Effects of Naturalistic Psychedelic Use on Depression, Anxiety, and Well-Being: Associations With Patterns of Use, Reported Harms, and Transformative Mental States

Psychedelics have rapidly moved from marginalised substances to the focus of substantial scientific, commercial and public interest, driven in part by a series of small clinical studies reporting rapid and durable improvements across several psychiatric and addictive disorders when these agents are administered with psychotherapeutic support. Previous large-scale survey studies have similarly suggested associations between lifetime psychedelic use and increased emotional well-being, reduced problematic substance use, greater connection with nature, and shifts in political and social attitudes, and have linked these effects to transformative acute states such as mystical-type or insight experiences. Despite this growing literature, the authors identify several gaps: few large surveys have used validated measures to examine associations with core symptoms of major depressive disorder and generalized anxiety disorder; no prior study has examined how the number of lifetime psychedelic experiences relates to changes in depression and anxiety; and the frequency and consequences of harms ascribed to naturalistic psychedelic use remain under‑characterised. Martinotti and colleagues designed the Psychedelics and Wellness Study (PAWS) to address these gaps. The study aimed to test whether past psychedelic use is robustly associated with current levels of emotional wellness—operationalised as depressive and anxious symptom severity and overall well‑being—while also examining how frequency of use, preferred agent, occurrence of transformative long‑term states, and self‑reported harms relate to those outcomes. The investigators framed the study as hypothesis‑generating, recognising the limitations of cross‑sectional, retrospective self‑report data but emphasising the value of a large sample with diverse naturalistic exposure patterns.

The PAWS study used a cross‑sectional, anonymous online survey targeting adults aged 18 and older who reported at least one lifetime psychedelic experience. Recruitment occurred via free online platforms, social media targeted to psychedelic groups, flyers and snowball sampling; no incentives were provided. Inclusion criteria were age ≥18 and at least one self‑reported psychedelic use; no further screening was applied. Participants completed an online consent form before the survey. The survey asked respondents to retrospectively rate their average depressive, anxious and overall well‑being status twice: referring to their state prior to any lifetime psychedelic use and to their state after psychedelic exposure. Depression was assessed using the 9‑item Patient Health Questionnaire (PHQ‑9), anxiety with the 7‑item Generalized Anxiety Disorder scale (GAD‑7), and well‑being with the five‑item HERO Wellness Scale. In addition, the investigators developed two study‑specific instruments: the 26‑item Psychedelic Change Questionnaire (PCQ‑26) to capture longer‑lasting changes in emotions, cognitions and behaviours often associated with acute psychedelic experiences, and the 8‑item Negative Consequences Inventory (NCI‑8) to query eight specific harms participants might attribute to psychedelic use (including increases in suicidal desire, criminal impulses, aggressive behaviour and misuse of various substances). Participants also reported demographics, preferred psychedelic agent (single choice), estimated number of lifetime uses, and history of microdosing. The survey excluded MDMA as a classic psychedelic but included ketamine as a ‘‘psychedelic‑like’’ agent for completeness. Statistical analyses began with descriptive statistics and normality checks. Although pre/post difference scores deviated from normality on standard tests, bootstrap simulations (500 samples) supported the use of parametric tests. Paired‑samples t‑tests compared pre‑ and post‑scores on PHQ‑9, GAD‑7 and HERO; effect sizes were expressed as Cohen’s d. Exploratory factor analysis (EFA) with parallel Monte Carlo simulation guided factor extraction from the PCQ‑26; factor loadings were interpreted with standard thresholds and internal consistency reported with Cronbach’s alpha. To examine predictors of change, the researchers computed residualised change scores (post score minus predicted score from regression, standardised) and used multiple linear regression to estimate independent associations of demographic variables, PCQ‑26 factors and patterns of use with changes in PHQ‑9, GAD‑7 and HERO. Statistical significance was set at alpha < 0.05 (two‑tailed); analyses were conducted in SPSS v27. The institutional review board determined the protocol exempt given anonymous, minimal‑risk survey methodology and the study was registered on ClinicalTrials.gov.

A total of 2,510 adults completed the survey. Participants ranged from 18 to 86 years, with approximately equal representation of males and females; 58% held a bachelor’s degree or higher. The sample reported a mean of 38.55 lifetime psychedelic uses (range 1–500). Psilocybin (51.6%) and LSD (30.1%) were the agents most frequently nominated as ‘‘preferred’’; 90 respondents (3.6%) reported only a single lifetime use. Retrospective pre‑to‑post comparisons showed large, statistically significant reductions in depressive and anxious symptoms and increases in emotional well‑being. PHQ‑9 and GAD‑7 scores decreased significantly (PHQ‑9: t(2,509) = 51.54, p < 0.001; GAD‑7: t(2,509) = 52.79, p < 0.001), and HERO scores increased significantly (reported as significant in the text; exact t‑value for HERO was not clearly extracted). Effect sizes were described as large. Exploratory factor analysis of the PCQ‑26 identified three correlated factors that together accounted for 59.6% of variance: Factor 1 (13 items; Cronbach’s α = 0.94) comprised items resembling enduring mystical/transformative states; Factor 2 (6 items; α = 0.83) captured emotional and somatic domains (e.g. rumination, sleep, appetite, irritability); and Factor 3 (3 items; α = 0.82) reflected prosocial emotions and motivations. Factor correlations ranged from r = 0.44 to r = 0.65. Using an aggregated improvement metric, 91.7% of respondents reported improvements on Factor 1 items, 66.2% on Factor 2 items, and 77.8% on Factor 3 items. Regarding harms, 330 participants (13%) endorsed at least one negative outcome they attributed to psychedelic use, yielding 476 negative item endorsements in total. The most frequently reported problems included cigarette smoking and problematic cannabis use, with smaller numbers endorsing increases in suicidal ideation, aggressive/impulsive behaviour or misuse of benzodiazepines and opioids. Participants who reported any negative outcome showed significantly smaller pre‑to‑post improvements on PHQ‑9, GAD‑7 and HERO compared with those reporting no harms (statistical comparisons reported as significant; p < 0.001 for associations with reduced improvement). Patterns of lifetime use were associated with outcome magnitude. Extent of lifetime psychedelic use was not related to pre‑exposure symptom levels, but greater lifetime use correlated with larger reductions in PHQ‑9 (r = -0.06, p = 0.006) and GAD‑7 (r = -0.04, p = 0.040) scores and larger increases in HERO scores (r = 0.10, p < 0.001). Curve fitting for HERO scores suggested a sigmoidal relationship (r2 = 0.52, p = 0.018) in which benefit increased and then plateaued after roughly five exposures; no significant non‑linear curves were identified for PHQ‑9 or GAD‑7. Participants reporting only a single lifetime psychedelic use (n = 90) still showed significant pre‑to‑post improvements: PHQ‑9 t(89) = 6.57, p < 0.001, d = 0.68; GAD‑7 t(89) = 6.82, p < 0.001, d = 0.72; HERO t(89) = 6.99, p < 0.001, d = 0.74. Comparisons by preferred agent found no robust superiority of psilocybin or LSD over other agents on pre‑to‑post changes. Ayahuasca (n = 145) was associated with a small advantage in HERO‑assessed well‑being versus other agents (group × time interaction p = 0.010, d = 0.22), though no differences emerged for PHQ‑9 or GAD‑7. Participants who selected ketamine (n = 69) showed group × time interactions for PHQ‑9 (p = 0.035, d = 0.31) and GAD‑7 (p = 0.047, d = 0.24), but these effects were attributable to worse baseline scores in the ketamine group; PHQ‑9 and GAD‑7 remained higher post‑exposure in ketamine‑preferring participants (PHQ‑9 p = 0.006; GAD‑7 p = 0.035), and HERO scores did not differ (p = 0.255). In multiple regression models using residualised change scores, higher scores on PCQ‑26 Factors 1 and 2 independently predicted greater reductions in PHQ‑9 and GAD‑7 and larger increases in HERO (all p < 0.001). Endorsing any negative outcome was independently associated with smaller improvements in PHQ‑9 and GAD‑7 (p < 0.001). The text reports standardised beta coefficients were used to compare predictor strength, though specific beta values are not clearly extracted for all predictors.

Martinotti and colleagues interpret their findings as supporting a sizable association between naturalistic psychedelic use and improved self‑reported depression, anxiety and emotional well‑being, consistent with prior clinical trials and survey studies. The investigators emphasise that benefits were evident even after a single lifetime use, yet tended to increase with greater lifetime exposure before plateauing for well‑being after roughly five exposures and showing a non‑linear pattern for symptom measures that appeared to level off later. They highlight that persistent shifts in transformative‑type states (the PCQ‑26 Factor 1) and changes in emotional/somatic domains (Factor 2) correlated independently with improved mental health, suggesting that enduring changes in these states may be important mediators or markers of benefit. The authors acknowledge several important limitations. Chief among them are the retrospective, self‑report design and the inability to verify psychedelic use or timing, which prevent causal inference. Self‑selection and recruitment methods likely produced a non‑representative sample prone to response bias, and demand characteristics could have inflated positive reports; the authors note, however, that the presence of a measurable minority reporting harms provides some counterbalance to concern about uniformly biased positive responding. The survey did not collect context of use (for example, recreational versus therapeutic settings), dose, set and setting, or other details that can influence outcomes. Crucially, the NCI‑8 did not assess certain serious but less frequent adverse events such as psychotic reactions or Hallucinogen Persisting Perceptual Disorder (HPPD), nor did the survey capture serotonin syndrome, limiting the assessment of full risk profiles. In terms of implications, the study team suggests several directions for future research: prospective longitudinal surveys to verify associations between frequency of naturalistic use and wellness; more detailed characterisation of harms, their risk factors and their impact on long‑term functioning; and head‑to‑head comparisons of different psychedelic agents in both naturalistic and clinical settings to clarify relative benefits. The authors also raise the practical question—left unresolved by their design—of whether observed dose–response associations reflect repeated dosing strategies for acute enhancement, maintenance/redosing to prevent relapse, or both, and whether fostering persistence of transformative states might reduce the need for repeated dosing. They caution that their findings are hypothesis‑generating and should be followed by more rigorous investigations before drawing definitive conclusions about clinical utility or public‑health implications.