New Opioid Analgesic Approvals and Outpatient Utilization of Opioid Analgesics in the United States, 1997 Through 2015

Grace Chai, Pharm.D.; Jing Xu, Ph.D., M.A.; James Osterhout, Ph.D.; Mark A. Liberatore, Pharm.D., R.A.C.; Kathleen L. Miller, Ph.D.; Carolyn Wolff, Ph.D.; Marisa Cruz, M.D.; Peter Lurie, M.D., M.P.H. and Gerald Dal Pan, M.D., M.H.S.


Anesthesiology. 2018;128(5):953-966. 

In This Article


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.