Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2026 May 2.
Published before final editing as: Exp Clin Psychopharmacol. 2026 Apr 30:10.1037/pha0000844. doi: 10.1037/pha0000844

Exploring Risk Tolerance Among Individuals Who Use Opioids

Sahana Lothumalla 1, Devin C Tomlinson 1, Maya Campbell 1, Mary Jannausch 1, Chelsea Wilkins 1, Nathan Menke 1, Maureen A Walton 1,2,3, Erin E Bonar 1,2,3, Jason Goldstick 3,4, Lewei A Lin 1,2,5, Lara N Coughlin 1,2,3
PMCID: PMC13134685  NIHMSID: NIHMS2148529  PMID: 42060407

Abstract

Background:

Higher overdose risk may be influenced by risk tolerance, which may influence risky behaviors for individuals who misuse opioids and/or stimulants. We investigate overdose-related risk tolerance in this population.

Methods:

Adults (18–75 years; N=181) with past-month opioid misuse and phone access, remotely completed a questionnaire on demographics, substance use, preferred opioid, and an overdose probabilistic discounting task (i.e., risk tolerance). Regression models compared opioid preferences (street vs. prescription) and/or co-stimulant use (vs. no use), on the outcome of risk tolerance.

Results:

In the sample, 53% and 47% preferred to use/misuse street and prescription opioids, respectively. Overdose risk tolerance was significantly higher for individuals who use street opioids compared to individuals who use prescription opioids (p<0.001). Risk tolerance was significantly higher among those who used fentanyl (Area Under the Curve[AUC] M±SD=0.56±0.23) compared to heroin (AUC M±SD=0.44±0.24; p=0.031), prescription opioids (AUC M±SD=0.24±0.21; p=0.032), and medication for opioid use disorder (AUC M±SD=0.38±0.27; p=0.004). Street opioids (F=35.09; p<0.0001) and stimulant use severity (F=4.51; p=0.035) were significantly associated with increased risk tolerance.

Conclusions:

Individuals who prefer fentanyl and those with high stimulant-use severity show the highest overdose risk tolerance and should continue to be prioritized for interventions to reduce overdose risk.

Keywords: opioids, stimulants, fentanyl, substance use

1). Introduction

Drug overdose has stood as the leading cause of unintentional deaths in the United States since the mid-2010s (Judd et al., 2023; National Institute on Drug Abuse, 2024). The Centers for Disease Control and Prevention indicates that over 107,000 drug overdose deaths occurred in the United States during 2023 (CDC, 2024). Although overdose deaths from synthetic opioids, such as fentanyl, have begun to decrease recently, overdose deaths associated with psychostimulants (e.g., methamphetamine, cocaine) have continued to increase (CDC, 2024), and synthetic opioids continue to be the major driver of overdose deaths. Although the majority of these deaths may involve intentional polysubstance use, some of these deaths reflect the increasingly blurred boundary between opioid and stimulant use, as fentanyl contamination within stimulant supplies has led to unintentional opioid exposure and overdose among individuals who do not identify as people who use opioids (LaRue et al., 2019; O’Donnell et al., 2020). This evolving drug landscape underscores the need to consider the intentional and inadvertent patterns of polysubstance exposure when examining current overdose trends.

Opioids encompass both synthetic and natural substances with varying properties, including prescription and street opioids. Both prescription opioids (e.g., oxycodone) and street opioids (e.g., illicitly manufactured fentanyl/heroin) are highly addictive (National Institute on Drug Abuse, 2021) and contribute to the widespread polysubstance use epidemic (Phillips et al., 2017). In the past five years, an uptick in co-use of stimulants and opioids has accelerated overdose fatalities (Ahmed et al., 2022; US Department of Health and Human Services, 2024). In 2022, roughly 43% of drug overdose-related deaths involved both opioids and stimulants (CDC, 2024), underscoring the high prevalence of polysubstance use among people who overdose.

Understanding decision-making around the use of potentially lethal substances in the context of high-potency opioids and stimulants is key to informing public health messaging around risks associated with substance use. Prior research shows that substance use disorders (SUDs) alter decision-making, disrupting risk tolerance (i.e., willingness to accept potential losses or uncertainties in order to pursue a specific goal) (Erickson, 2018). One study found that individuals who use heroin have lower loss aversion relative to those who do not use substances, suggesting that those who use heroin are less sensitive to potential negative outcomes (Ahn et al., 2014). These results suggest that individuals who misuse opioids may have an increased tolerance to engage in risky behaviors despite being aware of the consequences. Moreover, the study also found that heroin-dependent individuals demonstrated reduced loss aversion, whereas amphetamine-dependent individuals exhibited heightened reward sensitivity relative to healthy controls (Ahn et al., 2014). These findings suggest that opioid use and stimulant use may impair decision-making through diminished sensitivity to potential losses, while increasing responsiveness to rewards.

Among people who use substances, one way to measure risk tolerance in the context of uncertainty and specific to the consequence of overdose is the overdose probabilistic discounting task. The overdose probabilistic discounting task evaluates the influence of hypothetical fatal overdose risk on the likelihood of substance use (Dolan et al., 2021). In a past study, an overdose probabilistic discounting task was used to evaluate the influence of substance use on risk assessment for 69 adults who were currently using heroin. Results indicated that individuals who used heroin more frequently and/or exhibited more severe opioid dependence were less likely to discount the perceived risks of overdose, suggesting lower sensitivity to possible negative outcomes compared to those who use less frequently or demonstrate a lower dependence (Dolan et al., 2021). Unlike traditional behavioral economic paradigms, which often use hypothetical monetary rewards or losses to assess discounting, overdose probabilistic discounting incorporates fatal overdose risk, which mirrors real-world substance use contexts. Monetary-based discounting tasks, while useful, can miss important substance-specific decision-making patterns. Individuals often discount outcomes differently depending on what is at stake (commodity-specific effects), and using non-monetary outcomes, like overdose risk, better reflects the real decisions faced by people with substance use disorders (Exum et al., 2023; Rasmussen et al., 2024).

Understanding the nuanced relative risk tolerance amid uncertain outcomes among those who use opioids along with other substances, such as stimulants, can help inform avenues for harm reduction and overdose prevention. By identifying subgroups with higher tolerance for overdose risk, we may be able to tailor existing treatment and harm reduction efforts, such as medications for opioid use disorder (Degenhardt et al., 2024; Mattick et al., 2014), targeted naloxone distribution (Beaudoin et al., 2022), fentanyl test strip programs (Banerjee Gakrieger et al., 2018), and/or supervised consumption sites (Milaney et al., 2021) for opioid-associated overdose risk, and contingency management interventions (Khazanov et al., 2024) for stimulant-associated overdose risk. In clinical practice, integrating risk tolerance assessment into substance use screening may allow providers to match patients to interventions aligned with their behavioral profiles. Patients demonstrating elevated risk tolerance to fentanyl-associated overdose risk may be prioritized for fentanyl test strip education and referral to supervised consumption services, whereas those with elevated risk tolerance to stimulant-associated overdose risk may be directed toward contingency management programs. Embedding these treatment and harm reduction approaches within healthcare settings ensures that intervention delivery is informed by individualized risk profiles.

The primary goal of this exploratory study is to investigate overdose-related decision-making among individuals with opioid-involved polysubstance use (i.e., the use of opioids alongside other substances), with a particular focus on co-stimulant use to assess discrepancies in risk tolerance based on the type of opioid and frequency of usage.

2). Methods

2.1). Procedures

This paper focuses on a cross-sectional analysis of baseline data from a longitudinal study focusing on evaluating behavioral economic measures to understand opioid-involved polysubstance use (Campbell et al., 2025; Rodriguez et al., 2025). The study procedures received Institutional Review Board approval from the University of Michigan (HUM00229563). The study was not preregistered. Adults (N=197) were recruited to promote broad representation of people using multiple substances; participants were recruited in-person and through electronic healthcare records from a large university hospital, and nationally online through social media (Campbell et al., 2025; Rodriguez et al., 2025). Recruitment of the parent study began in August 2023 and was completed in October 2025 (Campbell et al., 2025). The first 200 enrolled participants were included in the current analyses, and 3 were removed due to failed authenticity checks (Rodriguez et al., 2025). Inclusion criteria were: 1) self-report past-month opioid use/misuse (score ≥1 on National Institute on Drug Abuse (NIDA) and Tobacco, Alcohol, Prescription medication (not as prescribed), and other Substance use (TAPS) Tool-2 Heroin or Prescription Opioids section); 2) self-report past-month use/misuse of another substance (score ≥ 2 on the NIDA TAPS-2 for alcohol OR score ≥ 1 on the NIDA TAPS-2 for another substance such as cannabis, cocaine, methamphetamine, sedative/sleep anxiety medication, prescription stimulants, etc.); 3) age 18–75; 4) have reliable phone access. Exclusion criteria were: 1) non-English speaking; 2) conditions precluding informed consent (e.g., acute psychosis); 3) identity unable to be confirmed (e.g., failure to confirm name, date of birth, and zip code before study assessments). Consistency checks specific to reported substance use around inclusion criteria were conducted between screening and baseline measures (Rodriguez et al., 2025).

2.2). Measures

The parent study focused on measures related to demographics, substance use (National Institute on Drug Abuse [NIDA], 2023), healthcare utilization ((Substance Abuse and Mental Health Services Administration (SAMHSA), 2022), motivations for use (Mahu et al., 2021), stigma behind use (L. R. Smith et al., 2016), and behavioral economic measures such as overdose probabilistic discounting (Dolan et al., 2021). For more detailed description of the parent study, see (Campbell et al., 2025). The present analysis focused specifically on measures related to demographics, substance use severity, preferred opioid, and overdose probabilistic discounting, as these variables addressed the present report’s primary aim of examining overdose-related risk tolerance.

2.2.1). Demographics

Descriptive characteristics assessed included age, sex, education (G. E. Smith et al., 1996), employment (National Institute on Drug Abuse, 2004), and income (Substance Abuse and Mental Health Services Administration (SAMSA), 2021). As part of our sample, we also collected race/ethnicity, sexual orientation, and housing status of our participants.

2.2.2). Substance Use Severity

Substance use severity based on the past four months was measured with NIDA-modified Alcohol, Smoking and Substance Involvement Screening Test (ASSIST) questionnaire for each of the following substances: cannabis, alcohol, cocaine, hallucinogens, street opioids, methamphetamine, as well as misuse of prescription opioids, stimulants, and sedatives ((Substance Abuse and Mental Health Services Administration [SAMHSA], 2021). Substance use severity scores of 0–3, 4–26, and 27 or more correspond to lower, moderate, and high risk, respectively. For the analyses on the preferred opioid and overdose probabilistic discounting task, individuals were dichotomized into groups of high risk vs. all others based on the NIDA-ASSIST scoring.

2.2.3). Preferred Opioid

Each participant was asked to identify a preferred opioid, which was categorized as either a street or prescription opioid. Fentanyl and heroin were designated as street opioids, and all other opioids (e.g., buprenorphine, codeine, hydrocodone, hydromorphone, methadone, morphine, oxycodone, and tramadol) were designated as prescription opioids. For a separate analysis, misused prescription opioids were further categorized by medications that can be used for opioid use disorder (MOUD) (i.e., buprenorphine and methadone), while the other misused prescription opioids (i.e., codeine, hydrocodone, hydromorphone, morphine, oxycodone, and tramadol) remained in the prescription opioid category.

2.2.4). Overdose Probabilistic Discounting Task

The overdose probabilistic discounting task was used to determine how the associated risk of death by overdose at different probabilities affected the decision to use or not use the preferred opioid (Dolan et al., 2021; Johnson & Bickel, 2008). Our study categorized participants to their preferred opioid and used the probabilistic discounting task to assess risk tolerances among preferred opioid categories. Participants rated the likelihood of using their preferred opioid based on how many other people have died from an overdose using the same preferred opioid. Participants selected a numeric response ranging from 0 (definitely would not use) to 100 (definitely would use) to rate the likelihood of using the preferred opioid in the given situations. The eight items in the task were based on the percentage of people who have hypothetically overdosed using the same batch of their preferred opioid (1%, 5%, 10%, 15%, 25%, 50%, 75%, 99%), and participants provided a numeric response for each of these percentages. The items were in the same order when presented in the survey.

Of the initial baseline sample (N=200), 3 were removed due to failure of authenticity checks (Campbell et al., 2025; Rodriguez et al., 2025). Of the N=197 participants, N=181 (92%) contributed discounting responses consistent with criteria used in previous reports on the overdose probabilistic discounting task (Dolan et al., 2021; Johnson & Bickel, 2008), which define the analytic sample for the present report. The criteria are: (1) use likelihood at one probability did not surpass previous, lower probabilities by 20% or more, and (2) use likelihood at 99% chance of overdose risk was not greater than use likelihood at 1% chance of overdose risk by greater than or equal to 10%. Seven participants failed only criterion 1, two participants failed only criterion 2, and seven participants failed both criteria.

Among the 181 participants with qualifying discounting responses, we derived outcome measures of overdose risk tolerance. Each participant’s responses (Y) for each probability of death by overdose (X) were divided into a series of trapezoids. The area under the curve (AUC) of each trapezoid was calculated via (x2 − x1)[(y1 + y2)/2], and all trapezoids for each participant were summed (range=(0, 1.0)). A higher AUC indicates an increased risk tolerance. The Area Under the Curve (AUC) was estimated by summing the trapezoids formed by quantities reported at neighboring levels of overdose mortality risk. A lower AUC indicates a decreased risk tolerance or lower likelihood of using the substance in the face of increasing risk for overdose.

2.2.5). Data Analysis

All analyses were conducted using SAS 9.4 (SAS Institute, Cary N.C.). First, in our primary analysis, AUC was compared between preferences for prescription and street opioid groups using a general linear regression (GLM). Second, AUC was derived from alternate categories: fentanyl, heroin, prescription opioids, and MOUD, and compared using a GLM. Lastly, an interaction between preferred opioid type and stimulant use severity was evaluated to compare AUCs. Specifically, we used a GLM to investigate the effect of preferred street/prescription opioid and stimulant severity (low/moderate risk and high-risk) on AUC. In a GLM, the significance of an individual effect is determined via a t-test. We report p-values and the effect size eta. Significance was defined as p<0.05. De-identified, study-specific data will be made available through the National Addiction & HIV Data Archive Program (NAHDAP), consistent with HEAL data sharing policies (HDP00944).

3). Results

3.1). Demographics, Substance Use, and Risk Tolerance among Individuals Using Street and Misusing Prescription Opioids

The overall sample (N=181) had a mean age of 41.4 years (SD= 9.7). More (N=106; 58.6%) identified as female or another gender than participants who identified as male (N=75; 41.4%). The majority of participants identified as White (N=148; 81.8%), followed by Black/African American (N=15; 8.4%), Asian (N=2; 1.1%), Native American (N=2; 1.1%), Pacific Islander (N=1; 0.6%), other race (N=1; 0.6%), and more than one race (N=10; 5.6%). In terms of ethnicity, 13 participants (7.2%) identified as Hispanic. Most participants reported never overdosing (N=148; 81.8%), while some (N=28; 15.5%) had overdosed once or twice, and few (N=5; 2.8%) had overdosed three or more times (p=0.0014). The overall sample was derived from social media recruitment (N=164), electronic health records (N=12), and in-person recruitment from a university hospital (N=5). For demographic characteristics by recruitment source, see Supplemental Table 1.

Supplemental Table 2 stratifies demographic characteristics by preferred opioid. Those who preferred street opioids were younger (M=40.0; SD=9.2) than those who preferred prescription opioids (M=43.1; SD=9.9; p=0.028). In this sample, more individuals who preferred street opioids earned < $25,000 per year (N=53; 55.2%) than those who preferred prescription opioids (N=34; 40.0%; p=0.04). More individuals whose preferred opioids are street opioids used street opioids daily/almost daily (N=76; 79.2%) compared to those who prefer prescription opioids (N=11; 12.9%; p<0.001). More individuals whose preferred opioid was a prescription opioid reported using prescription opioids daily/almost daily (N=27; 31.8%) compared to those whose preferred opioid was a street opioid (N=7; 7.3%; p<0.001).

Individuals who preferred street opioids reported higher average street opioid ASSIST scores (M=32.3, SD=7.5) than prescription opioid ASSIST scores for those who preferred prescription opioids (M=21.3, SD=11.1). This indicates that when participants’ risk scores aligned with their preferred opioid type, the AUC remained higher among those who use street opioids, reflecting greater overdose risk tolerance. More individuals (N=81; 84.4%) who prefer to use street opioids were considered to have high-risk street opioid substance use severity (based on ASSIST), compared to those (N=14; 16.5%) who prefer to use prescription opioids (p<0.001). More (N=31; 36.5%) individuals who prefer prescription opioids were considered high risk for prescription opioid severity compared to individuals (N=23; 24.0%) who prefer to use street opioids, although this difference is not statistically significant (p=0.09). Overall, the sample mean AUC was 0.40 (SD=0.28; Supplemental Table 2). The AUC for those who prefer to use street opioids was higher (M=0.52; SD=0.28) than those who prefer to use prescription opioids (M=0.28, SD=0.23; p<0.001; Figure 1; Supplemental Table 4).

Figure 1. Area Under the Curve by Preferred Opioid Type.

Figure 1.

Open in a new tab

Note: Individuals who preferred street opioids had a significantly higher AUC than individuals who preferred prescription opioids (i.e., more risk tolerant; M=0.52; M=0.28; p<0.001). Error bars represent standard error.

3.2). Comparing Street and Prescription Opioid Risk Tolerance

Opioid use was categorized based on use of fentanyl (N=62), heroin (N=34), MOUD (N=21), and prescription opioids (N=64). Those who preferred fentanyl had the highest AUC (M=0.56; SD=0.23; Figure 2; Supplemental Table 5), indicating the greatest risk tolerance for use. Fentanyl AUC was significantly higher than heroin (M=0.44, SD=0.24; p= 0.031), misuse of MOUD (M=0.38; SD=0.27; p= 0.004), and prescription opioids (M=0.24; SD=0.21; p=0.032). Individuals whose preferred opioid was the use of prescription opioids not as prescribed reported an AUC which was significantly lower than those whose preferred opioid was misuse of MOUD (p=0.032).

Figure 2. Area Under the Curve by Further Categorized Street and Prescription Opioids.

Figure 2.

Open in a new tab

*excluding Medications for Opioid Use Disorder

Note: Individuals who preferred fentanyl had a significantly higher AUC (M=0.56) than individuals who preferred heroin (M=0.44; p=0.031), medications for opioid use disorder (M=0.38; p=0.004), and prescription opioids (M=0.24; p=0.032). Error bars represent standard error.

3.3). Prescription/Street Opioid and Stimulant Co-Use

The co-use of opioids and stimulants (i.e., non-medical use of prescription stimulants, cocaine, and methamphetamine) was categorized by high and moderate/low risk stimulant use (based on ASSIST scores) in Supplemental Table 3. The interaction between opioid type (street vs. prescription) and high-risk stimulant use (vs. moderate/low-risk stimulant use) on AUC was not significant (p=0.13). Main effects analysis revealed opioid type (street vs. prescription) and stimulant severity (high vs. moderate/low) are significantly and independently associated with AUC. Individuals who prefer to use street opioids have significantly different AUC values than individuals who prefer to use prescription opioids (F=35.09; p < 0.0001; Figure 3; Supplemental Table 6). Individuals reporting high-risk stimulant use significantly differ from those with moderate/low risk (F=4.51; p=0.035; Figure 3).

Figure 3. Area Under the Curve Among Individuals Who Co-use Opioids and Stimulants.

Figure 3.

Open in a new tab

Note: Although the interaction between opioid type and stimulant severity was not significant (p=0.13), both opioid preference (F=35.09; p<0.0001) and stimulant severity (F=4.51; p=0.035) were independently associated with AUC. Error bars represent standard error.

4). Discussion

This exploratory study investigated overdose risk tolerance among individuals with opioid-involved polysubstance use, including assessing differences in risk tolerance based on the preferred opioid and co-use of stimulants. Risk tolerance and decision-making behaviors among participants revealed a greater risk tolerance for those who prefer to use street opioids. In particular, preference for fentanyl was associated with the highest AUC in both the street opioid category and when considered by itself when evaluated against heroin, prescription opioids, and MOUD. Higher AUC scores among those who prefer to use street opioids suggest that individuals may be more willing to accept the short-term effects (e.g., euphoria, relief, or avoidance of withdrawal) of opioids despite the potential subsequent harmful effects, including risk of overdose. To further differentiate risk, clinical settings could embed a brief overdose risk tolerance assessment into substance use screening to identify risk-tolerant individuals. Those with particularly high overdose risk tolerance could be prioritized for tailored interventions, including effective treatments for SUDs and targeted naloxone distribution (Beaudoin et al., 2022), referral to supervised consumption services (Milaney et al., 2021), and fentanyl and xylazine test strip education given the increasing presence of xylazine in the fentanyl drug supply (Banerjee Gakrieger et al., 2018; Johnson et al., 2021).

While the previous study (Dolan et al., 2021) assessed overdose risk tolerance among individuals with past-30-day heroin use, with probability-discounting tasks framed around heroin samples, our study expands this scope in several key ways. First, our sample size (N=181) is substantially larger than that of the prior study (N=69), increasing the generalizability of findings. Second, we distinguish fentanyl from other opioids, enabling comparisons of risk tolerance across fentanyl, heroin, prescription opioids, and MOUD. Finally, we incorporate stimulant co-use and stimulant use severity, which is highly relevant given that, although overall opioid-involved overdose deaths have begun to decline, deaths involving stimulants have continued to rise (CDC, 2024). This broader inclusion better reflects the current polysubstance use landscape and its implications for overdose risk tolerance.

Additionally, our findings also suggest that individuals who prefer to use street opioids or have high stimulant use severity have significantly higher risk tolerance than those who prefer prescription opioids or have low-risk severity. These preliminary findings suggest high stimulant use severity may be a predictive marker of increased risk tolerance. Past research indicates that stimulant use can heighten impulsivity (Moallem et al., 2018), but pharmacological and behavioral treatments can lower impulsivity (Kozak et al., 2019). In clinical practice, patients with high stimulant severity and opioid use may benefit from stimulant-specific treatment approaches such as contingency management programs (Khazanov et al., 2024), delivered alongside opioid-focused interventions, reflecting the polysubstance reality of overdose risk.

This study includes a few key limitations. First, these analyses should be considered exploratory and hypothesis-generating, particularly given the sample size for the various subgroups based on opioid preference. Additionally, the demographic composition of the sample may limit generalizability to more diverse populations, as the majority of participants identified as White and several racial and ethnic groups were underrepresented. Future work should prioritize recruitment of more racially and ethnically diverse samples to strengthen external validity and evaluate observed patterns of risk tolerance across diverse groups. Nonetheless, this novel data may inform future work exploring public health initiatives and their impact on overdose risk tolerance. Another limitation is that responses in the baseline survey are self-reported, which may introduce demand characteristics or recall bias; however, the self-administered, standardized assessments and assurances of confidentiality likely enhanced the validity of responses. Lastly, we operationalized fentanyl as a street opioid; however, we acknowledge that fentanyl use could also be the misuse of prescription fentanyl. Future investigations may examine differences in risk tolerance between prescription and non-prescription fentanyl. In addition, future work could incorporate objective measures (e.g., toxicological testing from healthcare provider/rehabilitation facilities data) to validate self-reported substance use, and verify types of substances used (i.e., confirmed presence of opioids/stimulants).

The present study highlights the variability in risk tolerance among those who use different opioids (prescription or street) and co-use stimulants. These initial findings can inform future public policy initiatives to create tailored campaigns and interventions to reduce opioid misuse/stimulant co-use and overdose deaths. Future research may explore the interplay between demographic factors, available resources, and risk tolerance to create targeted interventions for at-risk populations. Incorporating behavioral economic measures like the overdose probabilistic discounting task into clinical workflows could help shape prevention and harm reduction strategies by aligning them with patients’ behavioral risk profiles, rather than solely by substance type or reported use.

Supplementary Material

Supplemental Material

Public Significance Statement.

We investigate overdose-related risk tolerance in individuals who misuse opioids alongside other substances. We identified fentanyl preference and high stimulant use severity as the strongest independent predictors of overdose risk tolerance. These findings underscore the importance of delivering effective overdose prevention and substance use treatment interventions to these populations who are at the highest risk of fatal opioid and polysubstance-associated overdose.

Disclosures and Acknowledgments

Dr. Bonar (The Kathy Fant Brzoznowski Research Professor in Behavioral Health Technology Innovations) and Dr. Walton (The Toby Brzoznowski Research Professor in Behavioral Health Technology Innovations) are generously supported through the Broznowksi family. Dr. Tomlinson’s time was funded by National Institute on Alcohol Abuse and Alcoholism (NIAAA) Grant #T32AA007477 was awarded to DCT. Funding for this study was provided by NIDA through an NIH HEAL Initiative Award (1R01DA057591) to Dr. Coughlin and Dr. Lin.

All authors contributed meaningfully to the conceptualization, writing, and revision of the manuscript and have read and approved the final version.

We acknowledge and appreciate all participants who contributed to this research.

This manuscript has not been posted to a preprint archive prior to submission.

Funding:

Dr. Bonar (The Kathy Fant Brzoznowski Research Professor in Behavioral Health Technology Innovations) and Dr. Walton (The Toby Brzoznowski Research Professor in Behavioral Health Technology Innovations) are generously supported through the Broznowksi family. Dr. Tomlinson’s time was funded by National Institute on Alcohol Abuse and Alcoholism (NIAAA) Grant #T32AA007477 was awarded to DCT. Funding for this study was provided by NIDA through an NIH HEAL Initiative Award (1R01DA057591) to Dr. Coughlin and Dr. Lin.”

Footnotes

The authors declare no conflicts of interest.

Preliminary work for this paper was presented as a poster at an internal University of Michigan conference. Apart from this presentation, the data and ideas in this manuscript have not been disseminated, presented, or published elsewhere.

Works Cited

  1. Ahmed S, Sarfraz Z, & Sarfraz A (2022). Editorial: A changing epidemic and the rise of opioid-stimulant co-use. Frontiers in Psychiatry, 13, 918197. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Ahn W-Y, Vasilev G, Lee S-H, Busemeyer JR, Kruschke JK, Bechara A, & Vassileva J (2014). Decision-making in stimulant and opiate addicts in protracted abstinence: evidence from computational modeling with pure users. Frontiers in Psychology, 5, 849. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Banerjee Gakrieger MS, Goedel WC, Buxton JA, Lysyshyn M, Bernstein E, Sherman SG, Rich JD, Green TN, Mello MJ, Merchant RC, Mccormack RP, Baird J, & Green TC (2018). Effect peer-led behavioral intervention emergency department patients high risk fatal opioid overdose: randomized clinical trial. Harm Reduction Journal, 15(1). 10.1186/s12954-018-0213-2 [DOI] [Google Scholar]
  4. Beaudoin FL, Jacka BP, Li Y, Samuels EA, Hallowell BD, Peachey AM, Newman RA, Daly MM, Langdon KJ, & Marshall BDL (2022). Effect of a peer-led behavioral intervention for emergency department patients at high risk of fatal opioid overdose: A randomized clinical trial. JAMA Network Open, 5(8), e2225582. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Campbell M, Lafayette JI, Vallejo Vasquez K, Rivard Diamond, J., Nguyen C, Caputo D, Lothumalla S, Rodriguez CR, Tomlinson DC, Bonar EE, Goldstick JE, Walton M, Lee S, Menke N, Lin L, & Coughlin LN (2025). Preferences Predictors Driving Opioid-Involved Polysubstance Use Profiles Trajectories: Prospective Cohort Study. [DOI] [PMC free article] [PubMed]
  6. CDC. (2024, November 1). Understanding the opioid overdose epidemic. Overdose Prevention. https://www.cdc.gov/overdose-prevention/about/understanding-the-opioid-overdose-epidemic.html [Google Scholar]
  7. Degenhardt L, Clark B, Macpherson G, Leppan O, Nielsen S, Zahra E, Larance B, Kimber J, Martino-Burke D, Hickman M, & Farrell M (2024). Buprenorphine versus methadone for the treatment of opioid dependence: A systematic review and meta-analysis of randomised and observational studies. Drug and Alcohol Dependence, 260(110531), 110531. [DOI] [PubMed] [Google Scholar]
  8. Dolan SB, Johnson MW, Dunn KE, & Huhn AS (2021). The discounting of death: Probability discounting of heroin use by fatal overdose likelihood and drug purity. Experimental and Clinical Psychopharmacology, 29(3), 219–228. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Erickson CK (2018). The science of addiction (2nd ed.). WW Norton. [Google Scholar]
  10. Exum AC, Sutton CA, Bellitti JS, Yi R, & Fazzino TL (2023). Delay discounting and substance use treatment outcomes: A systematic review focused on treatment outcomes and discounting methodology. Journal of Substance Use and Addiction Treatment, 149(209037), 209037. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Johnson, & Bickel. (2008). An algorithm for identifying nonsystematic delay-discounting data. Experimental and Clinical Psychopharmacology, 16(3), 264–274. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Johnson J, Pizzicato L, Johnson C, & Viner K (2021). Increasing presence of xylazine in heroin and/or fentanyl deaths, Philadelphia, Pennsylvania, 2010–2019. Injury Prevention: Journal of the International Society for Child and Adolescent Injury Prevention, 27(4), 395–398. [DOI] [PubMed] [Google Scholar]
  13. Judd D, King CR, & Galke C (2023). The opioid epidemic: A review of the contributing factors, negative consequences, and best practices. Cureus, 15(7), e41621. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Khazanov GK, McKay JR, & Rawson R (2024). Should contingency management protocols and dissemination practices be modified to accommodate rising stimulant use and harm reduction frameworks? Addiction (Abingdon, England), 119(9), 1505–1514. [DOI] [PubMed] [Google Scholar]
  15. Kozak K, Lucatch AM, Lowe DJE, Balodis IM, MacKillop J, & George TP (2019). The neurobiology of impulsivity and substance use disorders: implications for treatment. Annals of the New York Academy of Sciences, 1451(1), 71–91. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. LaRue L, Twillman RK, Dawson E, Whitley P, Frasco MA, Huskey A, & Guevara MG (2019). Rate of fentanyl positivity among urine drug test results positive for cocaine or methamphetamine. JAMA Network Open, 2(4), e192851. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Mahu IT, Barrett SP, Conrod PJ, Bartel SJ, & Stewart SH (2021). Different drugs come with different motives: Examining motives for substance use among people who engage in polysubstance use undergoing methadone maintenance therapy (MMT). Drug and Alcohol Dependence, 229(Pt B), 109133. [DOI] [PubMed] [Google Scholar]
  18. Mattick RP, Breen C, Kimber J, & Davoli M (2014). Buprenorphine maintenance versus placebo or methadone maintenance for opioid dependence. Cochrane Database of Systematic Reviews, 2, CD002207. [DOI] [PubMed] [Google Scholar]
  19. Milaney K, Passi J, Zaretsky L, Liu T, O’Gorman CM, Hill L, & Dutton D (2021). Drug use, homelessness and health: responding to the opioid overdose crisis with housing and harm reduction services. Harm Reduction Journal, 18(1), 92. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Moallem NR, Courtney KE, & Ray LA (2018). The relationship between impulsivity and methamphetamine use severity in a community sample. Drug and Alcohol Dependence, 187, 1–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. National Institute on Drug Abuse. (2004). Drug abuse treatment outcome study--adolescent (DATOS-A), 1993–1995: [United States] [Dataset]. In ICPSR Data Holdings. Inter-university Consortium for Political and Social Research (ICPSR). 10.3886/icpsr03404.v3 [DOI] [Google Scholar]
  22. National Institute on Drug Abuse. (2021, June 1). Prescription opioids DrugFacts. National Institute on Drug Abuse. https://nida.nih.gov/publications/drugfacts/prescription-opioids [Google Scholar]
  23. National Institute on Drug Abuse. (2024, August 21). Drug overdose deaths: Facts and figures. National Institute on Drug Abuse. https://nida.nih.gov/research-topics/trends-statistics/overdose-death-rates [Google Scholar]
  24. National Institute on Drug Abuse [NIDA]. (2023, January 6). NIDA Quick Screen V1.0F 1. National Institute on Drug Abuse [NIDA]. https://nida.nih.gov/sites/default/files/pdf/nmassist.pdf [Google Scholar]
  25. O’Donnell J, Gladden RM, Mattson CL, Hunter CT, & Davis NL (2020). Vital signs: Characteristics of drug overdose deaths involving opioids and stimulants - 24 states and the District of Columbia, January-June 2019. MMWR. Morbidity and Mortality Weekly Report, 69(35), 1189–1197. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Phillips JK, Ford MA, & Bonnie RJ (2017). Evidence on strategies for addressing the opioid epidemic. National Academies Press. [Google Scholar]
  27. Rasmussen EB, Camp L, & Lawyer SR (2024). The use of nonmonetary outcomes in health-related delay discounting research: Review and recommendations. Perspectives on Behavior Science, 47(2), 523–558. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Rodriguez CR, Campbell M, Bonar EE, Goldstick JE, Walton MA, Lin LA, & Coughlin LN (2025). Improving data integrity in samples obtained from web-based recruitment: Protocol for the development of a novel system for assessing participant authenticity in a remote longitudinal cohort study of polysubstance use. JMIR Research Protocols, 14(1), e69956. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Smith GE, Ross RL, & Rost KM (1996). Psychiatric outcomes module: substance abuse outcomes module (SAOM). In Sederer LI & Dickey B (Eds.), Outcome Assessment Clinical Practice (pp. 85–88). [Google Scholar]
  30. Smith LR, Earnshaw VA, Copenhaver MM, & Cunningham CO (2016). Substance use stigma: Reliability and validity of a theory-based scale for substance-using populations. Drug and Alcohol Dependence, 162, 34–43. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. (Substance Abuse and Mental Health Services Administration [SAMHSA]. (2021, February 5). NIDA-modified ASSIST (NM ASSIST): Clinician’s screening tool for drug use in general medical settings. (Substance Abuse and Mental Health Services Administration [SAMHSA]. https://www.samhsa.gov/resource/ebp/nida-modified-assist-nm-assist-clinicians-screening-tool-drug-use-general-medical [Google Scholar]
  32. (Substance Abuse and Mental Health Services Administration (SAMHSA). (2022). 2022 National Survey on Drug Use and Health. (Substance Abuse and Mental Health Services Administration (SAMHSA). https://www.samhsa.gov/data/system/files/media-puf-file/NSDUH-2022-DS0001-info-codebook.pdf [Google Scholar]
  33. Substance Abuse and Mental Health Services Administration (SAMSA). (2021). 2020 National Survey on Drug Use and Health. Substance Abuse and Mental Health Services Administration (SAMSA). https://www.datafiles.samhsa.gov/sites/default/files/field-uploads-protected/studies/NSDUH-2020/NSDUH-2020-datasets/NSDUH-2020-DS0001/NSDUH-2020-DS0001-info/NSDUH-2020-DS0001-info-questionnaire-specs.pdfhttps://www.samhsa.gov/data/system/files/media-puf-file/NSDUH-2020-DS0001-info-codebook.pdf [Google Scholar]
  34. US Department of Health and Human Services. (2024). Overdose Prevention Strategy. US Department of Health and Human Services. https://www.hhs.gov/overdose-prevention/ [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplemental Material
NIHMS2148529-supplement-Supplemental_Material.docx (31.7KB, docx)

RESOURCES