Introduction
Psychedelics are receiving widespread attention as clinical trials show positive results with psychedelic-therapy for a wide range of psychiatric disorders – for recent meta-analysis see Zeifman et al. (Reference Zeifman, Yu, Singhal, Wang, Nayak and Weissman2022). Legal access to psychedelic-therapy is available in Australia and the US state of Oregon, and FDA approval for the first psychedelic intervention is also under review. In this context of liberalizing access, it is imperative that we develop our ability to predict the nature of psychedelic experiences – so as to mitigate risk and potentially maximize benefit.
Qualities of the subjective experience have been found to be predictive of subsequent positive and negative therapeutic outcomes (Herrmann et al., Reference Herrmann, Earleywine, De Leo, Slabaugh, Kenny and Rush2022; Roseman, Nutt, & Carhart-Harris, Reference Roseman, Nutt and Carhart-Harris2018), indicating the nature of experience may be an important determinant of subsequent therapeutic trajectories (Yaden & Griffiths, Reference Yaden and Griffiths2020). Specifically, so-called ‘peak’ or ‘mystical-type experiences’ (Barrett, Johnson, & Griffiths, Reference Barrett, Johnson and Griffiths2015; Bogenschutz et al., Reference Bogenschutz, Forcehimes, Pommy, Wilcox, Barbosa and Strassman2015; Garcia-Romeu, Griffiths, & Johnson, Reference Garcia-Romeu, Griffiths and Johnson2015; Johnson, Garcia-Romeu, Cosimano, & Griffiths, Reference Johnson, Garcia-Romeu, Cosimano and Griffiths2014; Roseman et al., Reference Roseman, Nutt and Carhart-Harris2018; Studerus, Gamma, & Vollenweider, Reference Studerus, Gamma and Vollenweider2010), ‘challenging experiences’ (Barrett, Bradstreet, Leoutsakos, Johnson, & Griffiths, Reference Barrett, Bradstreet, Leoutsakos, Johnson and Griffiths2016; Gashi, Sandberg, & Pedersen, Reference Gashi, Sandberg and Pedersen2021; Roseman et al., Reference Roseman, Nutt and Carhart-Harris2018), and ‘emotional breakthrough’ (Dougherty et al., Reference Dougherty, Clarke, Atli, Kuc, Schlosser, Dunlop and Ryslik2023; Haijen et al., Reference Haijen, Kaelen, Roseman, Timmermann, Kettner, Russ and Carhart-Harris2018; Kettner et al., Reference Kettner, Rosas, Timmermann, Kärtner, Carhart-Harris and Roseman2021; Lyons et al., in prep; Murphy et al., Reference Murphy, Kettner, Zeifman, Giribaldi, Kartner, Martell and Carhart-Harris2021; Peill et al., Reference Peill, Trinci, Kettner, Mertens, Roseman, Timmermann and Carhart-Harris2022; Roseman et al., Reference Roseman, Haijen, Idialu-Ikato, Kaelen, Watts and Carhart-Harris2019) have been related to therapeutic outcomes (Peill et al., Reference Peill, Trinci, Kettner, Mertens, Roseman, Timmermann and Carhart-Harris2022).
While recognized for its importance, the nature of the acute subjective experience induced by psychedelics has been notoriously difficult to predict (Aday, Davis, Mitzkovitz, Bloesch, & Davoli, Reference Aday, Davis, Mitzkovitz, Bloesch and Davoli2021). Researchers have long speculated that ‘set’ and ‘setting’ (Leary et al., Reference Leary, Metzner, Presnell, Weil, Schwitzgebel and Kinne1965), and additionally the psychosocial ‘matrix’ (Eisner, Reference Eisner1997) as important modulators of psychedelic response, and the field has generally settled on certain guidelines that are intended to support safe and positive experiences (Johnson, Richards, & Griffiths, Reference Johnson, Richards and Griffiths2008). However, recent studies are beginning to advance on these to propose and test the contributions of specific elements to the nature and therapeutic success of psychedelic experiences (e.g. Heinzerling et al., Reference Heinzerling, Sergi, Linton, Rich, Youssef, Bentancourt and Kelly2023).
A range of predictors are being investigated, including environment (Kaelen et al., Reference Kaelen, Giribaldi, Raine, Evans, Timmerman, Rodriguez and Carhart-Harris2018; Kettner et al., Reference Kettner, Rosas, Timmermann, Kärtner, Carhart-Harris and Roseman2021), dosage (Barsuglia et al., Reference Barsuglia, Davis, Palmer, Lancelotta, Windham-Herman, Peterson and Griffiths2018; Bremler, Katati, Shergill, Erritzoe, & Carhart-Harris, Reference Bremler, Katati, Shergill, Erritzoe and Carhart-Harris2023; Griffiths et al., Reference Griffiths, Johnson, Richards, Richards, McCann and Jesse2011; Holze et al., Reference Holze, Vizeli, Ley, Müller, Dolder, Stocker and Liechti2021; Kangaslampi, Hausen, & Rauteenmaa, Reference Kangaslampi, Hausen and Rauteenmaa2020; Lyvers & Meester, Reference Lyvers and Meester2012; Madsen et al., Reference Madsen, Fisher, Burmester, Dyssegaard, Stenbæk, Kristiansen and Knudsen2019), traits (Russ, Carhart-Harris, Maruyama, & Elliott, Reference Russ, Carhart-Harris, Maruyama and Elliott2019a; Russ, Carhart-Harris, Maruyama, & Elliott, Reference Russ, Carhart-Harris, Maruyama and Elliott2019b; Smigielski, Scheidegger, Kometer, & Vollenweider, Reference Smigielski, Scheidegger, Kometer and Vollenweider2019; Studerus, Gamma, Kometer, & Vollenweider, Reference Studerus, Gamma, Kometer and Vollenweider2012), and ‘pre-state’ (see Fig. 1 for visual representation). We operationally define ‘pre-state’ as including the following elements: (1) expectations, (2) intentions, (3) present mood and mental state, and (4) present inter-personal feelings toward those in their presence (Carhart-Harris et al., Reference Carhart-Harris, Roseman, Haijen, Erritzoe, Watts, Branchi and Kaelen2018). With ‘pre-state’, we sought to capture aspects of ‘set’ (and aspects of ‘setting’ as they translate into ‘set’) that could be assessed shortly before taking a psychedelic.
Predictors of psychedelic outcomes (from Carhart-Harris et al., Reference Carhart-Harris, Roseman, Haijen, Erritzoe, Watts, Branchi and Kaelen2018) in which our operational definition bears most relevance to the ‘pre-state’ component.
Psychological preparation is hypothesized to be important for response to psychedelics – as implied here (Haijen et al., Reference Haijen, Kaelen, Roseman, Timmermann, Kettner, Russ and Carhart-Harris2018), but approaches toward promoting this vary greatly across clinical trials (Thal et al., Reference Thal, Wieberneit, Sharbanee, Skeffington, Baker, Bruno and Bright2022). A scale that assesses state-level variables that are relevant to ‘pre-state’, such as ‘preparedness’, could aid in the development of preparatory interventions, bolstered by an evidence base that is presently lacking. To our knowledge, there have been four modern research studies assessing our operationally defined ‘pre-state’ (Haijen et al., Reference Haijen, Kaelen, Roseman, Timmermann, Kettner, Russ and Carhart-Harris2018; McAlpine, Blackburne, & Kamboj, Reference McAlpine, Blackburne and Kamboj2024; Russ et al., Reference Russ, Carhart-Harris, Maruyama and Elliott2019a, Reference Russ, Carhart-Harris, Maruyama and Elliott2019b).
Russ et al. (Reference Russ, Carhart-Harris, Maruyama and Elliott2019a, 2019Reference Russ, Carhart-Harris, Maruyama and Elliottb) developed a two-factor inventory intended to assess the constructs of ‘surrender’ and ‘preoccupation’, finding that preoccupation was correlated with adverse reactions, while surrender was correlated with ‘ego dissolution’ and mystical-type experiences. McAlpine et al. (Reference McAlpine, Blackburne and Kamboj2024) developed a 20-item scale of preparedness. When prospectively testing their proposed scale on a psilocybin retreat dataset (N = 46), they found that a group with high preparedness scores showed significantly better responses to ‘Depression Anxiety and Stress Scale’ (DASS-21) scored depression, anxiety, and stress. Haijen et al. (Reference Haijen, Kaelen, Roseman, Timmermann, Kettner, Russ and Carhart-Harris2018) conducted a rigorous prospective survey design with questionnaires administered 1 week prior, 1 day before, 1 day after, 2 weeks after, and 4 weeks following the day a participant planned to use a psychedelic. Using a self-constructed scale aimed at ‘pre-state’, this study found that ‘pre-state’ was positively correlated with mystical-type experiences and well-being improvements and negatively correlated with acute challenging experience.
The present study intends to refine and formally validate Haijen et al.'s scale, calling it the ‘Imperial Psychedelic Predictor Scale’ (IPPS). Specifically, we used extended data from Haijen's survey (N = 741) (Haijen et al., Reference Haijen, Kaelen, Roseman, Timmermann, Kettner, Russ and Carhart-Harris2018), a study of ceremony usage of psychedelics (N = 836) (Kettner et al., Reference Kettner, Rosas, Timmermann, Kärtner, Carhart-Harris and Roseman2021), and two controlled trials (N = 30 and N = 28) (Carhart-Harris et al., Reference Carhart-Harris, Giribaldi, Watts, Baker-Jones, Murphy-Beiner, Murphy and Nutt2021; Lyons et al., in prep) to assess the internal, convergent, and predictive validity of the IPPS. In the process of validation, the scale was reduced to items that reliably predict acute experience across independent datasets.
Method
Data collection
The data used in this study were collated from four studies approved by the Joint Research Compliance Office and the Imperial College Research Ethics Committee. Written consent was obtained prior to admission to the studies. Due to the observational nature of two source studies, all psychedelic drugs were taken by the individuals own accord, without any experimental control. Data from controlled research received favorable opinion from National Research Ethics Service London-West London, was sponsored and approved by Imperial College London's Joint Research and Complication Organisation, adopted by the National Institute of Health Research Clinical Research Network and reviewed and approved by the Medicines and Healthcare products Regulatory Agency. Storage and dispensing licenses for schedule 1 drugs were obtained to enable safe formulation by Guys and St Thomas’ Hospital Pharmacy. Psilocybin for the controlled trials was provided by COMPASS Pathways.
Psychological measures
Measures relevant to the present study are outlined below.
EBI
The ‘Emotional Breakthrough Inventory’ (EBI) is used to assess emotional breakthrough via a visual analogue scale from 0 to 100. Internal consistency of EBI is excellent (Cronbach's α = 0.932). Predictive validity of EBI was first demonstrated through correlation with well-being 2-weeks post psychedelic experience (r = 0.294, p = 0.005; Roseman et al., Reference Roseman, Haijen, Idialu-Ikato, Kaelen, Watts and Carhart-Harris2019) and has since been further supported (Kettner et al., Reference Kettner, Rosas, Timmermann, Kärtner, Carhart-Harris and Roseman2021; Murphy et al., Reference Murphy, Kettner, Zeifman, Giribaldi, Kartner, Martell and Carhart-Harris2021; Peill et al., Reference Peill, Trinci, Kettner, Mertens, Roseman, Timmermann and Carhart-Harris2022).
MEQ
The ‘Mystical Experience Questionnaire’ (MEQ-30) (Maclean, Leoutsakos, Johnson, & Griffiths, Reference Maclean, Leoutsakos, Johnson and Griffiths2012), used here, is a revised and shorter version of an earlier measure developed by Pahnke and Richards (Reference Pahnke and Richards1966). MEQ has excellent internal consistency (Cronbach's α = 0.933) and strong predictive validity with respect to measures associated with long-term well-being (Garcia-Romeu et al., Reference Garcia-Romeu, Davis, Erowid, Erowid, Griffiths and Johnson2019; Griffiths, Richards, Johnson, McCann, & Jesse, Reference Griffiths, Richards, Johnson, McCann and Jesse2008).
CEQ
The ‘Challenging Experience Questionnaire’ (CEQ) pulls from other psychedelic questionnaires (Griffiths, Richards, McCann, & Jesse, Reference Griffiths, Richards, McCann and Jesse2006; Strassman, Qualls, Uhlenhuth, & Kellner, Reference Strassman, Qualls, Uhlenhuth and Kellner1994; Studerus et al., Reference Studerus, Gamma and Vollenweider2010) to assess distress during a psychedelic experience. Each of the 26 questions of the CEQ are measured on a 5-point Likert scale and relate to one of the seven dimensions: ‘fear’, ‘paranoia’, ‘insanity’, ‘physical distress’, ‘isolation’, ‘death’, and ‘grief’. Internal consistency of the subscales ranges from fair to very good (Cronbach's α = 0.65–0.89). Internal validity of the scale as whole has been reported as excellent (Cronbach's α = 0.95; Davis et al., Reference Davis, Barrett, So, Gukasyan, Swift and Griffiths2021).
Communitas
The Communitas Scale, introduced by Kettner et al. (Reference Kettner, Rosas, Timmermann, Kärtner, Carhart-Harris and Roseman2021), assesses perceived togetherness and shared humanity (communitas) – a social dimension of psychedelic experience – through eight questions. The Communitas Scale shows excellent internal consistency (α = 0.92). Concurrent validity is shown through correlation with psychological well-being (r = 0.22) and path analyses indicating a mediating effect on long-term outcomes.
Surrender
The surrender scale was introduced by Russ et al. (Reference Russ, Carhart-Harris, Maruyama and Elliott2019a) to assess the ‘set’ of a psychedelic user before dosing. The ten items that constitute the surrender factor showed high internal consistency (α = 0.918), and the four items that constitute preoccupation showed good internal consistency (α = 0.757) and were demonstrated in two different datasets to be correlated and predictive of psychedelic experience (Russ et al., Reference Russ, Carhart-Harris, Maruyama and Elliott2019a, Reference Russ, Carhart-Harris, Maruyama and Elliott2019b). This study used a reduced, five-item version of this scale administered before dosing sessions in Haijen et al. (Reference Haijen, Kaelen, Roseman, Timmermann, Kettner, Russ and Carhart-Harris2018).
IPPS construction
Twelve items were originally used in Haijen et al. (Reference Haijen, Kaelen, Roseman, Timmermann, Kettner, Russ and Carhart-Harris2018) and were employed here in the development phase of the IPPS. These were created by senior author, RC-H, based on prior knowledge of psychedelic phenomenology and literature, conversation with colleagues at the Centre for Psychedelic Research, Imperial College London, and direct work within the field of psychedelic science and medicine. The items were devised based on observations of responses to psychedelics seen first-hand (e.g. in research studies and trials), principles implied by quantitative research findings, and a familiarity with the human research literature on classic serotonergic psychedelics plus a broader reading of psychological literature. The idea was to capture aspects of an individual's ‘pre-state’ related to (1) their preparedness for the experience, (2) their readiness to relinquish top-down psychological control and let go to the experience, (3) their feelings toward others with them for the experience (e.g. their therapists, facilitators, or guides), and (4) their intentions for having the experience.
Drug type and dose
Two datasets used observational data with a range of drug type and dose. Participants had the following options to indicate which psychedelic they had taken. Dose varied from the equivalent of 50–300 μg of lysergic acid diethylamide (LSD). The controlled datasets are from two trials in which 25 mg of psilocybin was administered.
Software
A combination of R version 4.2.2 and MATLAB version R2021b was used to complete the following analyses. Additional packages used in R include: dyplr, stats, and psych. R was used to complete principal component analysis (PCA) and regression analyses. MATLAB was used to complete correlation matrices and regression analyses.
Factor analysis
IPPS scores from all 12 questions were included in an exploratory factor analysis by PCA. The internal consistency of resulting factors was assessed with Cronbach's alpha. PCA indicated three factors: set, rapport, and intention.
Scale reduction
Two 12 × 6 correlation matrices were run testing the original 12 items of the IPPS with 6 outcomes of interest, namely, MEQ, CEQ, and EBI. Items were retained if they showed significant correlations with outcomes in both of the independent datasets (see online Supplementary materials).
Validation
Predictive validity
To compare the predictive validity of the reduced IPPS across different datasets, individual factors and total IPPS scores were related to outcomes of interest in recreational use (cohort), ceremony use (ceremony), and two sets of controlled data (depression trial, healthy volunteer). The depression trial involved two dosing sessions, with the IPPS administered just before each ingestion of psilocybin. We therefore treated each dosing session separately (N = 30). The healthy volunteer study involved a single dosing session with IPPS again administered prospectively (N = 28).
Three multiple regressions were completed with acute measures (CEQ, MEQ, and EBI) as the dependent variable. Subfactors of IPPS were set as independent variables. Each model was trained on cohort data and tested on concatenated ceremony and depression trial datasets (healthy volunteer was not used in this analysis due to item differences explained below).
After this train–test regression analysis, simple regression fits were conducted on the two controlled datasets to further assess the predictive validity of the IPPS. An incomplete version of the IPPS was used in the healthy volunteer dataset due to its premature state at the start of the study. The version used maintained the rapport factor, but not intention and with a modified set. The modified set omitted items two and three of the final scale (Table 1).
Final IPPS
This table displays final 9-item IPPS with factor loadings. Bold numbers indicate which factors the items belong to. Items 1–6 fall under set, item 7 intention, and items 8 and 9 rapport.
a This item is not included in one-on-one dosing settings.
Convergent validity
In light of our aim to predict positive outcomes based on set, the surrender scale was used to convergently validate the IPPS. This was done through correlations between total IPPS scores as well as IPPS factor scores.
Results
Demographics
Demographics for cohort (N = 741), ceremony (N = 836), depression trial (N = 30), and healthy volunteer (N = 28) datasets are displayed in Table 2.
Demographics
This table displays demographic data for each of the four datasets used in the study. Percentages within parentheses represent portion of total original dataset.
Factor structure
A PCA was conducted on cohort data to confirm the previous structure attained and screen it for reduction. A Kaiser–Meyer–Olkin (KMO) test of sphericity returned a value of 0.80 on the cleaned cohort data (N = 319), indicating the data were suitable for a factor analysis. Our analysis suggests the same essential structure found in Haijen et al. (Reference Haijen, Kaelen, Roseman, Timmermann, Kettner, Russ and Carhart-Harris2018), with an initial Cronbach's alpha of 0.72 for the unreduced, 12-item scale. Scale reduction by the correlation matrix technique described in the Methods resulted in removal of three items: ‘I am preoccupied with my work and/or life duties’, ‘I have strong expectations for the upcoming experience’, and ‘The environment/setting feels good for my upcoming experience’ (see online Supplementary materials for correlations matrices used in reduction). To assess the internal consistency of the revised 9-item scale, a PCA was conducted using ceremony data (N = 735, KMO = 0.85). Reduction of the scale improved Cronbach's alpha to 0.83. The results of the factor analysis yielded three sufficiently orthogonal factors, namely (1) set (Cronbach's alpha: 0.81) – on which items inquiring about a felt readiness for and openness to the experience loaded; (2) a factor we called rapport (Cronbach's alpha: 0.76) – on which items sampling feelings of interpersonal trust and environmental comfort loaded; and (3) an intention factor that included items sampling expectations and intentions (Table 1).
Convergent validity
Convergent validity was assessed through correlation of IPPS scores with the surrender scale (Russ et al., Reference Russ, Carhart-Harris, Maruyama and Elliott2019a; Reference Russ, Carhart-Harris, Maruyama and Elliott2019b). Total IPPS scores significantly correlated with surrender (r = 0.455, p < 0.0001), with set showing moderate correlative strength (r = 0.433, p < 0.0001).
Predictive validity
Used as a total score or factor scores, the IPPS values were significantly predictive of important outcomes in each of the naturalistic datasets (Table 3). In both, set negatively correlated with challenging experience (r = −0.303 and −0.251 for cohort and ceremony respectively) and in ceremony settings rapport positively correlated with EBI scores (r = 0.136).
Correlations between IPPS scores and relevant acute outcomes
This table displays correlations between total and factor scores on IPPS and relevant acute outcomes. Bold numbers indicate significant (p < 0.05) correlations.
Train–test regression
IPPS explained 6.76% of the variance in CEQ scores (adjusted R 2 = 0.063, F = 14.5, p < 0.0001), with set significantly predicting challenging experience (β = −0.062, p < 0.001). When predicting MEQ, IPPS explained 4.23% of the variance (adjusted R 2 = 0.0376, F = 9.05, p < 0.0001). In particular, set significantly predicted mystical-type experiences (β = 3.17, p < 0.001). Variance in EBI was explained to a lesser extent, accounting for 2.42% of the variance (adjusted R 2 = 0.02, F = 5.12, p = 0.0017) and was significantly predicted by rapport (β = −0.062, p = 0.005).
Controlled studies regression
To test whether the predictions discovered through multi-dataset correlation and regression analysis were present in controlled datasets, we used regression fitting on two controlled datasets to explore if CEQ, EBI, and MEQ could be predicted by the IPPS. A simple regression fit in depression trial showed rapport explained 8.65% of EBI variance (β = 3.48, p = 0.028). In healthy volunteers, a simple regression fit showed that set explained 16% of CEQ variance (β = −3.72, p = 0.043). Also in healthy volunteers, a multiple regression fit showed that set and rapport explained 40% of MEQ variance (p = 0.003) (set: β = 3.44, p = 0.088, rapport: β = 10.75, p = 0.074).
Retrospective potential
Administration of the IPPS before and after the psychedelic experience in cohort data provided an opportunity to investigate retrospective validity. We found that prospective IPPS scores significantly correlated with retrospective IPPS scores. This was true for total scores (r = 0.658, p < 0.0001), as well as individual factors set (r = 0.683, p < 0.0001), rapport (r = 0.462, p < 0.0001), and intention (r = 0.637, p < 0.0001).
Discussion
This study sought to examine and develop the construct validity of a new ‘pre-state’ subjective rating scale for predicting acute responses to psychedelic compounds. We call this scale the ‘Imperial Psychedelic Predictor Scale’, or IPPS. Through data reduction and assessments of internal, convergent, and predictive validity, a brief 9-item scale was derived that contains three orthogonal factors, namely, set, rapport, and intention that were differentially predictive of different aspects of acute psychedelic experience.
As hypothesized, a positive mindset, as indexed by set, encapsulating feelings of readiness, preparedness, and openness for the experience plus a sense of comfort, general good mood, and low anxiety was associated with higher scores on positive aspects of the acute experience, namely mystical experience and emotional breakthrough and lower scores on challenging experience. A generally consistent pattern was seen for rapport and intention.
The prediction of EBI by rapport as per previous work (Murphy et al., Reference Murphy, Kettner, Zeifman, Giribaldi, Kartner, Martell and Carhart-Harris2021) suggests that social relationships may play a key role in emotional breakthrough. Positive therapeutic relationships have been identified as an important component for effective therapy (Kamilar-Britt, Gordis, & Earleywine, Reference Kamilar-Britt, Gordis and Earleywine2023; Murphy et al., Reference Murphy, Kettner, Zeifman, Giribaldi, Kartner, Martell and Carhart-Harris2021), and ‘therapeutic alliance’ has been found to be a core common factor underlying therapeutic response to psychotherapy (Grencavage & Norcross, Reference Grencavage and Norcross1990; Tschacher, Haken, & Kyselo, Reference Tschacher, Haken and Kyselo2015). Finally, acute and end-of-retreat communitas – which was correlated with rapport in this study – has been related to subsequent improvements in mental health outcomes (Kettner et al., Reference Kettner, Rosas, Timmermann, Kärtner, Carhart-Harris and Roseman2021; Kettner et al., preprint; Watts et al., Reference Watts, Kettner, Geerts, Gandy, Kartner, Mertens and Roseman2022). These data suggest that rapport may be a valuable 2-item measure of ‘positive therapeutic relationship’ that is predictive of positive outcomes.
The strength of the IPPS’ predictive capacity in the train–test regression analysis may appear modest compared with other literature (Davis et al., Reference Davis, Barrett, So, Gukasyan, Swift and Griffiths2021; McAlpine et al., Reference McAlpine, Blackburne and Kamboj2024; Roseman et al., Reference Roseman, Haijen, Idialu-Ikato, Kaelen, Watts and Carhart-Harris2019; Russ et al., Reference Russ, Carhart-Harris, Maruyama and Elliott2019a, Reference Russ, Carhart-Harris, Maruyama and Elliott2019b). However, we only removed participants if key data were absent, while other studies have selected subsets (Davis et al., Reference Davis, Barrett, So, Gukasyan, Swift and Griffiths2021; Roseman et al., Reference Roseman, Haijen, Idialu-Ikato, Kaelen, Watts and Carhart-Harris2019) or performed other data cleaning (Russ et al., Reference Russ, Carhart-Harris, Maruyama and Elliott2019a, Reference Russ, Carhart-Harris, Maruyama and Elliott2019b). Studies have also utilized retrospective data for scale development (McAlpine et al., Reference McAlpine, Blackburne and Kamboj2024; Russ et al., Reference Russ, Carhart-Harris, Maruyama and Elliott2019a, Reference Russ, Carhart-Harris, Maruyama and Elliott2019b). A hindsight bias could easily strengthen apparent predictive relationships in these models. McAlpine et al. (Reference McAlpine, Blackburne and Kamboj2024) were able to demonstrate the scale can perform prospectively, showing prospective prediction of long-term outcomes within a dataset from a psilocybin retreat. However, our IPPS remains the only prospectively developed predictor of acute outcomes, showing statistical significance across four independent datasets. Additionally, the bulk of the participants in the present study were in naturalistic settings, in which dose and setting were not controlled. Considering the large effect that dose and setting (Aday et al., Reference Aday, Davis, Mitzkovitz, Bloesch and Davoli2021) can have on an experience, it is unsurprising that the effect of ‘pre-state’ variables is smaller in relative contribution. However, in scenarios that control for these components – and thus lessen the influence of uncontrolled confounds, we were able to detect stronger effects for the relationships between IPPS factors and acute outcomes, suggesting that the influence of ‘pre-state’ is stronger in controlled settings. Finally, brief scales have better response rates (Edwards et al., Reference Edwards, Roberts, Clarke, DiGuiseppi, Pratap, Wentz and Kwan2002), which is an extremely important consideration if the hope is that a scale be widely adopted and used e.g. as part of a registry of real-world data. This is especially important if the scale is to be used close in time to a ritualized procedure such as often occurs with psychedelic use or therapy, as it lessens the time commitment and burden psychological distraction.
Future work might seek to examine whether some personality traits interact with IPPS factors or converge with them to improve the ability to predict response. For example, prior work has identified trait ‘neuroticism’ as a predictor of challenging experience (Barrett, Johnson, & Griffiths, Reference Barrett, Johnson and Griffiths2017), and other studies have identified additional factors including high doses, young age, current life stress, polydrug use – including mixing psychedelics with high potency cannabis, adverse environmental conditions, and poor rapport or absent supervision (Bremler et al., Reference Bremler, Katati, Shergill, Erritzoe and Carhart-Harris2023; Carbonaro et al., Reference Carbonaro, Bradstreet, Barrett, MacLean, Jesse, Johnson and Griffiths2016; Simonsson, Hendricks, Chambers, Osika, & Goldberg, Reference Simonsson, Hendricks, Chambers, Osika and Goldberg2023), and even female gender (Bienemann, Ruschel, Campos, Negreiros, & Mograbi, Reference Bienemann, Ruschel, Campos, Negreiros and Mograbi2020), past diagnosis of personality disorder (Marrocu et al., preprint), and avoidant attachment style (Stauffer, Anderson, Ortigo, & Woolley, Reference Stauffer, Anderson, Ortigo and Woolley2021) as predictors of adverse responses to psychedelics. Our new tool may aid in safeguarding, highlighting if and when use might be contraindicated and should be delayed, such that e.g. set and rapport can be worked on and improved. However, it is important to recognize that the simple predictors identified in the present work could only explain a modest proportion of the variance in response, the majority remaining unexplained. This highlights the need for more work on prediction of response modeling in this space. It also implies that it may be premature to apply the IPPS as a (trustworthy) safe-guarding tool.
Limitations and future directions
This is a self-report study leaning primarily on correlation analyses. Previous efforts to predict long-term psychedelic outcomes have been mixed, so we focused on prediction of acute measures that could optimize safety and, after better models are developed, predict long-term therapeutically relevant outcomes as well. Future analyses such as structural equation modeling or path modeling may seek to identify causal structure in statistical relations to better predict key dependent variables such as mental health outcomes. Moreover, the influence of other core factors such as demographics, dose, setting, ‘matrix’ (Eisner, Reference Eisner1997), and ‘integration’ could be examined to see whether more powerful predictive models can be devised. In the meantime, the predictive power of the IPPS could be used to tailor treatment parameters, such as dosage or the amount or type of therapeutic support, with the knowledge that doing so can improve subsequent outcomes.
Another limitation pertains to how the scale was constructed i.e. based on a factor analysis on a relatively small pool of items generated by the senior author. More comprehensive approaches might include assessing online ‘trip’ reports, using interviews and surveys to inform item wording and to assess how items are understood across target populations, as well as the so-called Delphi technique, which involves utilizing the opinions of a carefully selected group of topic ‘experts’ (Sforzini et al., Reference Sforzini, Worrell, Kose, Anderson, Aouizerate, Arolt and Pariante2022). It is likely that we have overlooked salient factors in our choice of items for the IPPS. For example, none of our items explicitly measure expectations or expectancy. This construct could be treated either on a trait or state level. For example, in a recent double-blind randomized controlled trial of psilocybin therapy v. a conventional antidepressant for depression, expectancy was measured early in the trial as a single visual analogue scale item. Results showed that efficacy-related expectancy for psilocybin did not predict antidepressant response to psilocybin (Szigeti et al., Reference Szigeti, Weiss, Rosas, Erritzoe, Nutt and Carhart-Harris2024), yet the single-item approach worked very well for predicting response to a selective serotonin reuptake inhibitor – supporting its validity. Future research might consider assessing expectancy closer in time to drug in-take. In this respect, an expectancy item could be added to the IPPS.
Other limitations include the single item for the intention factor. This could be viewed as weak as there are no additional items with which to assess this factor's own internal validity. Accordingly, researchers may choose to omit this item to give an even more efficient scale (i.e. 8-items), or carry out further work to develop the construct validity of this specific factor. We also note that we have only one reverse scored item – i.e. ‘I feel anxious’ and thus, a similar critique could be made here.
Finally, we acknowledge that issuing ‘pre-state’ measures close to dosing could prime for certain responses and thus it is conceivable that this procedure could bias outcomes. One way this could be tackled in the future, would be to develop a scale with an equal number of items referring to positive and negative ‘pre-state’ conditions. In this regard, we acknowledge that e.g. with just one negative item in our present list i.e. ‘I feel anxious’, our scale could be accused of priming positive responses.
For scoring and use of the IPPS, as a first pass, we recommend using of set and rapport factors, independently, as predictors of acute or longer-term response (taking mean scores for each factor and excluding the single-item intention factor). Additionally, to be sensitive to negative priming, we recommend not including the single reverse score item of the ‘set’ sub-scale i.e. ‘I feel anxious’ prospectively. Thus, users of the IPPS should compute mean scores of the five positively scored items on the set factor and two items comprising the rapport factor, and test whether either (or both) is predictive of response.
Summary
In summary, the IPPS is an intentionally concise rating scale that can be used to efficiently predict the quality of an acute psychedelic experience. The full 9-item scale assesses set, rapport, and intention, and a briefer 7-item, 2-factor (set and rapport) use of the scale is advised. Here we showed that the IPPS was predictive of challenging and mystical-type experiences as well as emotional breakthrough, all known to be predictive of subsequent mental health changes post psychedelic use. Our scale contributes to a growing knowledge base around preparedness that could inform the development of preparatory interventions to improve outcomes. Ultimately, we hope our development and open provision of the IPPS will serve to improve our collective ability to predict – and eventually improve responses to psychedelics. We see special value here given the increasingly liberalized use of these compounds and increased prevalence of use.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S0033291724002204.
Acknowledgments
We would like to thank the online survey participants as well as the retreats that assisted with data collection. Special thanks go to Kenneth Jønck and Nicolai Lassen who provided the expertise in building the online platform Psychedelic Survey (www.psychedelicsurvey.com). We are currently running new online survey studies. For more information, see www.psychedelicsurvey.com. We are also grateful to staff who help run and support the two controlled trials that provided data to help validate the PPS. We are especially grateful to the founding funders of the Centre for Psychedelic Research at Imperial College London.
Funding statement
This research received no specific grant from any funding agency, commercial, or not-for-profit sectors. The depression trial from which data were used was supported by a private donation from the Alexander Mosley Charitable Trust and by the founding partners of Imperial College London's Centre for Psychedelic Research. The Insight study was also supported by the founding of the Centre for Psychedelic Research. R. C.-H. was supported by the Alexander Mosley Charitable Trust and later the founders of the Centre for Psychedelic Research and now the Ralph Metzner Endowed Chair at UCSF.
Competing interests
None.
Ethical standards
The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration of 1975, as revised in 2008.
Open access
