Substance Use Treatment Utilization Among Individuals With Substance Use Disorders in the United States During the COVID-19 Pandemic: Findings on the Role of Polysubstance Use,Criminal Justice Involvement,and Mental Illness From the National Survey on Drug Use and Health

Abstract

This study used the National Survey on Drug Use and Health to assess a nationally representative sample (N = 4596) weighted to represent 35.2 million adults with DSM-5 criteria-determined substance use disorders (SUDs). This study explored substance use treatment utilization in 2020, emphasizing populations with high vulnerability (e.g., criminal justice involvement (CJI) through parole or probation, polysubstance use, severe mental illness, and HIV/STI). Substance use treatment was broadly defined (any inpatient, outpatient/doctor’s office, self-help/other for alcohol/drugs). Our results indicated that among adults with SUDs in 2020, 7 million (20%) had multiple SUDs, 1.75 million (5%) had CJI, 5.3 million (15%) had a severe mental illness, and 1.8 million (5%) had a diagnosis of HIV/STI in the last year. Only 7% of individuals with SUD sought any substance use treatment in the past year. CJI (aOR: 13.39, 95% CI: [7.82, 22.94]), serious mental illness (aOR: 3.27, 95% CI: [1.93, 5.55]), and having both 2 (aOR: 2.10, 95% CI: [1.29, 3.42]) or 3 or more SUDs (aOR: 3.46, 95% CI: [1.82, 6.58]) were all associated with a greater likelihood of receiving treatment. Marriage (aOR: 0.43, 95% CI: [0.25, 0.74]) and having an income twice the poverty threshold (aOR: 0.53, 95% CI: [0.29, 0.94]) were associated with reduced odds of receiving any substance use treatment. Compared to those 18 to 25, older individuals had increased odds (2-4 times) of receiving treatment. Interventions are crucially needed to increase access to treatment among those with SUDs.

Keywords

Substance use treatment COVID-19 pandemic polysubstance use criminal justice involvement mental illness

Introduction

Increasing substance use and drug overdose deaths during the COVID-19 pandemic

The COVID-19 pandemic era was marked by increases in the initiation of substance use and increases in the amount of substances used to cope with stress or other negative emotions.^1,2 During the pandemic, people who use substances were also more likely to use substances in isolation or to use dangerous drug combinations, including fentanyl, increasing the risk of experiencing an overdose.³ Consequently, rates of drug overdose deaths rose dramatically during this time, particularly in the late spring of 2020.^4,5 Access to substance use treatment is vital to reducing substance use, drug overdoses, and associated adverse social and health consequences related to substance use. It is important to understand treatment access during this tumultuous time, and to elucidate substance use treatment utilization rates by subpopulations most vulnerable to adverse consequences.

Substance use treatment during the COVID-19 pandemic

Nationally, the most common types of treatment in the United States are non-professional self-help groups (such as 12-step meetings), outpatient treatment (such as in a private doctor’s office or mental health clinic), and inpatient residential programs.⁶ Evidence suggests that these treatment modalities effectively promote reduced use or continuous abstinence from alcohol⁷ and other substances.^8,9 Yet, for many individuals with SUD, access to treatment programs was limited during the pandemic^10,11; during this time, there were many social and economic barriers to receiving needed substance use treatment and harm reduction services due to social distancing protocols and other pandemic-related concerns.^10
-13 While the rise of telehealth appointments may have increased access to treatment for some individuals, telehealth appointments may not have been accessible to those who did not have a reliable cell phone or internet access.¹²

Substance use treatment among vulnerable populations

Many individuals experience co-occurring vulnerabilities that may complicate access to treatment, such as involvement with the criminal justice system, having a serious mental illness, being diagnosed with sexually transmitted diseases (STIs), or human immunodeficiency virus (HIV), or engaging in polysubstance use. With regard to criminal justice involvement (CJI), it has been estimated that 24% of incarcerated people meet the diagnostic criteria for alcohol use disorder when first arriving at prison, while 51% meet the criteria for drug use disorders.¹⁴ These rates are much higher than those reported in the general population; in 2022, 17.3% of the general population was estimated to have alcohol use disorder, while only 9.7% had a drug use disorder.¹⁵ With regard to HIV, it has been estimated that of adults who are HIV-positive, 48% have an SUD.¹⁶ In 2021, 6.4 million American adults had both a substance use disorder, and a serious mental illness.¹⁷ Additionally, these vulnerabilities may often interact with one another; for example, individuals with both a substance use disorder and severe mental illness are much more likely to become involved in the criminal justice system than those with neither circumstance, or only one.¹⁸

Prior research on the impact of these vulnerabilities on treatment utilization has been mixed. Some studies suggest that individuals with these comorbidities are more likely to receive SUD treatment, while others studies suggest the opposite. For instance, extant research shows that those involved in the criminal justice system and who use opioids are less likely to be referred to medications for opioid use disorder (MOUD) programs.¹⁹ However, those with comorbid mental health disorders are more likely to receive SUD treatment both in the general population,²⁰ and for those with CJI.²¹ With regard to HIV status, veterans with SUD and HIV were 5% more likely to initiate SUD treatment and 10% more likely be retained in treatment at six months, when compared to their counterparts who were not HIV-positive, though these effect sizes are quite small and may have been inflated by the large sample size of the study.²² Polysubstance use may also influence treatment reception, though prior research has also been mixed. Pre-pandemic evidence suggests that individuals who used opioids in co-use with other substances,²³ or who were diagnosed with opioid use disorder (OUD) in tandem with another SUD,²⁴ were less likely to receive MOUD treatment than those who only used opioids, or were only diagnosed with OUD. On the contrary, veterans with polysubstance use disorders were more likely to receive psychiatric treatment than those with single substance use disorders,²⁵ and those who co-used opioids along with other substances were more likely to receive intensive treatment services than those who only used opioids.²⁴ Taken together, this evidence suggests that these 4 co-vulnerabilities (mental illness, CJI, HIV/STI diagnosis, and polysubstance use) may impact substance use treatment reception, both for better and for worse.

Current study

Given the lack of consensus in previous findings, and the additional complication of the COVID-19 pandemic, this study used a nationally representative sample of adults with SUD to explore substance use treatment utilization in 2020. The aims of this study are 2-fold: (1) to evaluate rates of SUD treatment utilization in 2020 and (2) to assess SUD treatment utilization among individuals who also had high co-occurring vulnerability (e.g., CJI through parole or probation, polysubstance use, severe mental illness, and HIV/STI diagnosis). The analyses controlled for demographic covariates that may have also been related to the likelihood of treatment reception.

Methods

Data overview

This study utilizes data from the 2020 wave of the National Survey on Drug Use and Health (NSDUH). NSDUH is a cross-sectional, deidentified, publicly available, and nationally representative dataset, and is collected annually by the Substance Abuse and Mental Health Services Administration. The authors had no access to information which could be used to identify individual participants. NSDUH employs a sampling method stratified by states, census tracts, census blocks, area segments, and specific dwelling units. This stratified sampling method allows for calculating person-level analysis weights, which can be used to weight observed data to approximate the United States population. It is important to note that while sample weighting can approximate the means in the population, sample weighting cannot introduce variation that is not included in the observed sample. Extensive information on sample weighting for this particular cohort is detailed in a report by NSDUH.²⁶

All participants were recruited in 2020. The analytical sample includes only those participants who were 18 years or older and who met the criteria for having a SUD, according to the DSM-5. In total, this analytical sample includes 4596 people, weighted to represent 35.2 million adults. While it may have been relevant to the aims of the paper to directly compare 2020 NSDUH rates of treatment utilization with rates from prior years, the NSDUH team cautioned against comparing 2020 data with data from prior years. This is because of methodological changes made in 2020, to accommodate for the circumstances of the COVID-19 pandemic.⁶

Measures

Substance use disorders and treatment reception

SUD diagnosis was determined using the Diagnostic and Statistical Manual of Mental Disorders, fifth edition (DSM-V) criteria. Participants were evaluated for SUD with regard to 11 possible substances: alcohol, marijuana, cocaine, heroin, hallucinogens, inhalants, methamphetamine, prescription pain relievers, tranquilizers, stimulants, and sedatives. In contrast with the DSM-V, participants were only presented the diagnostic items if they met a certain threshold of use for each individual substance. For alcohol and marijuana, diagnostic items were presented only if participants endorsed that they had used alcohol and marijuana in the last year for more than 5 days or if they did not endorse the specific number of days that they had used alcohol or marijuana in the previous year. The diagnostic items were presented for the other substances if the participant had previously endorsed any use in the last year.

Nine to eleven items were presented regarding the participant’s use of each substance, dependent on the specific substance in question. These items included “Some people who [use the specific substance] try to cut down or stop but find they can’t. Was there more than one time in the past 12 months when you tried but were unable to cut down or stop [using the substance]?” and “During the past 12 months, did you have any long-lasting or repeated physical health problems that were caused or made worse by [using the substance]?” To be classified with the particular disorder, participants had to have responded affirmatively to at least 2 of the diagnostic criteria. Only participants who met the diagnostic criteria for at least 1 substance were included in the analysis.

Participants who met diagnostic criteria for at least 1 SUD were then asked if they received any SUD treatment in the last year. Participants responded affirmatively or negatively to the question, “During the past 12 months. . .have you received treatment or counseling for your use of alcohol or any drug, not counting cigarettes?” Substance use treatment was broadly defined (any inpatient, outpatient/doctor’s office, self-help/other for alcohol/drugs).

Demographic variables

Participants reported on 10 demographic variables shown to be linked with substance use treatment utilization: age, sex, education, income, race, employment, area of residence, health insurance, marital status, and having children in their household.^27
-32

Age, sex, race/ethnicity, and familial demographic variables

In the NSDUH, the only sex/gender variable assessed is biological sex, which is male and female. Males were treated as the reference group. Age was assessed using the categorization of 18 to 25, 26 to 34, 35 to 49, 50 to 64, or 60 or older, using the youngest age group as the reference group. Though NSDUH provides data on multiple races and ethnicities, race/ethnicity was truncated into 4 categories to account for small subsample sizes. These categories are Hispanic, Non-Hispanic White, Non-Hispanic Black, and Other (including Non-Hispanic Native American/Alaskan Native, Non-Hispanic Native Hawaiian/Pacific Islander, Non-Hispanic Asian, and Non-Hispanic more than 1 race). Non-Hispanic Whites were treated as the reference group.

Regarding familial circumstances, self-reported marital status and the number of dependents in the home were treated in the analysis as covariates. The NSDUH marital status response categories were truncated to be not currently married (including widowed, divorced, separated, or never married) and currently married. Those who were not currently married were treated as the reference group. To assess dependents, participants self-reported the number of children under 18 years old living in their household. For the sake of this analysis, this variable was truncated into a binary variable of no children in the household or 1 or more children in the household. Those with no children were treated as the reference group.

Markers of socioeconomic status, and other covariates

Covariate predictors included variables indicative of participants’ socioeconomic statuses, namely educational attainment, income within the federal poverty threshold, and employment status. Household income was recorded categorically according to the federal poverty guidelines as determined by the U.S. Census Bureau. These categories are: living in poverty (having a family income below the federal poverty threshold), having an income up to 2 times the poverty threshold, and having an income more than 2 times the federal poverty threshold. Those living in poverty were treated as the reference group.

Regarding educational attainment, participants self-reported their highest completed level of education (less than a high school diploma, some college, or college graduate or higher); less than a high school diploma was treated as the reference group. Lastly, self-reported employment status was categorical, namely, full-time employed, part-time employed, unemployed (including those who had a job/volunteer work but did not work in the last week, those who were unemployed but looking for work, and those who did not have a job for some other reason) and other (including those who were disabled, keeping house full-time, in school or a training program, or retired). Those working full-time were treated as the reference group.

In addition, predictors included variables that may indicate the availability of SUD treatment in participants’ areas of residence, namely the rurality of the county where each participant lived. The 2013 Rural-Urban Continuum Codes by the USDA were used to identify if participants lived in a small metropolitan areas, a large metropolitan area, or a nonmetropolitan (or rural) area.³³ Those living in a rural area were treated as the reference group. Furthermore, participants reported whether they currently had health insurance or were uninsured, a factor in seeking health care. Those without health insurance were treated as the reference group.

Vulnerable circumstances

Four indices of comorbid vulnerability were treated as predictors of treatment reception, namely CJI, polysubstance use, mental illness severity, and HIV/STI status. CJI was operationalized as being on probation or parole at any time in the last 12 months. Polysubstance use was operationalized as meeting the DSM-5 criteria for multiple SUDs. Count variables were calculated for each of the 11 possible SUDs which participants may have endorsed. This variable was truncated as 1 SUD, 2 SUDs, and 3 or more SUDS because only 4% of the sample met the criteria for the latter category. Those with only 1 SUD were treated as the reference group.

Past-year mental illness severity was categorized according to predictive models developed in 2012 by the NSDUH team (see Appendix E of the 2020 National Survey on Drug Use and Health Public Use File Codebook for detailed information).²⁶ These categories are no mental illness, mild mental illness, moderate mental illness, and severe mental illness. Those with no mental illness were treated as the reference group.

Finally, participants self-reported whether they had been diagnosed with HIV/AIDS or had been diagnosed with an STI in the past year. Those without a diagnosis in the last year were treated as the reference group.

Analytical strategy

To address the first aim of the study, descriptive statistics were calculated using the weighted data, to assess the rates of treatment reception among the Americans with SUD during 2020. In addition, Rao-Scott Chi-squares were calculated to assess whether proportions of the demographic and vulnerable experiences differed between those who received treatment, and those who did not.

To address the second aim, a binary logistic regression assessed the relationship between sociodemographic covariates, vulnerable circumstances, and SUD treatment reception. The binary variable of treatment reception was treated as the outcome, while the sociodemographic and circumstantial variables were treated as predictors. All analyses were performed using the statistical software SAS.

Results

Sample characteristics

Both observed, and sample-weighted descriptive statistics of all study variables are detailed in Table 1. Of the 35.1 million American adults with SUDs in 2020, only an estimated 2.6 million (7%) participated in any treatment program (see Figure 1A). The most frequent treatment setting was outpatient with a medical provider (4.2%), followed by self-help groups such as Alcoholics Anonymous and Narcotics Anonymous (3.7%), care received at a hospital, such as at the emergency room or overnight as an inpatient (1.7%), and at a dedicated inpatient residential substance use treatment facility (0.8%). The most frequent SUD was alcohol use disorder (72.8%), followed by marijuana use disorder (37.1%) and prescription pain reliever use disorder (4.4%) (see Figure 1B).

Table 1.

Counts of study variables and chi-squares comparing those who did and did not receive treatment.

	Observed frequencies		Sample-weighted frequencies
	(N = 4596)		(N = 35.2 million)
	Overall		Overall			No treatment reception		Treatment reception		χ2; F(df)
	N	%	N	%	SE of %	N	%	N	%
Any treatment reception	286	6.2	2 556 498	7.3	0.75	-	-	-	-	-
Age
18-25	1785	38.8	7 588 767	21.6	0.88	7 279 314	95.9	309 453	4.1	9.3; 2.3(4)
26-34	1136	24.7	7 752 309	22.1	1.00	7 126 606	91.9	625 703	8.1
35-49	1208	26.3	10 049 121	28.6	1.16	9 260 622	92.2	788 499	7.9
50-64	315	6.9	6 230 332	17.7	1.46	5 570 609	89.4	659 723	10.6
65+	152	3.3	3 524 461	10.0	1.13	3 351 341	95.1	173 120	4.9
Sex (Female)	2332	50.7	15 705 015	44.7	1.39	14 507 242	92.4	1 197 773	7.6	0.2; 0.2(1)
Education
Less than high school diploma	326	7.1	3 346 300	9.5	1.02	3 025 559	90.4	320 740	9.6	7.6; 2.5(3)
High school/GED graduate	935	20.3	8 465 875	24.1	1.30	7 713 983	91.1	751 892	8.9
Some college	1619	35.2	12 738 953	36.2	1.31	11 689 512	91.8	1 049 441	8.2
College graduate or higher	1716	37.3	10 593 861	30.1	1.19	10 159 437	95.9	434 425	4.1
Income
Living in poverty	786	17.1	6 210 167	17.7	1.13	5 287 624	85.1	922 542	14.9	26.0; 13.0(2)***
Up to 2× federal poverty threshold	826	18.0	6 593 951	18.8	1.04	6 230 149	94.5	363 803	5.5
2× federal poverty threshold or higher	2953	64.3	22 258 915	63.3	1.35	20 990 173	94.3	1 268 742	5.7
Race
Non-Hispanic White	3216	70.0	23 326 377	66.4	1.38	21 702 592	93.0	1 623 785	7.0	3.8; 1.3(3)
Non-Hispanic African-American/Black	371	8.1	4 189 354	11.9	1.00	3 967 827	94.7	221 527	5.3
Hispanic	537	11.7	4 733 546	13.5	1.05	4 361 153	92.1	372 393	7.9
Non-Hispanic other	472	10.3	2 895 712	8.2	0.76	2 556 920	88.3	338 792	11.7
Employment
Unemployed	937	20.4	6 860 971	19.5	1.13	6 091 115	88.8	769 857	11.2	12.9; 4.3(3)**
Part-time	646	14.1	4 299 746	12.2	0.89	4 044 840	94.1	254 907	5.9
Full-time	2357	51.3	17 339 951	49.3	1.40	16 486 907	95.1	853 044	4.9
Other	656	14.3	6 644 320	18.9	1.21	5 965 630	89.8	678 691	10.2
Rurality
Non-metro	796	17.3	19 394 221	55.2	1.37	18 225 007	94.0	1 169 215	6.0	6.9; 3.4(2)*
Small metro	1749	38.1	11 271 951	32.1	1.26	10 161 050	90.1	1 110 901	9.9
Large metro	2051	44.6	4 478 817	12.7	0.91	4 202 435	93.8	276 382	6.2
Health insurance	3963	86.2	29 981 573	85.3	0.93	27 820 375	92.8	2 161 198	7.2	0.06; 0.06(1)
Currently married	1334	29.0	12 543 030	35.7	1.39	12 044 879	96.0	498 152	4.0	16.2; 16.2(1)***
Any children in household	1472	32.0	11 094 139	31.6	1.23	10 418 331	93.9	675 808	6.1	1.5; 1.5(1)
Criminal justice involvement	169	3.7	1 605 650	4.6	0.63	802 927	50.0	802 723	50.0	0.1; 0.1(1)
Past-year mental illness
None	2297	50.0	19 128 019	54.4	1.39	18 359 548	96.0	768 471	4.0	27.4; 9.2(3)***
Mild	887	19.3	6 777 500	19.3	1.21	6 158 590	90.9	618 910	9.1
Moderate	588	12.8	3 944 778	11.2	0.77	3 550 285	90.0	394 493	10.0
Serious	824	17.9	5 294 692	15.1	0.91	4 520 069	85.4	774 624	14.6
Last year diagnosis of STI/HIV	246	5.4	1 816 017	5.2	0.54	1 636 567	90.1	179 450	9.9	0.84; 0.84(1)
Number of SUDs
1 SUD	3603	78.4	28 143 227	80.1	1.04	26,701,784	94.9	1 441 443	5.1	49.9; 25.0(2)***
2 SUDs	811	17.6	5 753 463	16.4	0.98	4 991 038	86.8	762 425	13.3
3 or more SUDs	182	4.0	1 248 299	3.6	0.41	895 669	71.8	352 630	28.3

Abbreviations: F, F statistic; df, degrees of freedom; χ², Rao-Scott Chi-Square of sample-weighted frequencies comparing proportions between those who did and did not receive treatment.

P < .05. **P < .01. ***P < .001.

Figure 1.

(A) Frequencies of specific SUDs in 2020; (B) Frequencies of treatment settings among those with SUDs in 2020.

Concerning the demographic variables, an estimated 45% of the weighted sample was female. Only 36% of the sample was married, and only 32% had children living in their homes. Most of the sample (85%) had health insurance in the last year, and 64% were Non-Hispanic White. Age was well-distributed: 22% of the sample was 18 to 25 years old, 22% was 26 to 34, 29% was 35 to 49, 18% was 50 to 64, and 10% was 65 or older. Educational attainment was also well-distributed across the sample. About 10% of the sample was estimated to have not attained a high school diploma, 24% had a high school diploma or completed a GED, 36% attended some college, and 30% completed a college degree or more. Income was measured according to the federal poverty guidelines; around an estimated one-fifth of the sample reported living in poverty (18%) or having income up to 2 times the federal poverty threshold (19%). In comparison, 63% had an income twice the federal poverty threshold or higher. In terms of employment, most participants were estimated to be employed full-time (49%) or employed part-time (12%). Over half of the sample lived in a rural, nonmetropolitan area (55%), while 32% lived in a small metropolitan area, and 13% lived in a large metropolitan area.

With regard to vulnerable circumstances, 1.75 million (5%) adults with SUD were estimated to be involved with the criminal justice system. An estimated 7 million adults (20%) had multiple SUDs. Specifically, 16% of the sample met the diagnostic criteria for 2 SUDs, and 4% met the criteria for 3 or more SUDs. 46% of the sample had a mental illness in the last year. Specifically, 6.8 million adults (19%) had a mild mental illness, 3.9 million adults (11%) had a moderate mental illness, and 5.3 million (15%) had a serious mental illness. See Table 1 for the raw and weighted counts of all study variables.

Rao-Scott Chi-Squares were calculated to assess whether proportions among the categorical predictors differed between those who received treatment and those who did not (see Table 1). Concerning the demographic variables, there were significant proportional differences for income (χ²(2) = 26.0, P < .001), employment (χ²(3) = 12.9, P = .005), rurality (χ²(2) = 6.9, P = .03), and marital status (χ²(1) = 16.2, P < .001). Regarding vulnerable circumstances, proportions differed by treatment reception for mental illness severity (χ²(3) = 27.4, P < .001) as well as polysubstance use (χ²(2) = 49.9, P < .001).

Binary logistic regression

The results of the binary logistic regression are detailed in Table 2. The results indicated that treatment utilization was more prevalent among older age groups than young adults aged 18 to 25. Specifically, those aged 26 to 34 were about 2.4 times more likely to receive treatment than young adults (aOR: 2.39, 95% CI [1.26, 4.54]), those aged 35 to 49 were about 2.6 times more likely to receive treatment than young adults (aOR: 2.62, 95% CI [1.42, 4.82]), and those aged 50 to 64 were almost 5 times more likely to receive treatment than young adults (aOR: 4.50, 95% CI [2.03, 9.98]).

Table 2.

Binary logistic regression predicting treatment reception.

	Adjusted odds ratio^a	95% CI
Intercept	0.01***	[0.00, 0.04]
Age (Reference: 18-25 y old)
26-34 y old	2.39**	[1.26, 4.54]
35-49 y old	2.62**	[1.42, 4.82]
50-64 y old	4.50***	[2.03, 9.98]
65+	2.66	[0.88, 8.03]
Sex (Reference: Male)	0.97	[0.61, 1.55]
Education (Reference: Less than H.S. diploma)
High school/GED graduate	1.36	[0.67, 2.78]
Some college	1.23	[0.60, 2.48]
College graduate or higher	0.86	[0.35, 2.15]
Income (Reference: Living in Poverty)
Up to 2× federal poverty threshold	0.53*	[0.29, 0.94]
2× federal poverty threshold or higher	0.81	[0.45, 1.44]
Race (Reference: Non-Hispanic White)
Hispanic	1.18	[0.60, 2.33]
Non-Hispanic African-American/Black	0.87	[0.36, 2.07]
Non-Hispanic Other	1.44	[0.73, 2.84]
Employment (Reference: Full-time)
Unemployed	1.06	[0.61, 1.84]
Part-time	0.92	[0.43, 1.97]
Other	1.26	[0.65, 2.45]
Rurality (Reference: Nonmetro)
Small metro	1.36	[0.64, 2.87]
Large metro	1.55	[0.78, 3.06]
Currently has any health insurance	1.35	[0.71, 2.57]
Currently married	0.43**	[0.25, 0.74]
Any children in household	0.91	[0.58, 1.43]
Criminal Justice Involvement	13.39***	[7.82, 22.94]
Past-year mental illness (Reference: None)
Mild	1.53	[0.85, 2.78]
Moderate	1.92	[1.00, 3.72]
Serious	3.27***	[1.93, 5.55]
Last year diagnosis of STI/HIV	1.30	[0.59, 2.85]
Number of SUDs (Reference: 1)
2	2.10**	[1.29, 3.42]
3 or more	3.46***	[1.82, 6.58]

Adjusted odds ratios control for all other predictors in the model.

P < .05. **P < .01. ***P < .001.

Income predicted substance use treatment, such that those with income up to 2 times the poverty threshold were less likely to receive treatment than those living in poverty (aOR: 0.53, 95% CI [0.29, 0.94]); Currently married respondents were about half as likely to receive treatment than married respondents (aOR: 0.43, 95% CI [0.25, 0.74]).

In addition, those who were on probation or parole, had a serious mental illness, or had multiple SUDs were all significantly more likely to receive treatment than those who did not experience these vulnerabilities. Those with CJI were roughly 13 times more likely to receive treatment than those without CJI (aOR: 13.39, 95% CI [7.82, 22.94]). Those with a serious mental illness were roughly 3.3 times more likely to receive treatment than those without mental illness (aOR: 3.27, 95% CI [1.93, 5.55]). Finally, those with 2 SUDs were roughly 2 times more likely to receive treatment than those with only 1 SUD (aOR:2.10, 95% CI [1.29, 3.42]) and those with 3 or more SUDs were roughly 3.5 times more likely to receive treatment than those with only 1 SUD (aOR:3.46, 95% CI [1.82, 6.58]). Sex, education, race, employment, rurality, health insurance, having children in one’s household, and STI/HIV diagnosis did not significantly predict treatment reception.

Discussion

This study assessed substance use treatment utilization in 2020, emphasizing populations with high vulnerability. The study found that of the estimated 35.2 million adults with SUDs in 2020, only 7% (SE = 0.75) received any form of treatment for their substance use. This rate of treatment reception is seemingly a reduction compared to the year immediately before the pandemic (though it should be cautioned that pandemic-related factors preclude direct comparison with NSDUH data collected during non-pandemic years); in 2019, 10.2% of American adults with SUDs received any treatment.³⁴

Though very few individuals with SUDs received treatment during the pandemic, those with additional vulnerabilities were generally more likely to receive substance use treatment in 2020, than those who did not have additional vulnerabilities. For instance, those on probation or parole were 13 times (aOR = 13.39) more likely to receive treatment than those not involved with the criminal justice system. Those with serious mental illness were about 3 times more likely (aOR = 3.27) to receive treatment for substance use than those without mental illness. Those with 2 SUDs, or 3 or more SUDs, were 2 times more likely, and 3.5 times more likely (aOR = 2.10; aOR = 3.46) to receive treatment than those with only 1 SUD, respectively. The authors postulate that there are 2 possible explanations for this pattern across the vulnerabilities.

First, it is possible that in a time when many support services were limited or nonoperational, those with additional vulnerabilities were more likely to have contact with referral sources to SUD treatment, or to actively seek out treatment, than those without additional vulnerabilities. In general, referrals to treatment for SUDs can come from various sources, including primary care medical providers,^35,36 mental health professionals,³⁷ and the criminal justice system.²⁹ During the pandemic’s peak, access to these possible referral sources became very limited as many non-emergent primary care appointments and other routine medical screenings were canceled or postponed.^38,39 However, because people with serious mental health concerns are more likely to have contact with structured supports such as emergency medical personnel⁴⁰ and inpatient psychiatric treatment⁴¹ than those without serious mental illness, this contact may have served as a possible referral source to SUD treatment. Additionally, because those with serious mental illness very often experience distress and impeded functioning in their daily lives,⁴² this population may be more likely to purposefully seek treatment for SUDs—even during a pandemic—than those who are not experiencing mental illness.

The same logic may also apply to those with multiple SUDs. Overall, those who engage in polysubstance use or have multiple SUDs are more likely to have a lower degree of functioning across multiple domains^43,44 and a higher likelihood of various adverse physical health outcomes,⁴⁵ including overdose.⁴⁶ Because of these adverse and difficult experiences, individuals with multiple SUDs may have had more reason to initiate or continue contact with healthcare professionals, even when care services were limited or difficult to access. Hence, those with multiple SUDs may have been more likely to have been referred to, and receive, treatment than those who only met the diagnostic criteria for 1 SUD.

Second, the criminal justice system likely served as a relatively stable referral source to treatment. People with SUD on probation often must participate in treatment as a condition of their release from incarceration.⁴⁷ Failing to participate in treatment may lead to serious consequences, such as reincarceration.⁴⁷ Thus, it is not surprising that being involved with the criminal justice system predicted receiving SUD treatment.

There are multiple limitations to this study. First, NSDUH relies on participants’ self-report of substance use which may result in under-identifying those who meet the diagnostic criteria for an SUD. Additionally, this data did not specify which participants sought treatment, but were unable to access it. While NSDUH does inquire about reasons for nonengagement in treatment, these items are only presented to those who self-identify as needing treatment, regardless of their SUD status. It is possible that vulnerable participants with SUD were simply more likely to self-initiate treatment during the pandemic, when compared with those without these vulnerabilities.

Second, in 2020, NSDUH methodological changes to accommodate the COVID-19 pandemic rendered data incomparable with NSUDH data from prior years. These changes included an adjustment to the way SUDs were defined, collecting data only in the first and fourth quarters of the year (rather than throughout the year), collecting data virtually, and a lower survey response rate. Third, sample weighting is not likely to capture events that are rare within the population, especially in small subpopulations. For instance, in this study, older adults are represented by fewer than 200 participants. Given such a small subsample size, it is likely that rarer events—such as HIV and serious mental illness—are underrepresented for this subpopulation in the weighted sample. Fourth, cross-sectional data does not allow for causal inferences or to assess changes in treatment reception over time.

Finally, it should also be noted that treatment was broadly defined in this study to include outpatient care with a medical provider, self-help groups, any care received at a hospital, or an inpatient residential treatment facility (see Figure 1B). Because treatment was broadly defined, there is likely a high degree of variation in treatment quality and retention in treatment programs. For example, some participants may have attended very few self-help group meetings, while others may have attended meetings very consistently. Additionally, some participants may have received care at a hospital only briefly for an acute emergency event but did not pursue further treatment after their release from the hospital. Even with these limitations, this study boasts a large, nationally representative sample of people with SUDs and provides insight into substance use treatment utilization in 2020. Future studies should consider both the quality and retention of treatment services during the pandemic, especially for those with additional vulnerabilities.

Conclusion

As overdose deaths skyrocketed during the COVID-19 pandemic, unprecedented social and economic issues increased barriers to treatment for SUDs. This study adds to the body of research regarding the experience of those with SUDs during the pandemic. In particular, this study clarifies discrepancies in previous findings regarding the role comorbid vulnerabilities may play in treatment initiation for SUD. Using a nationally representative sample, this study explored substance use treatment reception rates during the pandemic among those with SUDs, especially emphasizing populations with high vulnerability. Populations with high vulnerability were more likely to receive substance use treatment, perhaps because they were more likely to have contact with referral sources, or because the criminal justice system may have mandated SUD treatment.

Future research should assess rates of SUD treatment reception, and the effect of co-vulnerabilities on treatment reception, post-pandemic, as well as explore how specific barriers to treatment initiation differ between those with and without significant comorbidities. This information could then be used to inform case management interventions, designed to help clients with their specific individual needs, including possible referrals to substance use treatment.

While populations with high vulnerability were more likely to receive substance use treatment, 93% of U.S. adults with SUDs received no treatment in 2020. Interventions to improve screening and referrals for substance use treatment are crucially needed. Additionally, more public funding and resources are needed to ensure that treatment is available to all those who need it, not just those with the most acute vulnerabilities This is especially needed in times of crisis, such as the pandemic.

Footnotes

Author Contributions

HBA wrote the original draft and edited subsequent drafts. ARS conducted the formal analysis. JG reviewed and edited the manuscript. GH reviewed and edited the manuscript. AAJ conceptualized and supervised the study, designed the data analysis plan, and reviewed and edited the manuscript.

Funding:

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the National Institute on Drug Abuse [T32 DA017629 (Trainee: Apsley, H.B.) & K01DA051715 (PI: Jones, A.A.)]. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institute on Drug Abuse or the National Institutes of Health.

Declaration of conflicting interests:

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

References

Stack

Leichtling

Larsen

, et al. The impacts of COVID-19 on mental health, substance use, and overdose concerns of people who use drugs in rural communities. J Addict Med. 2021;15:383-389.

Czeisler

MÉ

Lane

Petrosky

, et al. Mental health, substance use, and suicidal ideation during the COVID-19 pandemic - United States, June 24-30, 2020. Morb Mortal Wkly Rep. 2020;69:1049-1057.

Niles

Gudin

Radcliff

Kaufman

HW.

The opioid epidemic within the COVID-19 pandemic: drug testing in 2020. Popul Health Manag. 2021;24:S-43-S-51.

Ahmad

Cisewski

Rossen

Sutton

. Provisional Drug Overdose Data. National Center for Health Statistics; 2023. Accessed January 31, 2023. https://www.cdc.gov/nchs/nvss/vsrr/drug-overdose-data.htm#citation

Cartus

Macmadu

, et al. Forecasted and observed drug overdose deaths in the US during the COVID-19 pandemic in 2020. JAMA Netw Open. 2022;5:e223418.

Substance Abuse and Mental Health Services Administration. Key substance use and mental health indicators in the United States: results from the 2020 national survey on drug use and health. Published online 2021. https://www.samhsa.gov/data/sites/default/files/reports/rpt35319/2020NSDUHFFR1PDFW102121.pdf

Kelly

Humphreys

Ferri

Alcoholics anonymous and other 12-step programs for alcohol use disorder. Cochrane Database Syst Rev. 2020;3:CD012880.

Humphreys

Barreto

Alessi

, et al. Impact of 12 step mutual help groups on drug use disorder patients across six clinical trials. Drug Alcohol Depend. 2020;215:1-7.

Helbig

Pixa

Buehringer

Hoyer

Effectiveness of outpatient cognitive-behavioral therapy for substance use disorders - evaluation of a specialized outpatient clinic. Verhaltenstherapie. 2017;27:7-14.

10.

Banks

Paschke

, et al. Opioid use disorder and COVID-19: treatment and recovery factors among vulnerable populations at the intersection of two U.S. epidemics. J Psychoactive Drugs. 2022;54:300-308.

11.

Pagano

Hosakote

Kapiteni

, et al. Impacts of COVID-19 on residential treatment programs for substance use disorder. J Subst Abuse Treat. 2021;123:1-9.

12.

Walters

Seal

Stopka

Murphy

Jenkins

WD.

COVID-19 and people who use drugs - a commentary. Health Behav Policy Rev. 2020;7:489-497.

13.

Glick

Prohaska

LaKosky

, et al. The impact of COVID-19 on syringe services programs in the United States. AIDS Behav. 2020;24:2466-2468.

14.

Fazel

Yoon

Hayes

AJ.

Substance use disorders in prisoners: an updated systematic review and meta-regression analysis in recently incarcerated men and women. Addiction. 2017;112:1725-1739.

15.

Substance Abuse and Mental Health Services Administration. Key substance use and mental health indicators in the United States: results from the 2022 national survey on drug use and health. Center for Behavioral Health Statistics and Quality (CBHSQ), Substance Abuse and Mental Health Services Administration (SAMHSA). Published online 2023. https://www.samhsa.gov/data/sites/default/files/reports/rpt42731/2022-nsduh-nnr.pdf

16.

Hartzler

Dombrowski

Crane

, et al. Prevalence and predictors of substance use disorders among HIV care enrollees in the United States. AIDS Behav. 2017;21:1138-1148.

17.

Substance Abuse and Mental Health Services Administration. Key substance use and mental health indicators in the United States: results from the 2021 national survey on drug use and health. Center for Behavioral Health Statistics and Quality (CBHSQ), Substance Abuse and Mental Health Services Administration (SAMHSA). Published online 2022. https://www.samhsa.gov/data/report/2021-nsduh-annual-national-report

18.

Prince

Wald

Risk of criminal justice system involvement among people with co-occurring severe mental illness and substance use disorder. Int J Law Psychiatry. 2018;58:1-8.

19.

Mantha

Mauro

Martins

SS.

State criminal justice policy context and opioid agonist treatment delivery among opioid treatment admissions, 2015. Drug Alcohol Depend. 2020;206:1-6.

20.

Harris

Bharat

Glantz

, et al. Cross-national patterns of substance use disorder treatment and associations with mental disorder comorbidity in the WHO world mental health surveys. Addiction. 2019;114:1446-1459.

21.

Guastaferro

Koetzle

Lutgen-Nieves

Teasdale

Opioid agonist treatment recipients within criminal justice-involved populations. Subst Use Misuse. 2022;57:698-707.

22.

Kraemer

McGinnis

Fiellin

, et al. Low levels of initiation, engagement, and retention in substance use disorder treatment including pharmacotherapy among HIV-infected and uninfected veterans. J Subst Abuse Treat. 2019;103:23-32.

23.

Ford

Bart

Grahan

Shearer

Winkelman

TNA

. Associations between polysubstance use patterns and receipt of medications for opioid use disorder among adults in treatment for opioid use disorder. J Addict Med. 2021;15:159-162.

24.

O’Brien

Henke

Schaefer

Lin

Creedon

TB.

Utilization of treatment by Medicaid enrollees with opioid use disorder and co-occurring substance use disorders. Drug Alcohol Depend. 2020;217:1-7.

25.

Bhalla

Stefanovics

Rosenheck

RA.

Clinical epidemiology of single versus multiple substance use disorders: polysubstance use disorder. Med Care. 2017;55 Suppl 9 Suppl 2:S24-S32.

26.

Substance Abuse and Mental Health Services Administration. 2021 National survey on drug use and health (NSDUH): methodological resource book section 11: person-level sampling weight calibration. 2023. https://www.samhsa.gov/data/sites/default/files/reports/rpt40289/2021NSDUHmrbSamplingWgt041023.pdf

27.

Apsley

Vest

Knapp

, et al. Non-engagement in substance use treatment among women with an unmet need for treatment: a latent class analysis on multidimensional barriers. Drug Alcohol Depend. 2023;242:1-8.

28.

Shearer

Winkelman

TNA

Khatri

UG.

State level variation in substance use treatment admissions among criminal legal-referred individuals. Drug Alcohol Depend. 2022;240:1-7.

29.

Cano

Salas-Wright

, et al. Socioeconomic inequalities and black/white disparities in US cocaine-involved overdose mortality risk. Soc Psychiatry Psychiatr Epidemiol. 2022;57:2023-2035.

30.

Pro

Hayes

Brown

Goree

Zaller

Individual and health policy factors associated with positive heroin and opioid treatment response: United States, 2018. Am J Public Health. 2022;112:S66-S76.

31.

Jones

Park

Allen

, et al. Racial differences in overdose training, naloxone possession, and naloxone administration among clients and nonclients of a syringe services program. J Subst Abuse Treat. 2021;129:1-5.

32.

Lewis

Hoffman

Garcia

Nixon

SJ.

Race and socioeconomic status in substance use progression and treatment entry. J Ethn Subst Abuse. 2018;17:150-166.

33.

USDA. Rural-urban continuum codes: Documentation. Published 2020. Accessed March 31, 2023. https://www.ers.usda.gov/data-products/rural-urban-continuum-codes/documentation/

34.

Substance Abuse and Mental Health Services Administration. Key substance use and mental health indicators in the United States: results from the 2019 national survey on drug use and health. Published online 2020:114.

35.

Konkle-Parker

Williams

McAfee

Schumacher

Parker

Low-burden universal substance use screening in a primary care clinic to lower implementation barriers. J Behav Health Serv Res. 2023;50:108-118.

36.

Wiercigroch

Sheikh

Hulme

A rapid access to addiction medicine clinic facilitates treatment of substance use disorder and reduces substance use. Subst Abuse Treat Prev Policy. 2020;15:4.

37.

Martin

Cimini

Longo

Sawyer

Ertl

MM.

Equipping mental health professionals to meet the needs of substance-using clients: evaluation of an SBIRT training program. Train Educ Prof Psychol. 2020;14:42-51.

38.

Centers for Medicare and Medicaid Services. CMS releases recommendations on adult elective surgeries, non-essential medical, surgical, and dental procedures during COVID-19 response. Published 2020. Accessed February 28, 2023. https://www.cms.gov/newsroom/press-releases/cms-releases-recommendations-adult-elective-surgeries-non-essential-medical-surgical-and-dental

39.

Croswell

Corley

Lafata

, et al. Cancer screening in the U.S. through the COVID-19 pandemic, recovery, and beyond. Prev Med. 2021;151:1-7.

40.

Azar

KMJ

Petersen

Shen

, et al. Serious mental illness and health-related factors associated with regional emergency department utilization. Popul Health Manag. 2020;23:430-437.

41.

Ose

Kalseth

Ådnanes

Tveit

Lilleeng

SE.

Unplanned admissions to inpatient psychiatric treatment and services received prior to admission. Health Policy. 2018;122:359-366.

42.

Hayes

Chang

Fernandes

, et al. Functional status and all-cause mortality in serious mental illness. PLoS One. 2012;7:1-9.

43.

Bourgault

Rubin-Kahana

Hassan

Sanches

Le Foll

Multiple substance use disorders and self-reported cognitive function in U.S. adults: associations and sex-differences in a nationally representative sample. Front Psychiatry. 2022;12:1-9. doi:10.3389/fpsyt.2021.797578

44.

Kroll

Wunderli

Vonmoos

, et al. Socio-cognitive functioning in stimulant polysubstance users. Drug Alcohol Depend. 2018;190:94-103.

45.

Connor

Gullo

White

Kelly

AB.

Polysubstance use: diagnostic challenges, patterns of use and health. Curr Opin Psychiatry. 2014;27:269-275.

46.

Schneider

Park

Allen

Weir

Sherman

SG.

Patterns of polysubstance use and overdose among people who inject drugs in Baltimore, Maryland: a latent class analysis. Drug Alcohol Depend. 2019;201:71-77.

47.

Galvin

Davidson

Kleiman

Substance involvement and probation outcomes: evidence from a cohort study. J Drug Issues. 2022;52:329-348.