Survey administration

This anonymous survey was created using and hosted by www.typeform.com. An outline of the survey is provided in Table 1. A link to the survey was posted along with a short advertising statement onto multiple discussion forums and social media sites which relate to psychedelics and psychedelic research. It was also distributed locally to the University of Utah Psychedelic Science Interest Group listserv. The advertisement requested participation from all English-speaking adults with experience using a classic psychedelic (DMT, LSD, psilocybin, or mescaline). Survey respondents provided informed consent by answering yes to the first question of the survey, indicating that they met inclusion criteria (18 or older of age; fluent in English reading/writing; and had not previously taken this survey) and desired to participate voluntarily in the survey. No financial compensation was promised or provided in exchange for participation in the survey. The Institutional Review Board of the University of Utah reviewed and approved this study. While the survey requested participants to categorize their naturalistic psychedelic use into distinct categories, achieving a perfect categorization was challenging due to the vast array of settings and support types that respondents likely encountered. The primary categories of use were presented as:

1. A.“A formal or traditional group setting: in a group of other participants with a therapist, leader, or guide (this does not include a setting where you were the only person using the psychedelic)”;
2. B.“A formal or traditional solo setting: where you were the only person using the psychedelic, and you had a therapist, leader, or guide”;
3. C.“A non-traditional and non-formal group setting: with an unguided group of others”; and
4. D.“A non-traditional and non-formal solo setting: you were primarily alone, without a guide”.

In the survey, no specific questions were asked regarding psychedelic dosing for the sessions in question.

Data analysis

Responses were manually reviewed to identify and remove entries unsuitable for data analysis. Of the 900 people who clicked on the link to access our survey, 592 individuals submitted the survey during the study period (June–September, 2020), and 14 responses were deemed unsuitable and were removed (because they either indicated that they did not meet inclusion criteria, or they did not complete the survey questions; see Fig. 1).

View Full Size

Consort diagram illustrating the survey sample broken down by the self-reported setting of their naturalistic psychedelic use

Citation: Journal of Psychedelic Studies 7, 2; 10.1556/2054.2023.00261

• Download Figure
• Download figure as PowerPoint slide

Background and demographic characteristics (age, race, gender, education, income, country of residence, formal use (e.g., with a therapist, guide, or shaman) vs informal use (i.e., without any of the aforementioned formal supports)) were compared between group and solo settings using the Mann-Whitney-U (MWU) test or Fisher Exact test, as appropriate (Fisher Exact test was used instead of Chi Square because in many categories frequencies were <5).

For direct comparisons between group use (e.g., use of psychedelics in a setting with more than one person present) and solo psychedelic use (e.g., use of psychedelics in a setting with no one else present), respondents were separated into subgroups: those who used psychedelics only in a group setting (n = 162); only in a solo setting (n = 59); or in both settings (n = 357). Those who experienced both settings were excluded from primary analysis, however descriptive statistics for these respondents are reported separately (Table 2). To account for gender as a potential confounder of perceived benefit in group vs solo settings, data were stratified according to male and female genders. Respondents who self-identified as other than male or female were excluded from statistical comparisons due to small sample size, further narrowing the final sample size for group (n = 158) and solo (n = 58) settings (Table 3). However, descriptive statistics for respondents of all genders are provided in Table 4. Binary outcomes (use for mental health purposes, overall benefit, physical injury, symptomatic improvement, increased or decreased medications, new or worsening symptoms, decreased hospitalizations) and continuous outcomes (overall number of uses, number of uses for mental health purposes) were then compared using a Cochrane-Mantel-Haenszel (CMH) test (stratified Chi Square) or van Elteren (VE) test (stratified MWU test) as appropriate. The Woolf test was used to demonstrate homogeneous odds ratios across between gender strata, an assumption of the CMH test. Of note, not all participants answered the question “How many times have you used classical psychedelics with the primary goal of treating a mental health issue or issues?” resulting in a lower sample size for this question (n = 81; Table 4). To account for multiple testing, a Benjamini-Yekutieli (BY) correction was applied to the resulting p-values for all comparisons made using the p.adjust function in R and an FDR threshold of 5%.

Of note, in addition to comparisons between group and solo use, this data set could also be used to draw comparisons between formal and informal use. While these comparisons are beyond the scope of this current publication, it may be a fruitful topic for future study.

For power analysis, we used simplified models: chi square (for Cochrane-Mantel-Haenszel tests) and Student's t-test (for van Elteren tests). We estimated that our study was 80% powered to detect a moderate effect size (Cohen's d = 0.43, w = 0.22; n = 216, alpha = 0.05, df = 3) (Table 3).

For correlation between number of uses and overall benefit, a total sample of 582 was used (comprised of all 578 eligible responses, in addition to 5 respondents previously excluded for not indicating in which setting they used psychedelics, minus 1 participant who did not answer the “overall more helpful than harmful” question; Fig. 1). This is because no comparison was being made between group and solo setting, so it was not necessary to subdivide the sample as above. After verifying its assumptions, logistic regression was used to analyze the relationship between total number of psychedelic uses and the outcome of “helpful” or “harmful”.

Respondent characteristics

Demographics are shown in Table 5. The respondents were majority white (83.5%), male (65.6%), from the United States (68.3%), and had completed at least some college (82.8%). Participants were also asked whether they used psychedelics in a formal setting or informal setting. Most common was only ever having used psychedelics in informal settings (69%), followed by having used in both informal and formal settings (28%), and lastly having used in only formal settings (4%).

Of 578 participants, 59 reported only ever having used psychedelics in a solo setting, 162 reported only ever having used psychedelics in a group setting, and 357 reported having used in both solo and group settings. When stratified by whether they had used psychedelics in a group or solo setting, the only significant difference was gender: 59% of group users identified as male, whereas 85% of solo users identified as male (p = 0.001).

Mental health intentions and outcomes associated with psychedelic use

Participant responses to categorical questions relating to mental health intentions and outcomes are shown in Table 3. Ninety (90) respondents reported having used psychedelics with the primary goal of treating a mental health issue. For this purpose, psychedelics were more likely to be used in a solo setting than in a group setting (COR 0.35 (0.19–0.67), adjusted p = 0.03). Psychedelic use in a solo setting was more likely to be associated with subjective symptom improvement than in a group setting (COR 0.22 (0.11–0.42), adjusted p = 0.00021). After accounting for multiple testing, there was no significant association between group versus solo use and report of decreased psychiatric medications (adjusted p = 0.095). A total of 5 individuals reported a decreased frequency of psychiatric hospitalizations after psychedelic use, however, there was no significant association between group vs solo use and reported psychiatric hospitalizations (COR 0.60 (0.090–4.07), adjusted p = 1.00).

Subjective reports of new psychiatric symptoms were rare (5 total individuals), and no significant association was found between group vs solo setting and reports of new psychiatric symptoms (COR 0.43 (0.056–3.28), adjusted p = 1.00). Similarly, reports of worsened psychiatric symptoms were rare (2 individuals) and there was no significant association between group vs. solo use and reports of worsening symptoms (COR 0.11 (0.0063–2.01), adjusted p = 1.00). Only one respondent reported that psychedelic use was harmful to their mental health, and one respondent reported they needed an increase in psychiatric medications after psychedelic use.

Table 6 shows data for questions with numerical (continuous) answers. When stratified by gender, respondents report a higher median number of uses for the purpose of mental health in a solo setting than in a group setting, and there was no significant difference in the median number of total psychedelic uses per respondent between group and solo settings (see Table 6 for details).

Logistic regression was attempted to analyze the relationship between total number of psychedelic uses and perceived overall benefit (“helpful” vs “harmful”). However, given that a negative outcome was so rare (n = 10 out of 582 respondents), meaningful analysis was not possible. For the correlation between number of uses and overall helpful outcome, the model produced an OR of 1.0042 (95% CI [0.99, 1.02]), which was not significant (p = 0.64).

Physical danger or harm associated with psychedelic use

Eight individuals (4 females, 4 males) reported physical danger or harm in association with psychedelic use in group settings (Table 3). There was no significant association between group vs. solo use and report of physical danger or harm (COR infinite; CI infinite; adjusted p = 1.00). Table 7 shows the responses that were excluded from the results and the rationale for excluding them.

Group vs. solo comparisons

So-called “set and setting”—i.e. the context and mindset that precede and surround a psychedelic experience—are widely regarded as variables which strongly influence the effect of psychedelics (Carhart-Harris, 2018; Gukasyan et al., 2021). This study gathered self-report survey responses on the naturalistic use of psychedelics in different types of settings (in a group vs solo; and with vs. without formal support) and explores the differences between psychedelic use in these different contexts, especially differences pertaining to subjective mental health outcomes.

Across all settings, respondents were more likely to find psychedelic use helpful versus harmful to their mental health. Subjective negative effects—including new/worsened psychiatric symptoms, increased need for psychiatric medications, overall sense the psychedelics were harmful to mental health, and physical danger/harm—were rare overall and were not significantly associated with group only versus solo only use.

Those with the specific intention of treating mental health issues using psychedelics were about half as likely to use them in a group setting than in a solo setting (COR 0.35 (0.19–0.67), adjusted p = 0.03), including when stratified by gender. Solo psychedelic use was significantly more likely to be associated with subjective mental health symptom improvement than group use. The vast majority of recent clinical research into psychedelics has utilized a format that would be described as solo and formal in this study (Reiff et al., 2020) and it is possible that this setting and the facilitated space for personal introspection is more therapeutically impactful. However, several other plausible explanations could explain this difference: there could be an expectation bias towards solo use being more healing; group use could bias towards lighter, more social experiences; and the intentions going into group use might less often be directed towards inner healing (our study compared all group and solo users, regardless of whether the primary intention was treating mental health). There is also significant precedent of group-based psychedelic experimentation in recreational settings which accounts for a significant proportion of naturalistic psychedelic use which is not explicitly therapeutically focused (supported by our own finding that use for mental health purposes is less likely in a group setting). Lastly, it is conceivable that poorer baseline mental health could influence a person's choice of group versus solo setting (and poorer baseline mental health may also influence symptomatic improvement). Our baseline demographics did not take into account baseline mental health.

Despite a difference in symptom improvement between group and solo settings, there was no significant association between setting and perception of overall benefit. The reasons for the discrepancy between symptomatic improvement and overall benefit are not entirely clear. Although improvement of mental health symptoms implies respondents had symptoms to begin with, it could be conceivable that respondents reported psychedelics as overall more helpful than harmful simply because symptoms did not worsen. However, there were two separate survey questions to address symptoms: “Mental health symptoms improved,” and “Mental health symptoms worsened.” These two questions were answered differently, suggesting that respondents understood a difference between lack of worsening symptoms and actual improvement in symptoms.

The decoupling of perceived helpfulness and reported symptomatic improvement may be related to several factors. Research has suggested that depressive symptomatic improvement is not actually synonymous with real world functioning and quality of life (Langlieb & Guico-Pabia, 2010): i.e. perceived quality of life and functionality can improve independently of depressive symptomatology. There may have been benefits along these lines felt in group settings that were not captured by the survey questions about reduced mental health symptoms. For example, surveyants may have answered “overall more helpful” based on a general increased sense of well-being, regardless of whether or not they had symptomatic improvement. It is also possible that the survey did not assess certain symptom or experiential domains that did change following psychedelic administration.

There were no significant associations between group vs. solo use and report of physical danger or harm, change in psychiatric medications, new or worsening symptoms, and change in frequency of psychiatric hospitalizations. This supports the reasonableness of expanding clinical research to explore psychedelic-assisted therapy with group medicine administration, and clinical studies have begun both at our institution (“Study of Psilocybin Enhanced Group Psychotherapy in Patients With Cancer (HOPE)” 2020) and elsewhere (Anderson et al., 2020).

Psychedelics are less likely to be used for mental health purposes in a group setting than in a solo setting (COR 0.35 (0.19–0.67), adjusted p = 0.03). Furthermore, there was a significant difference in median number of uses for a mental health purpose in a solo setting versus group setting (F and M group medians = 2 and 2 respectively; F and M solo medians = 4 and 4.5 respectively; adjusted p = 0.00084). Though the reason for this isn't certain, this is conceivably due to the relative ease of informal group assembly (festivals, concerts) compared to intentional planned group experiences with therapeutic intent.

Our data show that female-identifying respondents are more likely to report group psychedelic use than solo use; and male-identifying respondents were more likely to report solo use. Accordingly, median number of total psychedelic uses among females is higher in group than solo settings (8 and 4.5 respectively). In non-psychedelic arenas, there is some limited evidence that women tend to seek out (non-psychedelic) group therapeutic settings more often and may even have better outcomes in those settings as compared to men (Ogrodniczuk, Piper, & Joyce, 2004). It is possible that variable preference could be confounding the results for group vs. solo naturalistic psychedelic use (especially where the gender discordance between groups was so prominent). Therefore, we stratified our data according to gender to account for this potential confounder.

Formal vs. informal settings

Among respondents using psychedelics to target mental health, a low percentage of both group and solo users reported utilizing a formal setting. This discordance seems likely related to barriers accessing formal psychedelic sessions (e.g., legality, cost, distance, privacy, or inability to locate sessions). For instance, in the U.S. currently, the only clearly legal way of accessing classic psychedelics for treating mental health is through research trials, which offer limited availability. It is also possible that some people are accessing psychedelics in the U.S. through legal exceptions for religious/spiritual use, and then using this access to aim for mental health improvement. Indeed, many formal group-based psychedelic sessions are organized in this country with predominantly spiritual intentions (e.g., Santo Daime or UDV ayahuasca ceremonies) which may conceive of mental health benefits as downstream from spiritual growth and de-emphasize the former (Marlan, 2019).

Number of uses and outcome

Our logistic regression model did not demonstrate a correlation between the number of psychedelic uses and likelihood of a helpful outcome. One reason for this may be the rarity of an overall subjectively harmful response (less than 2% of all respondents) likely impeded the possibility of detecting differences among differing number of uses of psychedelics. When a rare event is represented only a few times in the data (in our case n = 10), it is known that conventional logistic regression may grossly underestimate the likelihood of these events (Allison, 2012; King & Zeng, 2001). The rarity of subjectively overall harmful outcomes with psychedelic use is noted elsewhere in the literature, as with one study which found evidence of a “broad, robust and sustained positive impact on mental well-being in those that have a prior intention to use a psychedelic compound” (Mans et al., 2021).

One potential reason for such rarity of harmful outcomes (especially among those with a more total uses) may be that people would presumably not continue using psychedelics to treat their mental health unless they found it helpful.

Future studies would likely be more effective in modeling correlations by measuring response with a Likert or sliding scale. This would allow for increased sensitivity in measuring degree of benefit as well as enable other regression methods.

Group/solo only vs. mixed setting comparisons

Descriptive statistics for respondents with experience using psychedelics in both group and solo settings are provided in Table 2. Additionally, it is worth discussing in more depth the likelihood of an overall negative effect on mental health in this mixed setting group. Among the 221 respondents who used psychedelics only in either a solo or group settings, only 1 individual (solo settings only) reported an overall negative effect on mental health. But of the 357 respondents with experience in both solo and group settings, 10 reported an overall negative effect on mental health. Although this difference did not reach statistical significance (OR 6.33 (0.89–276.05), p = 0.059), it is feasible that experience in both solo and group contexts could represent less discriminating or intentional “set and setting,” possibly leading to more negative outcomes. Accordingly, regarding “setting,” only 1% of those who experienced both group and solo settings reported strictly formal use, verses 9% and 5% of group only and solo only respondents, respectively. Regarding “set” however, 48% reported prior use of psychedelics with intent to treat mental health, in between the 34% and 59% of group only and solo only respondents, respectively (Table 2). Whether set or setting could account more possible frequent overall negative outcomes remains unclear.

Limitations

Lack of experimental control is a significant limitation of this observational study. As we have limited information about the conditions under which respondents used psychedelics, we cannot account for possible environmental confounders.

Participants were recruited online, and this method of sampling tends to skew toward white individuals, consistent with our sample, which was 89% white (Dillman, Smyth, and Melani, 2014). There are also socio-political reasons that psychedelic use and psychedelic communities have historically been disproportionately white. For example, Jahn et al. posit that criminalization of substance use may act as a deterrent to psychedelic use especially among racial/ethnic minorities (Jahn, Lopez, de la Salle, Faber, & Williams, 2021).

Recruiting through online forums also creates the potential for selection bias, given that those with positive experiences using psychedelics may remain more interested in participating in online communities surrounding the topic. Similarly, group vs. solo settings select for different individual intentions and different environments that variably align with specific mental health therapeutic goals: this limits direct comparison from survey-response data.

No distinction was made as to which type of classic psychedelic was used in this survey (e.g., psilocybin, peyote, LSD, DMT, ayahuasca), nor what dosage. Therefore, we are unable to account for potential differences in the type of psychedelic used in group versus solo settings and any impact this may have had on the outcomes of said use. However it is worth noting that certain psychedelics—ayahuasca and peyote for instance—are more commonly used in more formal group settings. The survey did not specifically disambiguate psychedelic use between microdosing versus larger “macrodose” or experience-altering doses of psychedelics, meaning responses should be interpreted as possibly including microdose experiences.

Results in this survey were self-reported and subject to recency, social desirability, confirmation, and recall biases (Althubaiti, 2016).

Due to a low sample size (n = 5) of respondents who self-identified as other than male or female, these respondents are reported in survey question tables but were not included in statistical comparisons.

The construct of race remains challenging to analyze, and in this survey individual respondents were asked to select all races with which they identified. This led to many racial combinations, with 24 distinct categories for comparison. Comparing so many categories of race, each with such a small sample size, was deemed impractical. Therefore, the percentage of all respondents identifying with a given race is reported, resulting in overlap among participants between racial categories (Table 5).

In addition, ethnicity (e.g., Hispanic vs. Non-Hispanic) was conflated with race as asked in the survey question (ie, “How would you describe your race, ethnicity, or origin?”). This could create a scenario in which individuals of Hispanic ethnicity but not of Latinx race selected the Latinx racial category, thus creating more overlap. However the significance of this is uncertain, and the general public may not have a consistent understanding of the difference between ethnicity and race.

We acknowledge these reasons for potential overlap limit the statistical power for comparing race between group and solo settings, but results for the top six most reported races are nevertheless reported. In the future we may use the racial/ethnic categories used by the NIH, which preclude overlapping categories, making analysis more feasible. Notably however, even in studies where participants are required to select only one race, participants may in fact identify with other groups as well, resulting in overlap between groups not accounted for in statistical analysis.

Some of the reduction or improvement questions (e.g., “Mental health symptoms improved,” “Psychiatric medication was decreased/stopped,” “Overall frequency of psychiatric hospitalization decreased”) assumed the presence of a treatment or symptom where there perhaps was none (Table 1). For example, “Psychiatric medication was decreased/stopped” assumes they were on any medication prior to use. Respondents not taking any medication would be incapable of reducing it, so including these respondents underestimates a true reduction effect.

The question, “Have you ever used a classical psychedelic with the primary goal of treating a mental health issue?” does not indicate the proportion of prior experiences aimed at improving mental health. For example, no distinction would be made between an individual who only used psychedelics for mental health on only 1 occasion versus someone who used psychedelics for mental health in every instance. This limits the ability of our study to draw conclusions about intentions going into psychedelic use and the outcomes of that use.

Limiting respondents' ability to assess overall helpfulness of psychedelics to a binary outcome (“helpful” or “harmful”) resulted in an extremely small set of contrary responses (1.7%). More nuanced data such as helpfulness rated on a Likert scale may have uncovered interesting correlations to continuous measurements (e.g., total number of uses, number of “burns,” etc.) (Allison, 2012).