton, M.D, at the U.S. Food and Drug Administration, and Christopher M. Jones, Pharm.D., M.P.H., from the Office of the Assistant Secretary for Planning and Evaluation in the U.S. Department of Health and Human Services, Washington, D.C., for providing helpful comments and feedback on the manuscript.
Research Support
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. This study was carried out as part of the authors’ official duties at the U.S. Food and Drug Administration, Silver Spring, Maryland. However, the views expressed in this manuscript are the authors’ own, and do not reflect the official positions of the U.S. Food and Drug Administration.
Competing Interests
The authors declare no competing interests.