Racial/ethnic differences in prescription opioid misuse and heroin use among a national sample, 1999–2018

Abstract

Objective:

To characterize racial/ethnic differences in past-year prescription opioid misuse and heroin use

Methods:

Data on 1,117,086 individuals age 12 and older were from the 1999–2018 National Survey on Drug Use and Health. We compared relative prevalences across 6 racial/ethnic groups for prescription opioid misuse analyses and 4 racial/ethnic groups for heroin analyses. Unadjusted and gender- and age-adjusted prevalences are reported for 5 time periods (1999–2002, 2003–2006, 2007–2010, 2011–2014, 2015–2018). Survey-weighted Poisson regression models with robust variance were used to estimate risk ratios by race/ethnicity and to test for time trends.

Results:

Prescription opioid misuse was significantly higher among non-Hispanic White individuals than among Black, Hispanic, and Asian individuals across all time periods, yet was highest among Native American individuals in every time period. The relative difference between White and both Hispanic and Asian individuals significantly widened over time, whereas the gap between Black and White individuals significantly decreased. Early in the study period, heroin use was highest among Black and Hispanic individuals. Heroin use among White individuals first surpassed all other groups in 2007–2010 and continued to steadily increase, more than doubling from 1999–2002 to 2015–2018.

Conclusions:

While heroin use has risen among all racial/ethnic groups, the demographics of heroin use have changed significantly in the past two decades such that prevalence is now highest among White individuals. Opioid prevention and treatment initiatives should both be informed by the changing demographics of heroin use and seek to reduce opioid-related harms and expand treatment access equitably for all racial/ethnic groups.

1. INTRODUCTION

The U.S. is currently experiencing an opioid epidemic -- in 2018, nearly 10 million people misused prescription opioids, approximately 800,000 used heroin, and 2 million people had an opioid use disorder (OUD) (SAMHSA, 2019). Some evidence suggests that the opioid crisis is manifesting differently across racial/ethnic groups – in particular, the marked rise in opioid misuse and fatalities among non-Hispanic White individuals has received significant media coverage (Frakt and Monkovic, 2019; Netherland and Hansen, 2016). Opioid-related mortality was relatively similar across racial/ethnic groups in 1999 but has risen sharply among non-Hispanic White individuals, such that their opioid overdose fatality rate was double that of both Black and Hispanic individuals by 2014 (Alexander et al., 2018; James and Jordan, 2018). To date, the rate of total opioid overdose deaths, as well as prescription opioid-specific and heroin-specific deaths, remain highest among non-Hispanic White individuals relative to other racial/ethnic groups (Wilson et al., 2020).

The disproportionate risk for opioid misuse among White individuals has been attributed to historically higher rates of opioid prescribing among this subgroup (Hoffman et al., 2016; Ly, 2019). While some have speculated that racial/ethnic minorities have been “spared” the brunt of the crisis due to lower opioid prescribing rates (Frakt and Monkovic, 2019), this more accurately can be traced to a legacy of under-treatment of chronic pain among racial/ethnic minorities (Hoffman et al., 2016; Lee et al., 2019; SAMHSA, 2020). A meta-analysis found that Black individuals were 29% less likely than White individuals to receive opioid analgesics for similar pain conditions, whereas Hispanic individuals were 22% less likely (Meghani et al., 2012). Differential prescribing has been linked to providers’ inaccurate, negative stereotypes of racial/ethnic minorities (e.g., belief that racial/ethnic minorities have a greater likelihood of becoming dependent or selling medications, a lower accuracy of pain reporting, or a higher pain tolerance) (Santoro and Santoro, 2018).

To date, a limited number of studies have investigated racial/ethnic differences in prescription opioid misuse and heroin use in national samples. In particular, several studies have classified respondents as either “non-Hispanic White” or “Other race/ethnicity,” reporting that heroin use and heroin use disorder have increased significantly more among White individuals than non-White individuals during 2001 to 2013 (Jones et al., 2015; Martins et al., 2017). Other studies have compared non-Hispanic White, Black, and Hispanic individuals, such as a study using 2004–2013 NSDUH data found that prescription opioid misuse was consistently higher among White adolescents relative to Black and Hispanic adolescents (Vaughn et al., 2016). Disparities in opioid misuse among Native American and Asian individuals have been particularly overlooked. Prior studies that have examined these groups often have limited statistical power to robustly assess racial/ethnic differences, due to the confluence of small sample size for these racial/ethnic groups and the very low prevalence of opioid misuse behaviors. Rates of fatal opioid overdose among Native American individuals are second only to White individuals as of 2018 (Wilson et al., 2020), yet few studies have examined prevalence of opioid misuse among Native American individuals. National survey and commercial claims data suggests that Asian individuals have significantly lower rates of prescription opioid misuse (Campbell et al., 2018; Wu et al., 2013) compared to White individuals and the lowest rates of fatal opioid overdose (Wilson et al., 2020), consistent with general epidemiological trends of lower substance use rates among Asian individuals (Fong and Tsuang, 2007). However, national data indicates that treatment admissions, particularly for prescription opioid misuse, increased 30% among Asian individuals during 2000–2012 (Sahker et al., 2017).

Opioid-related mortality data trends suggest that mortality differences across racial/ethnic groups may be diminishing, particularly between White and Black individuals (Lippold et al., 2019). However, national trends in prescription opioid misuse and heroin use by racial/ethnic groups using contemporary data remain unexamined and may not directly track opioid mortality trends. Notably, the growing presence of fentanyl in the illicit opioid supply (Pardo et al., 2019) as well as the availability of naloxone, an overdose reversal medication (Guy et al., 2019), may have changed the relationship between rates of opioid misuse and fatal opioid overdose. Furthermore, fentanyl is increasingly present in the illicit stimulant (e.g., cocaine) supply (Park et al., 2021; Shover et al., 2020), meaning that some fentanyl overdose deaths may represent individuals who misused stimulants rather than opioids. As such, direct examination of trends in prescription opioid misuse and heroin use are warranted.

To address this gap, we use national survey data from 1999–2018 to characterize differences in both past-year prescription opioid misuse and heroin use across racial/ethnic groups among individuals ages 12 and older. By pooling across 20 years of NSDUH data, we have sufficient sample size to examine 6 racial/ethnic groups (White; Black; Hispanic; Asian; Native American; other race/multiracial) in prescription opioid misuse analyses and 4 racial/ethnic groups (White; Black; Hispanic; other race/multiracial) in heroin analyses. Our findings provide insights regarding the differential burden of the opioid crisis across racial/ethnic groups in the U.S.

2. METHODS

2.1. Study Population

Data were from the 1999–2018 NSDUH series, formerly titled the National Household Survey on Drug Abuse (NHSDA). The NSDUH is an annual, nationally-representative household survey of drug use among the civilian, non-institutionalized U.S. population ages 12 and older. Computer-assisted interviewing has been used since 1999 to confidentially administer substance use items and other sensitive questions. Since 2002, respondents have been given an incentive payment of $30. Response rates were relatively similar across years (e.g., 1999=69%, 2003=77%, 2007=74%, 2011=74%, 2015=71%). Our study sample comprised the full survey sample from the 1999 to 2018 NHSDA/NSDUH surveys, yielding a total sample of 1,117,086 individuals; the annual number of survey respondents was similar across years (approximately 55,000).

We note that the NSDUH underwent several revisions during the study period. In 2014, the NSDUH revised the sampling frame that had been used from 1999–2013. Key changes included revising state sample sizes to be more proportional to state populations and redistributing sample size across age groups to ensure adequate coverage of older adults. The NSDUH questionnaire was revised in 2015 – relevant changes included updating the prescription drug misuse module (detailed below); additional changes are detailed in NSDUH documentation (Center for Behavioral Health Statistics and Quality, 2015). This study was deemed exempt from review by RAND’s IRB.

Measures

Race/ethnicity:

The NSDUH provides a 7 category race/ethnicity variable classifying respondents as: non-Hispanic White; non-Hispanic Black; Hispanic; non-Hispanic Asian; non-Hispanic Native American/Alaskan Native; non-Hispanic Native Hawaiian/Other Pacific Islander; or some other race/multiracial.

Heroin use:

Past-year heroin use was defined as any heroin use in the past 12 months.

Prescription opioid pain reliever misuse:

Prior to 2015, past-year prescription opioid pain reliever misuse was assessed with the item: “How long has it been since you last used any prescription pain reliever that was not prescribed for you or that you took only for the experience or feeling it caused?” Beginning with 2015, misuse was asked as “Have you ever, even once, used any prescription pain reliever in any way a doctor did not direct you to use it?”

Demographics:

Covariates include age (categorized as: 12–17, 18–25, 26–34, 25–49, and 50+ years old) and gender (male, female).

Analysis

Analyses were conducted with respect to five time periods: 1999–2002, 2003–2006, 2007–2010, 2011–2014, and 2015–2018. Empirical descriptive statistics were used to determine the number of race/ethnic categories that each analysis could support. Based on the threshold of at least 30 positive observations per time period, 6 race/ethnic categories were selected for prescription opioid misuse analyses (White; Black; Hispanic; Asian; Native American; other race/multiracial) and 4 categories were selected for heroin analyses (White; Black; Hispanic; other race/multiracial). We report both unadjusted and age- and gender-adjusted prevalence estimates of heroin use and prescription opioid misuse, stratified by racial/ethnic group and time period. Unadjusted results are survey-weighted prevalences; adjusted results were obtained as predictive margins from regression models that included age and gender as covariates. Specifically, for both past-year heroin use and prescription opioid misuse, Poisson regression with robust variance was used to estimate the adjusted risk ratio (aRR) by race/ethnicity. Models included race/ethnicity interacted with time period, as well as age and gender. Age- and gender-adjusted prevalence estimates of heroin use and prescription opioid misuse for each time period were obtained as predictive margins, using Stata’s margins command, specifying the -over- option for time period. Additionally, we present aRRs for each time period and racial/ethnic group, specifying White individuals as the reference group. In subsequent models, we examine race-specific linear trends in RRs across time periods by specifying time period as a continuous, rather than categorical, variable; given 2015 survey revisions, we include a test of time period trends in prescription opioid misuse from 1999–2014 as a sensitivity analysis. All models used NSDUH analytic weights, which were normalized by rescaling weights to the same population total in each year (2007 U.S. population) to weight each year equally. This study is novel in its approach of pooling across 20 years of NSDUH data; while the NSDUH has undergone revisions during the study period, our primary focus was on relative differences between groups within a given time period and we do not expect survey revisions to differentially affect racial/ethnic groups. Analyses were conducted in Stata version 15.1.

3. RESULTS

3.1. Racial/ethnic differences in prescription opioid misuse

We present both unadjusted (Table 1) and age- and gender-adjusted (Table 1, Figure 1) prevalence estimates for past-year prescription opioid misuse. As age distributions and, to a lesser extent, gender distributions differ across race/ethnic groups, adjusted prevalence estimates are of primary interest. We note that, relative to unadjusted estimates, the adjusted estimates were larger for non-Hispanic White individuals and smaller for all other groups, likely reflecting that age distributions skew older for White individuals relative to other groups.

Table 1.

Unadjusted and age- and gender-adjusted prevalence estimates for past-year prescription opioid misuse and heroin use, by race/ethnicity and time period.

	1999–2002			2003–2006			2007–2010			2011–2014			2015–2018
	Unadj. %	Adj. %		Unadj. %	Adj. %		Unadj. %	Adj. %		Unadj. %	Adj. %		Unadj. %	Adj. %
PRESCRIPTION OPIOIDS
Non-Hispanic White	3.7%	3.8%		5.2%	5.6%		5.3%	5.7%		4.4%	4.8%		4.4%	4.7%
Non-Hispanic Black	2.8%	2.6%		3.5%	3.3%		3.8%	3.6%		4.2%	3.9%		3.8%	3.5%
Hispanic	4.0%	3.4%		4.9%	4.1%		4.4%	3.7%		4.6%	3.8%		4.1%	3.5%
Asian	2.0%	1.8%		2.4%	2.2%		2.8%	2.6%		2.0%	1.8%		1.7%	1.6%
Native American	5.0%	4.6%		7.1%	6.6%		8.1%	7.7%		6.7%	6.3%		5.5%	5.1%
Other race/multiracial	5.5%	5.0%		5.2%	4.8%		6.6%	6.1%		5.6%	5.0%		5.9%	5.3%
HEROIN
Non-Hispanic White	0.16%	0.16%		0.15%	0.16%		0.23%	0.25%		0.33%	0.36%		0.38%	0.41%
Non-Hispanic Black	0.22%	0.21%		0.24%	0.23%		0.20%	0.19%		0.22%	0.21%		0.30%	0.28%
Hispanic	0.23%	0.19%		0.28%	0.23%		0.19%	0.15%		0.21%	0.17%		0.24%	0.21%
Other race/multiracial	0.09%	0.09%		0.09%	0.08%		0.10%	0.09%		0.14%	0.12%		0.16%	0.15%

Note: All estimates are obtained from regression models that account for NSDUH survey design. Prevalence estimates were obtained as predictive margins.

Figure 1.

Age- and gender-adjusted prevalence and risk ratio estimates for past-year prescription opioid misuse, by race/ethnicity and time period.

Relative to White individuals, adjusted prevalence estimates of prescription opioid misuse were lower among Black, Hispanic, and Asian individuals across all time periods. While prevalence peaked for White individuals at 5.7% (during 2007–2010), the corresponding peak rates were 3.9% among Black individuals (2011–2014), 3.8% among Hispanic individuals (2011–2014), and 2.5% among Asian individuals (2007–2010). Regression results indicate that, relative to White individuals, relative risk of prescription opioid misuse ranged from 0.33–0.48 for Asian individuals, from 0.59–0.81 for Black individuals, and from 0.65–0.87 for Hispanic individuals (Table 2). The relative difference between White individuals and both Hispanic and Asian individuals widened over time (Table 3), as the prevalence among White individuals was higher at baseline and grew significantly faster per four-year time period (linear time interaction: Hispanic aRR=1.03 [1.00–1.06]; Asian aRR=1.09 [1.02–1.16]).¹ Conversely, the gap between Black and White individuals significantly declined across time, as the prevalence among White individuals increased more slowly per time period than among Black individuals (linear time interaction aRR=0.95 [0.92–0.98]).²

Table 2.

Unadjusted and age- and gender-adjusted relative risk (RR) estimates (relative to non-Hispanic White individuals) for past-year prescription opioid misuse and heroin use, by race/ethnicity and time period.

	1999–2002		2003–2006		2007–2010		2011–2014		2015–2018
	Unadj. RR [95% CI]	Adj. RR [95% CI]	Unadj. RR [95% CI]	Adj. RR [95% CI]	Unadj. RR [95% CI]	Adj. RR [95% CI]	Unadj. RR [95% CI]	Adj. RR [95% CI]	Unadj. RR [95% CI]	Adj. RR [95% CI]
PRESCRIPTION OPIOIDS
Non-Hispanic Black	0.75 [0.68, 0.84]	0.66 [0.60, 0.74]	0.67 [0.61, 0.74]	0.59 [0.53, 0.64]	0.72 [0.65, 0.79]	0.63 [0.57, 0.69]	0.94 [0.85, 1.04]	0.81 [0.74, 0.90]	0.86 [0.79, 0.94]	0.74 [0.68, 0.81]
Hispanic	1.09 [0.98, 1.21]	0.87 [0.78, 0.97]	0.94 [0.86, 1.02]	0.73 [0.67, 0.80]	0.82 [0.75, 0.89]	0.65 [0.60, 0.70]	1.03 [0.94, 1.12]	0.80 [0.73, 0.87]	0.94 [0.87, 1.01]	0.73 [0.68, 0.79]
Asian	0.55 [0.42, 0.72]	0.48 [0.37, 0.62]	0.46 [0.37, 0.58]	0.39 [0.31, 0.49]	0.53 [0.41, 0.69]	0.47 [0.36, 0.61]	0.45 [0.36, 0.57]	0.38 [0.30, 0.48]	0.39 [0.32, 0.49]	0.33 [ 0.27, 0.41]
Native American	1.36 [1.03, 1.81]	1.20 [0.91, 1.59]	1.36 [1.06, 1.74]	1.18 [0.92, 1.51]	1.51 [1.16, 1.97]	1.35 [1.04, 1.75]	1.52 [1.21, 1.92]	1.32 [1.05, 1.66]	1.26 [1.02, 1.56]	1.09 [0.88, 1.34]
Other race/multiracial	1.50 [1.20, 1.88]	1.31 [1.05, 1.63]	0.99 [0.84, 1.15]	0.87 [0.75, 1.01]	1.23 [1.03, 1.47]	1.08 [0.91, 1.28]	1.26 [1.08, 1.47]	1.04 [0.89, 1.20]	1.35 [1.18, 1.55]	1.13 [0.98, 1.29]
HEROIN
Non-Hispanic Black	1.41 [0.87, 2.28]	1.28 [0.79, 2.09]	1.60 [1.02, 2.53]	1.44 [0.91, 2.28]	0.87 [0.55, 1.39]	0.78 [0.49, 1.24]	0.66 [0.45, 0.98]	0.59 [0.40, 0.86]	0.80 0.53, 1.20]	0.70 [0.46, 1.05]
Hispanic	1.46 [0.93, 2.30]	1.15 [0.73, 1.81]	1.92 [1.21, 3.04]	1.45 [0.91, 2.30]	0.81 [0.50, 1.33]	0.62 [0.38, 1.02]	0.63 [0.47, 0.86]	0.48 [0.36, 0.66]	0.65 [0.46, 0.91]	0.51 [0.36, 0.71]
Other race/multiracial	0.60 [0.34, 1.04]	0.52 [0.30, 0.91]	0.62 [0.34, 1.12]	0.53 [0.30, 0.96]	0.42 [0.25, 0.71]	0.37 [0.22, 0.62]	0.41 [0.24, 0.71]	0.35 [0.20, 0.60]	0.43 [0.29, 0.64]	0.36 [0.24, 0.53]

Note: All estimates are obtained from regression models that account for NSDUH survey design. Bold denotes RR that are significant at the 0.05 significance level.

Table 3.

Linear trend test comparing prevalence trends across racial/ethnic groups.

	Relative Risk	95% CI
Prescription Opioid Misuse
Primary linear trend test: 1999–2018
White vs Black	0.95	[0.92, 0.98]
White vs Hispanic	1.03	[1.00, 1.06]
White vs Asian	1.09	[1.02, 1.16]
White vs Native American	1.01	[0.95, 1.08]
White vs Other race/Multiracial	1.01	[0.96, 1.07]
Sensitivity test: 1999–2014
White vs Black	0.94	[0.90, 0.98]
White vs Hispanic	1.04	[1.00, 1.09]
White vs Asian	1.06	[0.96, 1.17]
White vs Native American	0.97	[0.88, 1.07]
White vs Other race/Multiracial	1.06	[0.98, 1.14]
Heroin Use
Primary linear trend test: 1999–2018
White vs Black	1.23	[1.06, 1.43]
White vs Hispanic	1.31	[1.15, 1.50]
White vs Other race/Multiracial	1.11	[0.94, 1.30]

Note: All estimates are obtained from regression models that adjust for age and gender and use survey sampling weights. Bold denotes RR that are significant at the 0.05 significance level. Time period is coded as a continuous variable (i.e., 1–5) to estimate linear trends. Estimated RRs represent the relative change in prevalence per four-year time period. Given 2015 survey revisions, we include a test of time period trends in prescription opioid misuse from 1999–2014 as a sensitivity analysis.

Relative to White individuals, rates of prescription opioid misuse were higher at every time period among Native American individuals and higher at nearly every time period among other race/multiracial individuals. Prevalence peaked during the 2007–2010 period for both Native American (7.7%) and other race/multiracial individuals (6.1%). Adjusted regression results indicate that relative risk of misuse was 1.35 times higher during 2007–2010 and 1.32 times higher during 2011–2014 among Native American individuals relative to White individuals (Table 2). Relative risk was significantly elevated for other race/multiracial individuals during 1999–2002 (aRR=1.31 [1.05–1.63]), relative to White individuals. Trend tests did not provide evidence of a linear increase or decrease, relative to White individuals, for either Native American or other race/multiracial individuals.

3.2. Racial/ethnic differences in heroin use

We present both unadjusted (Table 1) and age- and gender-adjusted (Table 1, Figure 2) prevalence estimates for past-year heroin use. As observed for prescription opioid misuse, adjusted estimates were larger than unadjusted estimates for White individuals and smaller for all other groups. During 1999–2002 and 2003–2006, heroin use was highest among Black and Hispanic individuals, although not significantly higher than among White individuals (Figure 2). Use among White individuals significantly surpassed all other groups in 2011–2015 and continued to increase, more than doubling from 1999–2002 (0.16%) to 2015–2018 (0.41%). Adjusted regression results indicate that the relative risk of heroin use was significantly lower among other race/multiracial individuals, compared to White individuals, at every time period (aRR ranged from 0.35 [0.20–0.60] to 0.53 [0.30–0.96]). Similarly, heroin use was significantly lower among Hispanic individuals during 2011–2014 (aRR=0.48 [0.36–0.66]) and 2015–2018 (aRR=0.51 [0.36–0.71]) and among Black individuals during 2011–2014 (aRR=0.59 [0.40–0.86]). Prevalence trends across 1999–2018 differed significantly by race/ethnicity, as heroin use increased 1.23 times faster per four-year time period among White individuals relative to Black individuals (linear time interaction aRR=1.23 [1.06–1.43]) and 1.31 times faster per four-year time period relative to Hispanic individuals (linear time interaction aRR=1.31 [1.15–1.50]; Table 3). Trend tests did not provide evidence of a linear increase or decrease, relative to White individuals, among other race/multiracial individuals.

Figure 2.

Age- and gender-adjusted prevalence and risk ratio estimates for past-year heroin use, by race/ethnicity and time period.

4. DISCUSSION

This study is the first to comprehensively characterize racial/ethnic differences in both past-year prescription opioid misuse and heroin use with respect to multiple racial/ethnic groups over a 20-year period. Our findings indicate that while heroin use has risen among all racial/ethnic groups, the demographics of heroin use have changed significantly such that prevalence is now highest among non-Hispanic White individuals. However, similar to recent trends in opioid mortality rates, some racial/ethnic differences in prescription opioid misuse have narrowed in recent years, as prevalence rose among Black individuals while declining among White individuals. Furthermore, this study is one of the first national studies of racial/ethnic differences in opioid misuse to include both Native American and Asian individuals, finding evidence that prescription opioid misuse is significantly lower, relative to White individuals, among Asian individuals and significantly elevated among Native American individuals.

A significant contributor to the observed racial/ethnic trends in prescription opioid misuse is likely the longstanding racial/ethnic differences in opioid prescribing (Meghani et al., 2012). Opioid prescribing has historically been highest amongst non-Hispanic White individuals (Hoffman et al., 2016; Ly, 2019), with Black and Hispanic patients significantly less likely to receive opioid analgesics across many settings, including the emergency department (Lee et al., 2019), outpatient settings (Ly, 2019), and after childbirth (Badreldin et al., 2019). This may have contributed to a higher risk for misusing prescription opioids among White individuals relative to other racial/ethnic groups. Prior qualitative and survey work attributed lower prescribing rates, in part, to perceptions among prescribers that Black individuals were at higher risk for misuse (Hoffman et al., 2016; Jones, 2000), yet our results underscore that this stereotype is unfounded, as rates of prescription opioid misuse have consistently been highest among non-Hispanic White individuals. The observed narrowing of racial/ethnic differences in prescription opioid misuse in recent years may reflect an attenuation of racial/ethnic differences in prescribing (Harrison et al., 2018).

While limited work to date has examined opioid use among either Native American or Asian individuals in national samples, our findings are consistent with previous work that has shown that rates of substance use behaviors are generally higher, relative to White individuals, among Native American individuals (Whitesell et al., 2012) and generally lower among Asian individuals (Fong and Tsuang, 2007). Risk factors for prescription opioid misuse among Native American individuals may differ from those among White individuals. While high rates of clinical prescribing of opioids among White individuals is recognized as a key risk factor for subsequent misuse, limited evidence exists regarding relative prescribing rates among Native American individuals (Meghani et al., 2012). As Whitesell et al. (2012) review, Native American individuals may experience unique risks arising from socioeconomic factors, constrained access to health services, and a legacy of historical trauma/assimilation. Overall, our findings demonstrate important heterogeneity among racial/ethnic groups that are often overlooked or lumped together, highlighting the importance of examining disparities among smaller racial/ethnic groups using national data.

The disproportionate increase in heroin use among White individuals may be an unintended consequence of policies aimed at reducing opioid prescribing and misuse (Cicero et al., 2014). Prior studies have linked state and federal policies targeting prescription opioids to a rise in heroin use (Alpert et al., 2018; Cicero and Ellis, 2015; Cicero et al., 2012; Evans et al., 2018), and qualitative work found that individuals switched to or supplemented with heroin when prescription opioids became too difficult/expensive to obtain (Cicero et al., 2014). Indeed, as the supply of prescription opioids contracted, the availability of heroin sharply increased, accompanied by a decline in price and an increase in purity (Office of National Drug Control Policy, 2014). The demographic shift that resulted in the relative prevalence of heroin use, in addition to prescription opioid misuse, now being highest among White individuals is consistent with the hypothesis of a substitution effect to heroin use among individuals with prescription opioid misuse.

Racial/ethnic trends in heroin use and prescription opioid misuse should also be considered in the context of Case and Deaton’s “deaths of despair” theory. They demonstrated that mortality rates for U.S. non-Hispanic White individuals aged 45–54 increased from 1998 through 2013 driven by increases in drug overdoses, suicides, and alcohol-related liver mortality (Case and Deaton, 2015). More recently, mortality rates have risen among all White individuals, not strictly middle-aged individuals (Centers for Disease Control and Prevention, 2020). They theorized that this excess mortality was attributable, in part, to rising economic hardships, including wage stagnation and diminishing employment opportunities, that were particularly impacting rural White adults (Case and Deaton, 2015; Case and Deaton, 2017). Recent work has demonstrated that these so-called “deaths of despair” are increasingly impacting racial/ethnic minorities, who may experience more pronounced economic hardship than White individuals, as life expectancy fell between 2014 and 2016 for Black and Hispanic individuals (Woolf and Schoomaker, 2019). A notable exception is that while overdose mortality rates among Native American individuals have risen dramatically over the past two decades, overall mortality rates have steadily decreased (Centers for Disease Control and Prevention, 2020). Overall, rising overdose fatalities certainly contribute to reversals in life expectancy gains observed across multiple racial/ethnic groups in the U.S., yet these mortality trends likely reflect a complex interplay of multiple factors (Muennig et al., 2018).

Opioid prevention and treatment efforts should be informed by the changing demographics of heroin use (i.e., rising prevalence among White individuals) as well as the changing geographic distribution, as heroin is increasingly a rural and suburban issue, rather than predominantly an urban issue (Cicero et al., 2014). Given that historic heroin harm-reduction and treatment programs (e.g., needle exchanges, methadone clinics) had focused on urban, racial/ethnic minority neighborhoods, recent growth in buprenorphine treatment has included rural and suburban areas (Stein et al., 2018). Yet emerging evidence points to growing racial/ethnic disparities in treatment access, with racial/ethnic minorities (particularly Black individuals) less likely to receive buprenorphine compared to their White counterparts (Abraham et al., 2019; Krawczyk et al., 2017; Lagisetty et al., 2019; Stein et al., 2018). It is imperative that all racial/ethnic groups have equitable access to prevention initiatives and treatment services.

4.1. Limitations of the current study

Several limitations warrant mention. Measures of opioid misuse are self-reported and may be subject to social desirability or recall bias. While the NSDUH race/ethnicity categories did not change over the study period, the survey items used to define this variable varied somewhat across years with the inclusions of several additional racial/ethnic subcategories; this may have marginally affected racial/ethnic group composition across survey years. For heroin analyses, the other race/multiracial group is heterogenous, comprising individuals who identify as Asian, Native American/Native Alaskan, Pacific Islander/Native Hawaiian, and multiracial. Additionally, the NSDUH’s prescription drug questions were revised in 2015 – see (Center for Behavioral Health Statistics and Quality, 2016) for details. Due to this redesign, prescription opioid misuse estimates from 2015 onward are not necessarily comparable to prior estimates; however, our primary focus was on comparisons across groups within a given time period. Furthermore, we conducted sensitivity analyses of time trends in prescription opioid misuse based on 1999–2014 data, excluding data after the 2015 redesign. Finally, the NSDUH excludes individuals who are incarcerated or experiencing homelessness; given the ongoing legacy of racial bias regarding Drug War policies and criminal justice system outcomes, our estimates may not quantify the full magnitude of racial/ethnic disparities.

CONCLUSION

This study – which characterized racial/ethnic differences in both past-year prescription opioid misuse and heroin use over a 20-year period, including among Native American and Asian individuals – provides important insights regarding the differential burden of the opioid crisis across racial/ethnic groups in the U.S. With the continued rise of fentanyl and other synthetic opioids, increasing heroin use among all racial/ethnic groups over the past decade is of notable concern. Opioid prevention and treatment initiatives should both be informed by the changing demographics of heroin use and seek to reduce opioid-related harms and expand treatment access equitably for all racial/ethnic groups.