Estimating the impact of the COVID-19 pandemic on rising trends in drug overdose mortality in the United States, 2018-2021

Abstract

Purpose

During the COVID-19 pandemic, social and economic disruption such as social isolation, job and income losses, and increased psychological distress, may have contributed to the increase in drug-overdose mortality. This study aims to measure the impact of the pandemic on monthly trends in drug-overdose mortality in the United States.

Methods

We used the 2018–2020 final and 2021 provisional monthly deaths from the National Vital Statistics System and monthly population estimates from the Census Bureau to compute monthly mortality rates by age, sex, and race/ethnicity. We use log-linear regression models to estimate monthly percent increases in mortality rates from January 2018 through November 2021.

Results

The age-adjusted drug-overdose mortality rate among individuals aged older than or equal to 15 years increased by 30% between 2019 (70,459 deaths) and 2020 (91,536 deaths). During January 2018–November 2021, the monthly drug-overdose mortality rate increased by 2.05% per month for Blacks, 2.25% for American Indians/Alaska Natives, 1.96% for Hispanics, 1.33% for Asian/Pacific Islanders, and 0.96% for non-Hispanic Whites. Average monthly increases in mortality were most marked among those aged 15–24 and 35–44 years.

Conclusions

The COVID-19 pandemic had a substantial impact on the rising trends in drug-overdose mortality during the peak months in 2020 and 2021.

Introduction

Since 1999, more than 932,000 people have died due to drug overdose in the United States [1]. The age-adjusted death rate involving synthetic opioid (other than methadone), which largely consist of illicitly manufactured fentanyl, increased more than eleven fold, from 1.0 per 100,000 population in 2013 to 11.4 in 2019 [2]. In 2018, in 28 states and DC, fentanyl was detected in 73.9% of opioid-involved overdose deaths [3]. The COVID-19 pandemic, declared on March 11, 2020 [4], and subsequent disruptions in treatment, combined with increased psychological distress and social and economic stressors including social isolation and unemployment, may have fueled the opioid epidemic and rise in drug overdose mortality [5], [6], [7], [8]. During the pandemic, drug overdose [9], [10], [11], [12] and opioid-related overdose deaths [13], [14], [15], [16], [17], [18], [19], [20] have significantly increased. Median emergency department (ED) visits for all drug overdoses in the United States increased from 13,371 (711.1/100,000 ED visits) in 2019–15,604 (940.2/100,000 ED visits) in 2020 [12]. In 2020, 91,799 drug overdose deaths occurred, compared with 70,630 drug overdose deaths in 2019 [1,21]. The age-adjusted drug-overdose mortality rate increased by 31% between 2019 (21.6/100,000 population) and 2020 (28.3/100,000 population) [1].

There had been differences in drug overdose mortality by sex, age group, and race and ethnicity before the pandemic [21], [22], [23], [24]. The drug overdose mortality rate increased from 8.2 per 100,000 in 1999 to 29.6 in 2019 for males, and from 3.9 in 1999 to 13.7 in 2019 for females [21]. In 2019, adults aged 35–44 had the highest rate of drug overdose deaths (40.5/100,000), followed by adults aged 45–54, 25–34, 55–64, 15–24, and 65 and over [21]. Age-adjusted drug overdose mortality was highest among non-Hispanic Whites (25.9/100,000), followed by Blacks (23.7/100,000), American Indian and Alaska Natives (AIANs) (19.2/100,000), Hispanics (12.7/100,000), and lowest among Asian and Pacific Islanders (APIs, 4.1/100,000) in 2019 [22]. The average rate of increase in drug overdose mortality during 1999–2017 was the fastest for non-Hispanic Whites (7.6%/y), followed by AIANs (6.1%), APIs (5.9%), Blacks (3.6%), and Hispanics (3.3%) [23].

During the COVID-19 pandemic, drug overdose mortality increased more for males, than females [14,15]. There have been heterogeneous findings on differential effect of COVID-19 by age and race and ethnicity [[13], [14], [15],20]. Given disparities in social determinants of health, the pandemic might have disproportionately affected drug overdose mortality rates among racial and ethnic minorities [23,25]. One study found that racial and ethnic minorities experienced higher increases in drug overdose mortality than non-Hispanic Whites [20]. However, another study found that opioid overdose deaths continuously increased only for Blacks during March 2020 to March 2021 in Massachusetts, while it remained stable for AIANs, Asians, and Hispanics [16]. The heterogeneity of results among studies on the COVID-19 effect by age and race/ethnicity might arise from study period and regional differences.

This study extends previous analyses by estimating monthly percent changes in drug overdose mortality rates per 1 million population for U.S. adults aged 15 years and older during January 2018 through November 2021 by age, sex, and race/ethnicity. We also identified and analyzed peak periods in monthly drug overdose mortality during the pandemic.

Methods

The 2018–2020 final and 2021 provisional monthly deaths by age, sex, race/ethnicity, and cause of death were obtained from the National Vital Statistics System (NVSS)’s mortality files. NVSS data are collected and disseminated by the CDC's National Center for Health Statistics (NCHS) through the registration system, in which individual States and independent registration areas including the District of Columbia, New York City, and five territories are responsible for the registration of vital events - births, deaths, marriages, divorces, and fetal deaths [26]. Recent annual mortality datasets have included about 2.5 million records, which is based on the 2003 revision of the U.S. Standard Certificates and Reports [27]. Since final annual mortality data for a given year are typically released 11 months after the end of the calendar year [28], we used provisional mortality data from January to November, 2021, which were the latest data available from the NVSS [19,20,28]. The 2018–2020 monthly population estimates by age, sex, and race/ethnicity were obtained from the Census Bureau [29]. Monthly population estimates for 2021 were derived based on the annual growth rate of the population from 2020 to 2021.

We used deaths records for individuals aged 15 and older in the NVSS from January 2018 to November 2021. Drug overdose mortality rates were calculated by dividing the number of drug overdose deaths by the corresponding population and reported as deaths per 1,000,000 population. Drug overdose deaths were classified using the International Classification of Diseases, 10th Revision (ICD-10) underlying cause-of-death codes: X40-44 (unintentional), X60-64 (suicide), X85 (homicide), Y10-Y14 (undetermined intent) [21].

We estimated monthly trends in drug overdose mortality to give researchers and policymakers an early indication of shifts in mortality trends and to provide actionable information sooner than final mortality data [28]. We used log-linear regression models to estimate monthly percent increases in mortality rates from January 2018 through November 2021 by modeling the logarithm of the mortality rates as a linear function of time (month), which yielded monthly exponential rates of change in mortality rates. The estimated monthly percent changes can be used for public health surveillance to assess trends in the rate of change, providing how fast drug mortality rates increase or decrease. We computed age-adjusted drug overdose mortality rates per 1,000,000 population by month for the overall population, males, and females. Monthly drug overdose mortality rates were age adjusted by the direct method using the age distribution of the 2000 U.S. Standard Population. Standard errors of age-adjusted mortality rates were computed, and sex-, age-, race/ethnicity-specific mortality rate ratios were calculated. Statistical significance in group differences in monthly trend by sex, age group, and race/ethnicity were tested using a Hausman test after the seemingly unrelated estimation, suest, the estimations from all subgroups to be pooled together [30] for statistical significance. All analyses were conducted by Stata 17 [31].

Results

The drug-overdose deaths among individuals aged older than or equal to 15 years increased by 30% between 2019 (70,459 deaths) and 2020 (91,536 deaths). Figure 1 shows the monthly trend in drug overdose mortality per million population by age, sex, and race/ethnicity from January 2018 through November 2021. Average monthly drug overdose mortality among adult aged 35–44 was the highest, followed by the rates among adults aged 45–54, 25–34, and 55–64; individuals aged 15–24 and 65+ had lower rates than other groups (Fig. 1A). Average monthly drug mortality rates were higher for males than for females (Fig. 1B). Between February 2020 and May 2020, during the first peak of drug overdose mortality shown in the figures, the age-adjusted drug-overdose mortality rate increased by 48% overall, 52% for males, and 39% for females, while during the second peak of drug overdose mortality between October/November 2020 and March/April 2021, the increase in the drug overdose mortality rate was higher for females (34%) than for males (29%). During the first peak, drug-overdose mortality rates increased most rapidly for individuals aged 15-24 (63%), followed by individuals aged 25–34 (58%), 35–44 (52%), 45–54 (41%), 55–64 (36%), and 65+ (17%).

Fig. 1.

(A–C), Monthly drug overdose mortality rates per 1 million population among adults aged 15+ years by age, sex, and Race/Ethnicity, United States, January 2018 through November 2021. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Average monthly drug overdose mortality rates were highest for non-Hispanic Whites before the pandemic, followed by Blacks, AIANs, Hispanics, and APIs (Fig. 1C). During the pandemic, racial and ethnic minorities experienced a marked increase in drug overdose mortality. Between February 2020 and May 2020, the drug-overdose mortality rate increased by 52% for Blacks, 51% for AIANs, 49% for APIs, compared with 45% for non-Hispanic Whites and Hispanics. During the second peak between October/November 2020 and March/April 2021, API (53%) and AIANs (50%) showed higher increases in mortality, while other racial/ethnic groups experienced a 25%–40% increase in drug overdose mortality.

Table 1 shows estimated monthly percent changes in drug overdose mortality rates per one million population for U.S. adults aged 15 or older, using log-linear regression analysis. During January 2018–November 2021, the average monthly rate of increase in age-adjusted drug overdose mortality was 1.28% for adults aged 15 or older, 1.40% for males and 1.01% for females. Adults aged 15–24 (1.47%) and 35–44 (1.44%) had a faster average monthly increase in drug overdose mortality than other age groups. We found marked racial/ethnic disparities in the pace of monthly increases in drug overdose mortality. The monthly drug-overdose mortality rate increased by 2.05% per month for Blacks, 2.25% for AIANs, 1.96% for Hispanics, 1.33% for APIs, and 0.96% for non-Hispanic Whites.

Table 1.

Estimated monthly percent changes (EMPC) in age, sex, and race-specific drug overdose mortality rates per 1 million population for U.S. adults aged 15 years and older, January 2018 through November 2021

	Age 15+*	Age 15–24	Age 25–34	Age 35–44	Age 45–54	Age 55–64	Age 65 or older	Male*	Female*	NHW	Black	AIAN	API	Hispanic
EMPC (95% CI)	1.278 (1.084, 1.473)	1.468 (1.148, 1.790)	1.126 (0.897, 1.355)	1.442 (1.248, 1.637)	1.168 (0.970, 1.367)	1.258 (1.094, 1.422)	1.122 (0.957, 1.286)	1.401 (1.201, 1.602)	1.014 (0.829, 1.200)	0.961 (0.779, 1.144)	2.045 (1.829, 2.262)	2.251 (1.928, 2.576)	1.329 (0.978, 1.681)	1.955 (1.726, 2.185)
Slope† (SE)	0.013 (0.001)	0.015 (0.002)	0.011 (0.001)	0.014 (0.001)	0.012 (0.001)	0.013 (0.001)	0.011 (0.001)	0.014 (0.001)	0.010 (0.001)	0.010 (0.001)	0.020 (0.001)	0.022 (0.002)	0.013 (0.002)	0.019 (0.001)
Intercept	2.976	2.072	3.277	3.333	3.279	3.054	1.747	3.269	2.565	3.112	2.936	2.593	1.182	2.308
R-Square	0.789	0.645	0.676	0.826	0.749	0.836	0.801	0.808	0.721	0.705	0.886	0.808	0.552	0.863
% Increase (Feb-May 2020)	43.13	62.79	58.23	52.36	40.98	36.31	16.55	52.18	39.01	45.01	52.46	50.99	49.20	45.26
Rate ratio	NA	3.79	3.52	3.16	2.48	2.19	1.00	1.00	0.75	1.00	1.17	1.13	1.09	1.01

^⁎For overall adults aged 15+, males, and females, we calculated monthly age-adjusted drug overdose mortality rate using the 2000 US standard population

^†All slopes were statistically significant at P < .001 using log-linear regression

Notes. AIAN = American Indian and Alaska Native; CI = confidence interval; EMPC=EXP (slope)*100-100; NHW = Non-Hispanic White; SE = standard error.

Discussion

In this study, we found that COVID-19 disproportionately increased monthly drug overdose deaths among males, younger age groups, and racial and ethnic minorities. We also found that the rate of increase in drug overdose mortality during the pandemic was highest among the youngest group, that is, individuals aged 15–24. Our findings on age-group differential effects of the pandemic were consistent with the previous finding on the increase in opioid overdose deaths in Nevada that was mostly attributed to deaths among individuals aged 10–29 [13]. However, our findings differed from previous findings that there was no significant difference in the impact of COVID-19 on opioid overdose deaths by age group during 2017 and 2020 in Ohio [15], or that adults aged 30–40 and 50–60 were most affected in Milwaukee [14].

Our finding regarding higher rates of increase in drug overdose mortality among males compared to females are consistent with previous findings [14,15]. Our findings based on racial and ethnic stratified models are consistent with the previous studies which showed that Blacks, Hispanics, and AIANs experienced higher increases in drug overdose mortality than non-Hispanic Whites during the pandemic [15,20]. For example, a recent study of annual trends found that drug overdose mortality increased for non-Hispanic Blacks from 24.7 per 100,000 population in 2019–36.8 in 2020 (a 48.8% increase), for AIANs from 28.9 in 2019 to 41.4 in 2020 (a 43.3% increase), and for Hispanics from 12.4 in 2019 to 17.3 in 2020 (a 40.1% increase), compared with rates for non-Hispanic Whites from 25.0 in 2019 to 31.6 in 2020 (a 26.3% increase) [20].

Drug overdose mortality, in our study, showed the greatest increase from February to May 2020 and the greatest decrease from May 2020 to October/November 2020. The decreasing trend might be explained by implementation of temporary policies by federal and state governments to improve access to substance use disorder treatment. For example, the Centers for Medicare & Medicaid Services allowed telehealth reimbursement for Medicare beneficiaries under the 1135 waiver authority and Coronavirus Preparedness and Response Supplemental Appropriations Act on March 6, 2020 [32]. Some states required telehealth visits with Medicaid or private insurance to be paid at the same rate as analogous in-person visits, allowed for behavioral telehealth with waving requirements on prior in-person contact, relaxed state privacy laws, and allowed NPs and PA to obtain buprenorphine prescribing waivers [33,34]. Further studies are needed to estimate the impact of each intervention on access to care and drug mortality changes.

In our study, drug overdose mortality remained the highest among Blacks since August 2019 and the mortality rates among AIANs surpassed and remained higher than Whites since March 2021. Racial/ethnic disparities in drug overdose deaths might be explained by existing socioeconomic and health disparities among racial/ethnic groups and structural racism, in conjunction with COVID-19 [23,25]. For example, non-Hispanic Blacks and AIANs were more likely to experience job-related income loss, food insecurity, and depression than non-Hispanic Whites during the pandemic [7,35,36]. In particular, Blacks were less likely to have mitigation effects on mental health through government interventions such as stimulus fund or unemployment insurance [35]. Given higher rates of illicit drug use among Blacks and AIANs than Whites [37]. and lack of access to medications for opioid use disorder in short-term residential treatment [38]. it is important to implement appropriate policy interventions on prevention, treatment, and recovery support services to address underlying causes of racial and ethnic health disparities and to stem the rising tide in drug overdose deaths among U.S. adults.

Increased drug overdose mortality among youths might be explained by increase in use of illicitly manufactured fentanyl due to low cost, fentanyl contamination, and high potency [2,39]. Harm reduction intervention among young adults such as fentanyl test strips might be helpful in changing overdose risk behavior [40]. During the pandemic, adolescents might have experienced increased psychological stress and substance use disorders (SUDs) through social isolation due to school closure and social distancing, and their transitional stage of life course with psychiatric vulnerability [41]. According to Substance Abuse and Mental Health Services Administration, only about 3.5% of adolescents aged 12–17 and 3.7% of youth aged 18–25 received SUD treatment at a specialty facility in 2020, compared with 7.4% of adults aged 26 or older among those who needed substance use treatment [42]. Considering that the COVID-19 pandemic might have increased SUDs, policymakers should address an existing lack of access to SUD treatment among adolescents and youth through interventions such as expanding telehealth services including medication for opioid use disorder [43]. Further studies need to focus on adolescents and young adults to evaluate the impact of the pandemic on their drug overdose, treatment, and mortality.

Limitations

This study has limitations. Provisional death counts might be underestimated relative to final counts since provisional counts are often incomplete and causes of death may be pending investigation [28,44]. Future studies might consider adopting the predicted provisional death counts, adjusted for delayed reporting [44]. and comparing with the final death counts when data are available. Moreover, there might have been misclassification of AIANs, Asians, and Hispanics on death certificates and in provisional mortality statistics, possibly leading to an underreporting of racial/ethnic disparities in mortality rates [28]. In particular, more than 40% of AIANs who self-identified as AIANs were misclassified as Whites on the death certificate, and correction for death certificate misclassification increased mortality rate ratios of AIANs to Whites [45]. Finally, although our study calculated drug overdose deaths by combining unintentional, suicide, homicide, and undetermined deaths to obtain enough deaths for monthly trends, the trends might differ for specific causes of drug overdose mortality. The proportion of unintentional deaths due to drug overdose was 87%–88% during 2018–2019 and increased to 91%–92% during 2020–2021. The proportion of suicide deaths due to drug overdose was 7% during 2018–2019 and decreased to 4%–5% during 2020–2021. Future studies might consider analyzing trends in drug overdose mortality from specific causes during the pandemic.

Conclusions

The COVID-19 pandemic had a disproportionate impact on racial and ethnic minorities and individuals in younger age groups in the context of rising trends in drug-overdose mortality during the peak months in 2020 and 2021. Drug-overdose mortality rates increased faster among racial and ethnic minorities compared to non-Hispanic Whites.

CRediT authorship contribution statement

Hyunjung Lee: Conceptualization, Methodology, Software, Formal analysis, Data curation, Visualization, Validation, Writing – original draft, Writing – review & editing. Gopal K. Singh: Conceptualization, Methodology, Validation, Writing – review & editing.