A new survey of American adults suggests that illicit opioid use in the United States is much more common than the government’s numbers indicate. In the survey, conducted via the online platform Respondi in June 2024, 7.5 percent of respondents reported they had used (or might have used) illicitly manufactured fentanyl (IMF) in the previous 12 months, 25 times the rate suggested by the government-sponsored National Survey on Drug Use and Health (NSDUH).
RAND Corporation economist David Powell and University of Southern California economist Mireille Jacobson, who published their results in JAMA Health Forum on Friday, say the reasons for that huge disparity are unclear. “A number of previous studies also have reported higher rates of illicit opioid use, challenging the accuracy of the federal estimate,” a RAND press release notes. A 2014 report by Beau Kilmer and eight other drug policy analysts at RAND, for example, estimated that something like 1.5 million Americans were “chronic heroin users” in 2010, when the NSDUH suggested a total of about 620,000 Americans used heroin.
RAND suggests such disparities “may relate to the way the federal survey asks participants about illicit opioid use.” Powell and Jacobson note that “about half of NSDUH surveys are conducted in-person,” which may inhibit respondents’ candor. The Respondi survey, by contrast, was conducted entirely online, which may have encouraged honesty by enhancing the participants’ sense of privacy and making them less likely to shape their answers based on social expectations.
The phrasing of the questions may also help explain the dramatic divergence in estimates. The NSDUH asks, “Have you ever, even once, used illegally made fentanyl?” If the respondent says yes, he is asked, “How long has it been since you last used illegally made fentanyl?”
In the Respondi survey, by contrast, “participants were asked about use of nonprescription opioids within the past 12 months, with heroin and IMF given as examples.” They “could respond in 1 of 3 ways: (1) yes, I intentionally used illicit opioids; (2) yes, I may have unintentionally used illicit opioids; or (3) no.” Respondents who picked 1 or 2 “were subsequently asked about IMF use within the past 12 months with the following 3 options: (1) yes, I intentionally used illicitly made fentanyl; (2) yes, I may have unintentionally used illicitly made fentanyl; or (3) no.”
As Powell and Jacobson concede, the inclusion of unintentional fentanyl use, which they thought was appropriate given the vagaries of the black market drug supply, may have inflated their numbers because “individuals who had used an illicit substance but were unsure whether it contained fentanyl could have selected this response.” But nearly 5 percent of the participants reported intentional IMF use, which is still more than 16 times the rate reported by the NSDUH.
In addition to arguing that the NSDUH is subject to underreporting, critics of the survey have long noted that it omits groups, such as jail or prison inmates and people without fixed addresses, in which the prevalence of illegal opioid use is apt to be especially high. Powell and Jacobson’s survey did not address that issue. In fact, they note that the participants had to have internet access, which may have affected the sample “in systematic ways.” But that bias, they say, “would likely lead us to underestimate illicit opioid use.”
The NSDUH sample is much larger than the number of people who participated in the Respondi survey: 67,500 vs. 1,515. Still, the Respondi sample was larger than those routinely used in public opinion polling. Powell and Jacobson note that Respondi has a reputation for “high-quality nationally representative panels.”
Overall, 11 percent of respondents reported past-year use of illicit opioids, including fentanyl and heroin. Within that group, about 70 percent said that use was intentional.
The survey also asked about the participants’ first exposure to opioids. Among the people who reported past-year use of illegally manufactured opioids, 39 percent said their first exposure “involved opioids prescribed to them,” while 36 percent said it “involved prescription opioids not prescribed to them.”
Although Powell and Jacobson are interested in the potential connection between opioid prescriptions and subsequent illicit use, they note that “we cannot claim that initial exposure caused subsequent illicit opioid use.” Any such causal inference would be reckless, since a large share of American adults—one-third over just a two-year period, according to a 2018 survey—have received opioid prescriptions. Still, it is notable that most illicit opioid consumers in this survey had not received such prescriptions prior to using illegal drugs.
The survey asked illegal opioid users to assess their risk of an overdose. Twenty-four percent said an overdose was very likely, while 33 percent thought it was unlikely. As one might expect, the breakdown was different among fentanyl users: about 33 percent and 18 percent, respectively.
According to an estimate by the Centers for Disease Control and Prevention, “synthetic opioids,” the category that includes fentanyl, were implicated in about 65,000 U.S. deaths during the year ending in June 2024. “If we conservatively assume no fentanyl use among the 21.5% of the population that is younger than 18 years,” Powell and Jacobson say, “the national illicit fentanyl use rate was 5.9%,” which “implies an annual overdose death rate of 0.32% among the population using illicit fentanyl.”
One implication of the Respondi survey’s results, in other words, is that fentanyl use is less dangerous than the NSDUH’s numbers suggest. According to the latter survey, 0.2 percent of Americans 12 or older were past-year fentanyl users in 2023. That would make the “annual overdose death rate” within that group something like 9 percent rather than 0.32 percent.
“Overall,” Powell and Jacobson write, “17.4% of people reporting fentanyl use thought that it was unlikely that they would overdose from opioid use, implying that most people using IMF recognize the heightened risk of overdose from such consumption. Although speculative, the implied awareness about risk suggests that this population may be receptive to interventions that reduce the likelihood of overdose.” Those interventions, they note, include making naloxone, an opioid antagonist that quickly reverses overdoses, “available over the counter” and “distributing fentanyl test strips” to reduce uncertainty about the composition of black market drugs.
Although “polysubstance deaths” are becoming “increasingly common,” RAND notes, “illegally manufactured fentanyl remains involved in most overdose deaths. Despite the importance of illicit opioids in the current substance-use landscape, relatively little is known about the prevalence of illicit opioid use.”
