As fatal overdoses from synthetic opioids continue to rise, we need to understand decision-making processes underlying heroin and synthetic opioid use. This study evaluated the influence of sample impurity and fatal overdose risk on hypothetical heroin use. Individuals who currently use heroin (n = 69) were recruited online. Participants completed two probability-discounting tasks evaluating the likelihood of using a sample of heroin based on the likelihood of sample impurity and likelihood of fatal overdose, where greater discounting represented reduced use likelihood. Prior to completing the probability-discounting tasks, participants were randomized to read one of four prompts varying by the presence of information on heroin effects and active (e.g., fentanyl) or inert impurities. Influence of prompts on discounting processes and associations among probability-discounting measures, opioid use behaviors, and dependence severity were evaluated. Heroin use likelihood decreased with increased impurity or overdose risk and in a generally orderly fashion. Discounting was greater (i.e., reduced heroin use likelihood) when overdose risk, compared to sample impurity, was manipulated. Less discounting was associated with more severe opioid dependence. Discounting did not differ among prompts for either task. Individuals might adjust their heroin-use behavior to reduce harm with risk-related information. Greater discounting elicited by overdose relative to impurity risk suggests that equating adulteration and overdose risk is essential for harm reduction. Expanded access to drug checking services, which inform impurity and overdose risk, can reduce fatal overdoses. Due to fear of legal sanctions for these services, legislation and judicial decisions should explicitly protect these services. (PsycInfo Database Record (c) 2021 APA, all rights reserved).