Medical Use and Misuse of Prescription Opioids in the US Adult Population: 2016–2017

Pamela C. Griesler; PhD, Mei-Chen Hu, PhD; Melanie M. Wall, PhD; Denise B. Kandel, PhD


Am J Public Health. 2019;109(9):1258-1265. 

In This Article

Abstract and Introduction


Objectives: To characterize prescription opioid medical users and misusers among US adults.

Methods: We used the 2016–2017 National Surveys on Drug Use and Health to compare medical prescription opioid users with misusers without prescriptions, misusers of own prescriptions, and misusers with both types of misuse. Multinomial logistic regressions identified substance use characteristics and mental and physical health characteristics that distinguished the groups.

Results: Among prescription opioid users, 12% were misusers; 58% of misusers misused their own prescriptions. Misusers had higher rates of substance use than did medical users. Compared with with-prescription-only misusers, without-and-with-prescription misusers and without-prescription-only misusers had higher rates of marijuana use and benzodiazepine misuse; without-and-with-prescription misusers had higher rates of heroin use. Compared with without-prescription-only misusers, without-and-with-prescription and with-prescription-only misusers had higher rates of prescription opioid use disorder. Most misusers, especially with-prescription-only misusers, used prescription opioids to relieve pain. Misusers were more likely to be depressed than medical users.

Conclusions: Prescription opioid misusers who misused both their own prescriptions and prescription opioid drugs not prescribed to them may be most at risk for overdose. Prescription opioid misuse is a polysubstance use problem.


Despite declines in prescribing[1,2] and nonmedical use,[3,4] prescription opioid (PO) deaths increased from 3442 deaths in 1999 to 17 029 in 2017.[5,6] These deaths represent 35.8% of all opioid overdose deaths, including synthetic opioids (illicit fentanyl) and heroin.[5] The longitudinal population data necessary to understand which PO users are most at risk for negative outcomes are unavailable. Official health and death records do not indicate whether opioids were used as prescribed by a physician or misused, but only whether POs or other drugs contributed to overdoses or deaths. We calculated that, in 2014 and 2015, 5 substances other than POs (alcohol, cocaine, heroin, benzodiazepines, and antidepressants) were involved in 60% of natural and semisynthetic opioid deaths;[7] in 2017, the proportion was similar (61.4%). Identifying the characteristics of PO misusers compared with those who use opioids only as prescribed (medical use) is crucial for understanding who is most at risk for adverse outcomes from POs and for targeting prevention and treatment efforts.

Until 2015, national epidemiological data were restricted to PO misuse. No nationally representative data were available on prescribed PO use to permit comparisons of misusers and medical users. The only available information was from individuals in drug treatment and primary care clinics.[8,9] Misusers were more likely than were medical users to use and be dependent on licit and illicit drugs.[9] Medical users reported higher pain levels and more medical visits.[9] Findings from the National Survey on Drug Use and Health (NSDUH), the annual assessment of drug use in the US population, indicate that PO misusers who obtained prescription opioids from a physician report more frequent PO use and poorer health but less illicit drug use and deviance than do those who obtained prescription opioids from a nonmedical source.[10,11]

As of 2015, the NSDUH continued to assess PO misuse but introduced questions that made it possible to infer medical use.[12] One report compared medical users with PO misusers with and without a PO use disorder in 2015,[13] and another compared 5 levels of opioid exposure, from no opioid use to heroin use in 2015 and 2016.[14] Misusers, especially those with a PO use disorder, had higher rates of substance use, mental health problems, criminal records, and lower economic resources than did medical users.[13,14]

Using data from the 2016–2017 NSDUH surveys, we examined PO medical users and misusers, and we refined the definition of misuse to consider PO sources: whether one's own prescription or from a nonmedical source. Misusers were differentiated according to whether they misused their own prescribed opioids exclusively, misused POs without a prescription from a nonmedical source exclusively, or misused both ways. We characterized medical users and the 3 misuser groups by substance use and disorder, mental and physical health, and sociodemographic characteristics, and implemented multivariate analyses to identify unique correlates of each group. We also examined gender differences.[15,16] The results have implications for prevention and treatment, because misusers of their own prescriptions could be identified by physicians, even when patients misuse from nonmedical sources. The findings based on survey data provide insight into which PO users may contribute to overdoses.