This study aimed to investigate the impact of the pandemic on substance use patterns and psychological health of Brazilians. To address this, we assessed how the three years of pandemic could affect drug use habits in a Brazilian population and if changes could be associated with mood state symptoms and/or sociodemographic factors. Despite no changes were observed in the prevalence of use of all drugs, during the three time points investigated, ASSIST scores did not show the same pattern. Alcohol, cannabis, hallucinogens and cocaine/crack users reported a decrease in ASSIST scores in C2 (during a pandemic peak) compared to the C1 (during the beginning of pandemic), which was maintained during C3 (already decreed the end of national health emergency of the pandemic in Brazil), except for cocaine/crack. This might be explained by the emotional state of this sample, possibly related to fear of the pandemic. This aligns with the fact that C2 data were collected shortly after a significant surge COVID-19 cases and deaths, accompanied by stricter contingency measures, such as curfews, restrictions on non essential activities restrictions and circulation of people, among other measures implemented nationwide (23). In contrast, by the time of C3, approximately 84% of the Brazilian population had received at least one dose of the COVID-19 vaccine (C2: 23%, C1: 0%) (24), which may have alleviated pandemic-related fear and contributed to changes in the observed patterns. These findings suggest that pandemic policies, including vaccination and social distancing measures, may have directly influenced behavioral trends over time.

This could also be related to resilience to the pandemic after nearly two years. Studies (11, 25) have shown that participants who reported functional social support exhibited fewer symptoms of anxiety or depression and lower drug use (11, 25). A range of psychobiological mechanisms may underlie the relationship between resilience to stress and drug use. For instance, stress or adversity can alter certain functions of the mesolimbic pathway, increasing vulnerability to drug use (26–28). Social determinants, such as positive social interactions and social support, also play a critical role in fostering resilience. Positive social interactions can encourage adaptive coping strategies (29, 30), whereas insufficient or “negative” social support can contribute to maladaptive coping mechanisms, including risk-taking behaviors and drug use (26, 31).

Another factor potentially contributing to a sense of security during the pandemic was the implementation of vaccination programs (32, 33), which began in Brazil in 2021. With relatively low vaccine hesitancy among Brazilians (34), the reduction in COVID-19 cases and deaths led to the relaxation of contingency measures. Reduced isolation may have influenced the relationship between substance use and mental health, possibly mediated by vaccination rates. For instance, as observed in our previous study (13), higher levels of social distancing were associated with a lower percentage of drug use, suggesting that limited exposure to substances and heightened concern over the consequences of seeking or using drugs may have played a role.

However, different results are observed in other countries, like in Australia and the US (25, 35). The National Wastewater Drug Monitoring Program of Australia found that alcohol, cocaine and heroin levels in water had a decrease between April and August of 2022, while cannabis, oxycodone, fentanyl and MDMA levels showed an increase (25). The 2022 National Survey on Drug Use and Health reported a very slight increase from 2021 to 2022 of the following drugs: alcohol, cannabis, cocaine, meth, tranquilizers or sedatives and opioids. Tobacco showed a decrease, and hallucinogens did not change (35). In the present study we did not find the same chronological pattern compared to those ones, but the mitigation and timelapse of this sanitary phenomenon was not completely comparable.

Conversely, ASSIST also presents an intervention by the end of the questionnaire, showing which risk level the respondent is included in, which might have introduced a bias regarding reduced drug use scores. Although the brief intervention was tested and validated using a randomized controlled trial in four different countries (Brazil, India, Australia and USA) (36), their results showed that even the controls, who did not receive an intervention, had reductions in scores. Thus, the plain application of a questionnaire could present an effect in drug use score, which could happen in other studies as well, considering that the exposure to a survey on drug use could bring awareness to drug use and make respondents rethink their habits.

It is relevant to note that our sample seems to differ somewhat from other studies performed at the national level in Brazil. The most recent studies that report heavy episodic drinking and tobacco use reported in the last month (meaning between September of 2021 and February of 2022, and between December of 2022 and April of 2023) show an increasing trend of heavy episodic drinking and smoking (37, 38). Among heavy episodic drinkers and smokers, men were more prevalent than women, and both had an increase from one year to the other. Education years were different for alcohol and tobacco users. Drinkers had a higher number of education years than smokers, which is opposite from our results (37, 38). In our sample, higher education is more prevalent in alcohol and tobacco users. This may be related to the emotional state of young adults who are pursuing undergraduate degrees, as they have high levels of stress and depression symptoms, loneliness and worry feelings, as we can observe in studies with this population (10). Another aspect is that marital status has shown a relationship with the ASSIST scores for some drugs. A Japanese study found results that come in line with this, showing single people were more likely to perceive alcohol consumption and unhealthy lifestyle (39).

A study in a Canadian sample of people who use drugs (data collected between the second half of 2020 and the beginning of 2021) found that around 39% of their sample reported a decrease in income since the onset of the pandemic, and they also reported an overall drug use increase for drinking, use of cigarettes, e-cigarettes and cannabis (40). This comes in agreement with our results regarding income. Another study (32) from the USA reported that loss of income was among reasons for a decrease in drug use, as well as other stressful factors, such as social isolation and staying home. It must be highlighted that in the present research - and also one those ones presented, it is not clear if the reduction in income could be the cause or consequence of drug use and/or with, for instance, the limited social interaction due to the pandemic.

The DASS-21 severity ratings demonstrate a reduction of scores from C1 to C3. One hypothesis that could explain this phenomenon is that the population became gradually more accustomed to the context of the pandemic, which somewhat subsided symptoms related to their emotional state. Another hypothesis could be due to the relaxation of contingency measures. The increase in prevalence of people reporting more normal ratings for the three subscales further supports these inferences. A few cohort studies showed that anxiety levels were associated with higher risk of drug misuse (41), while depressive symptoms and loneliness were associated with binge drinking and functional social support was inversely associated with recreational drug use (11). If we compare these findings to our sample, for which a decrease in ASSIST for some drugs and DASS-21 scores was observed through time, we might infer that their emotional state has influenced their use, although we cannot pinpoint the cause to more resilience or more social support.

If we compare these findings to our sample, where a decrease in ASSIST scores for some drugs and DASS-21 scores was observed over time, we might infer that their emotional state influenced their substance use. However, it is difficult to attribute this to increased resilience or enhanced social support without deeper insights. Incorporating qualitative data, such as interviews or open-ended surveys, could provide a richer understanding of individual coping mechanisms and the role of community support systems in shaping these trends. Such data could help clarify how participants navigated stressors and whether specific protective factors contributed to the observed changes.

In our study, depression, anxiety and stress were associated with younger age, single/divorced/widowed, lower income, and lower education, moreover, women and lower social distancing with anxiety and stress. For this population, education and income had an effect mainly on anxiety and stress scores. In general terms, the greater the education and income levels, the lower the stress, anxiety and depressive symptoms, indicating a protective role of higher degrees of education and income. The results are similar to a study by Calegaro’s et al. (42), where the authors report an association between symptoms of anxiety and low levels of education. A discrepancy was found in another study (43) that discussed the probable presence of health anxiety, as higher education levels were related to the psychological symptoms. Another similar result is the gender effect, in which Calegaro et al. (42) reported higher levels of all subscales whereas we report higher levels of anxiety and stress in women (42). On this note, the literature shows that women have some psychological and socioeconomic traits which make them more prone to internalizing mental health disorders (44). A study investigating the risk factors for depression and distress (45) found that being single or widowed and living alone were linked to both symptoms, which comes in line with our data.

An interesting approach is that the most stressful and anxiogenic period of the pandemic (between March and May of 2020) (46) was not included in our first time point, as data collection was performed in September-October of 2020. Nevertheless, it is also perceived in the literature that stressful or harmful events may lead to depression later on (22, 47). With the first report, we might have seen a subchronic effect of the worst period of the pandemic, with moderate levels of depressive symptoms, while stress and anxiety were already waning.

The Spearman’s correlation could be a useful tool to confirm the relationship of the ASSIST scores and DASS-21 scores as both were progressively reduced through time points, which suggests the amelioration of symptoms. Though coefficients were not very strong, alcohol, tobacco, hypnotics, and opioids showed a few stronger coefficients through C1 to C3, indicating these users might be more vulnerable to stressful events. A cross-sectional study that was also conducted in Brazil (42) used a different approach, in which its focus was directed to the DASS-21 scale and PTSD symptoms. They also assessed the results of the DASS-21 and the link with the presence of drug use, for which alcohol, tobacco, cocaine, ecstasy or LSD, benzodiazepines and opioids users prevalence was much lower than in our sample. They reported an association between depression and PTSD symptoms with alcohol consumption, use of tobacco, cannabis and benzodiazepines, and stress with alcohol consumption and use of benzodiazepines.

In view of some populations strata presenting some changes in different drug classes, such information is important so the governmental organs can plan in a more adequate manner to mitigate the impact of drug use, as it is advocated by the National Plan of Policies on Drugs, previewed in the Law n° 13,840/2019 (48). This National Plan recognizes the importance of preventing drug use and implementing measures and programs to reduce the hazardous consequences of drug use and abuse, as much as enhancing healthcare, social assistance and promoting better access to public services. Therefore, there might emerge a need to pursue other routes not explored by this study, e.g., when considering the complexities of polydrug use and novel psychoactive substances (NPS) use screening and therapies, which could have a range of implications and challenges (14), which are relevant in the current context. NPS detain a whole level of complications for its detection and consequences to the organism, in view of their rapid evolution, variability in composition, and limited availability of reliable detection methods, complicating both screening and law enforcement efforts (49, 50).

Additionally, they can be much stronger or unpredictable in its effects (51), and even more, be used in conjunction with other drugs, raising the fatality risk, such as the case of fentanyl analogues, which are able to provoke a strong respiratory depression and deaths (52). Polydrug use amplifies risks by increasing the likelihood of adverse drug interactions, complicating clinical management, and contributing to worse mental health outcomes (53). Many overdose deaths are caused by the adulteration of street drugs with other drugs or psychoactive substances, such is the case with cocaine or methamphetamine and lacing with substances like fentanyl, ketamine, levamisole, and others (54–56). Adulterants further exacerbate these risks, as their presence in illicit drugs can lead to severe health consequences and pose significant hurdles for public health initiatives. And not only that, therapy strategies can be strongly impaired if polydrug use is not even considered or known, as metabolites can remain on the body for some time and likely affect recovery/relapse (57).

Evidence-based guidelines addressing these issues, as seen in countries with advanced drug monitoring systems, highlight the need for more robust surveillance, public education campaigns, and targeted interventions (58–60). Strengthening the capacity to detect and respond to these emerging threats should be a priority within national strategies, including this National Plan, to ensure the development of effective and adaptive measures for drug use prevention and harm reduction.