Ayahuasca Treatment Outcome Project (ATOP): One-Year Results from Takiwasi Center and Implications for Psychedelic Science

Abstract

Objective:

This article focuses on the outcomes at 1 year post-treatment of a naturalistic evaluation of services provided through the Takiwasi Centre, an accredited Peruvian therapeutic community offering an ayahuasca-assisted, integrative treatment program for addiction rehabilitation.

Method:

Participants (n = 52) completed structured interviews and a battery of validated instruments. Outcome measures included the Addiction Severity Index (Version 5), the Beck Anxiety and Depression Inventories, the World Health Organization Quality of Life Brief Version (WHOQOL-BREF), and the World Health Organization Quality of Life Spiritual Religious and Personal Belief (WHOQOL-SRPB) scale. Likert rating scales were used to assess perceived importance and significance of different aspects of the program and overall participant satisfaction.

Results:

The group change from baseline to the 1-year follow-up was significant and in the anticipated direction for alcohol and drug use severity, depression and anxiety, and some dimensions of quality of life. There was considerable individual variation in outcomes and treatment duration. The majority of participants rated all aspects of the program as important, including the spiritual and therapeutic significance of the ayahuasca experience, which was rated as very significant.

Conclusions:

Based on the positive 1-year outcomes, and within the limitations of an uncontrolled observational study design, the findings suggest promise for the effectiveness of the use of ayahuasca in a multifactorial treatment context for individuals with significant treatment histories, high levels of comorbidity, and treatment motivation. Results highlight considerable variation in individual experience that merit in-depth qualitative analysis. Implications for ayahuasca-assisted and other psychedelic-assisted treatment alternatives are discussed.

Ayahuasca is the Quechua-language name for a psychoactive plant beverage traditionally used therapeutically in Amazonian regions of South America. The therapeutic use of ayahuasca is of growing international interest primarily for its potential in reducing symptoms of depression, anxiety, and substance use disorders, among other mental health conditions (see Maia et al., 2023, for a review of recent studies). Observational studies have suggested that ayahuasca may be effective as part of a treatment protocol for reducing problematic substance use (Argento et al., 2019; Berlowitz et al., 2019; Loizaga-Velder & Pazzi, 2014; Thomas et al., 2013), although studies to date have not assessed long-term effects on substance use and related outcomes. To address this gap, the Ayahuasca Treatment Outcome Project (ATOP) developed a mixed-methods protocol to measure outcomes in a long-term residential treatment program for substance use disorders (Rush et al., 2021). Data collection occurred at Takiwasi Centre, an addiction recovery clinic in Peru that combines the use of ayahuasca and other traditional Amazonian medicine practices with individual and group psychotherapy in a therapeutic community (TC) milieu. This article presents the quantitative outcomes at 1 year post-discharge and explores associations with different aspects of program participation, with implications for psychedelic-assisted therapies for substance use health.

The evidence for the safety and potential mechanisms by which ayahuasca may confer benefits has been recently reviewed by Rush et al. (2022) and Maia et al. (2023). Based on a wide range of different types of observational, survey, clinical, and preclinical studies, both reviews highlight therapeutic promise for substance use and other mental disorders, physical health conditions, and overall well-being among healthy community members. Both reviews also call for considerably more research on ayahuasca-assisted therapies. In two naturalistic studies of ayahuasca-assisted interventions, Berlowitz et al. (2019) and Thomas et al. (2013) reported reduced substance use and cravings compared with baseline, whereas surveys of people who drink ayahuasca suggest that it may be associated with reduced alcohol and tobacco use or lower severity of alcohol use (see Maia et al., 2023). The mechanisms of action for how ayahuasca may contribute to substance use health are not yet clear, although its potent psychoactive components include the classic psychedelic N,N-dimethyltryptamine (DMT) as well as harmala alkaloids, which in combination are known to confer short-term antidepressant and anxiolytic effects (Osório et al., 2015; Palhano-Fontes et al., 2019; Sanches et al., 2016). Completed randomized controlled trials that have tested lysergic acid diethylamide (LSD) and psilocybin, both classic psychedelics, for alcohol and tobacco use disorders, indicate that psychedelic-assisted therapies that include a strong focus on pre- and post-session psychotherapeutic interventions may contribute to reduced problematic substance use and dependence (Bogenschutz et al., 2022; Johnson et al., 2017; Krebs & Johansen, 2012). To further investigate the contextual use of ayahuasca in a therapeutic setting, the ATOP team collected both quantitative and qualitative data among participants at Takiwasi Centre between 2016 and 2020 concerning treatment experiences and outcomes related to substance use and mental health.

Research setting

The Takiwasi Centre is an all-male¹ addiction recovery, nonprofit institution in the Upper Peruvian Amazon. The treatment model is referred to as a “therapeutic tripod”: traditional Amazonian medicine, psychotherapy, and communal living. In many respects, Takiwasi includes the core features of a hybridized TC (Perfas, 2019), including long-term stays and the role of communal living in the therapeutic process. Although there is adherence to the “community as method” principle, Takiwasi has a more hierarchical organization than classic TCs, which are peer led; Takiwasi is led by trained therapists and coordinated by administrative staff (De Leon, 2019). The treatment stay at Takiwasi is completely voluntary, although participants are required to participate in all aspects of the program, with the exception of religious activities. The program requires a 9-month commitment with three main phases of the process; however, many patients come in with shorter-term commitments to accommodate their needs. The institution has a subsidy program for people with low income who cannot afford full treatment fees—50% of annual treatment costs are covered by the organization's subsidy program.

The treatment philosophy puts a strong emphasis on spirituality, particularly in the framework of traditional Amazonian medicine. Takiwasi does not identify as a religious institution; however, there is a notable Catholic influence maintained by the current leaders and staff, including an onsite chapel and a local Catholic priest (padre) who, until early 2019, performed Sunday masses, baptisms, and other Christian rites for those wishing to participate (Dubbini et al., 2020; Dupuis, 2018; O'Shaughnessy, 2017). Longer-term treatment goals are adapted to each individual, and this may or may not include complete abstinence as an indicator of recovery (Giove Nakazawa, 2002). The program stipulates that after at least 2 years post-discharge, and depending on the individual case, participants may return to selective low-risk alcohol use. However, consumption of other addictive substances including cannabis is considered inappropriate to recovery. For further background on Takiwasi Centre, see Berlowitz et al. (2017); O'Shaughnessy (2017); Politi et al. (2018, 2019); Mendive et al. (2023); and Rush et al. (2021). For a commentary on the activities of Takiwasi outside of the therapeutic community context, see Dupuis (2018).

Method

Study design

The ATOP study is designed with the recognition that the treatment protocol at Takiwasi is a complex intervention, and like all TCs, presents multiple potential experiences and processes that may influence therapeutic outcomes (Aslan & Yates, 2015). To help address this complexity, a mixed-methods approach was used to measure relevant quantitative outcomes over time, as well as qualitative reports on the significance of different treatment components and other topics such as motivation for seeking treatment. The full study protocol is reported in Rush et al. (2021) including all inclusion/exclusion criteria, measures and their timing, recruitment, data collection, and analysis procedures. The research protocol was approved by the Comité de Ética PRISMA, Lima, Perú. A prospective cohort design was used incorporating a set of quantitative measures and a structured qualitative interview. Data were collected at program intake, discharge, and 3-, 6-, 12-, 18-, and 24-month follow-ups post-discharge, although in this report we focus on the 12-month outcomes. The baseline assessment included the Mini-International Neuropsychiatric Interview (MINI) to define status regarding substance use and other mental disorders for purposes of study inclusion and sample description (Lecrubier et al., 1997; Sheehan et al., 1998).

Following completion of the MINI and other aspects of eligibility determination, the following measures were administered: Addiction Severity Index, 5th Version (ASI-5); Global Appraisal of Individual Needs (GAIN-I)–Substance Use Grid; Beck Depression and Anxiety Inventories (BDI and BAI); World Health Organization Quality of Life Brief Version (WHOQOL-BREF), and the World Health Organization Quality of Life Spiritual Religious and Personal Belief (WHOQOL-SRPB) scale; Treatment Entry Questionnaire (TEQ-12; at intake only); and Client Satisfaction Questionnaire (CSQ-8). See Supplemental Table 1 in Appendices for more details and references concerning the study measures. (Supplemental material appears as an online-only addendum to this article on the journal's website.) For most of the instruments, validated Spanish versions were available. The TEQ-12 English version was translated and back-translated according to accepted international guidelines. The Substance Use Grid of the GAIN-I, which captured substance use and treatment history, was also translated, and several terms adapted to reflect street names of various drugs in the Peruvian and Latin American context. The TEQ-12 is a 12-item baseline-only measure of participant motivation for treatment. Unlike other measures of treatment motivation, the TEQ draws on self-determination theory for understanding client experiences of social controls and the importance of self-determined motivation for long-term behavior change, resulting in three subscales: identified motivation (e.g., sees value in changing and personally chooses); introjected motivation (e.g., internalized feelings of guilt, shame, and anxiety); and external motivation (e.g., pressure/demands from social network or context) (Urbanoski & Wild, 2012; Wild et al., 2006). The follow-up assessment was initially performed through videocalls and then converted to an online platform with support provided as needed by a team member (SG). There may be slight variation in the precise timing of the scheduled follow-up and in the completion of all measures.

Study recruitment and measuring treatment participation

Between 2016 and 2020, 138 participants completed an intake assessment. A total of 47 left the program before participating in any ayahuasca ceremony as part of their treatment and thus were excluded from the study cohort. Of the 91 participants included, 81 whose discharge was within the cutoff date for the present analysis were evaluated. From the 81 evaluated at discharge, 69 were scheduled for 1-year follow-up interviews (12 were not yet at due date), of which 52 (75.3%) were completed. In all, 17 out of 69 had not yet responded to the request to schedule a 1-year follow-up. See Figure 1 for the study flow diagram.

Figure 1.

Study flow diagram

Treatment participation was measured in two ways: first, calculating the number of weeks of treatment for each patient (i.e., retention); second, the nature and extent of engagement in program activities, for example, the number of ayahuasca sessions, days of dieta (i.e., isolated plant medicine retreats), and integration sessions with staff. Takiwasi provided these data for ATOP from its internal information system (Saucedo Rojas et al., 2019).

Statistical analyses and significance testing were undertaken with IBM SPSS Statistics for Windows, Versions 25 and 27 (IBM Corp., Armonk, NY). Linear correlations were based on Pearson r and for interval-level data, Spearman's rho. Group means were compared with t tests or one-way analyses of variance; all used a level of statistical significance at p < .05 or lower.

Results

Table 1 shows the demographic and clinical profile of participants who reached the 1-year follow-up time point. Highlights include the wide age range and variation in employment status, nationality, and religion. The majority had previous treatment experience (77%) and had considerable mental health comorbidity at intake (75% with at least one mental disorder). There was also considerable diversity in the principal substance of concern, although many reported using multiple substances (data not shown).

Table 1.

Demographic and clinical characteristics of Takiwasi participants in 1-year study sample (n = 52)

Variable	n	%
Demographic characteristics
Age (M = 33.5 years)
19–29	20	38.5
30–39	18	34.6
40–59	14	26.9
Employment
Yes	20	38.5
No	27	51.9
Student	5	9.6
Nationality
Peru	21	40.4
Europe	15	28.8
Other South American	13	25.6
North America	3	5.8
Religion
Catholic	30	57.7
None	21	40.4
Other	1	1.9
Clinical profile
Treatment history
0	8	15.4
1–2	20	38.5
≥3	20	38.5
Unknown	4	7.7
No. of mental diagnoses
0	13	25.0
1	19	36.5
2	8	15.4
≥3	12	23.1
Principal substance
Alcohol	13	25.0
Opioids	13	25.0
Cocaine/crack/coca paste	12	23.1
Cannabis	10	19.2
Other	4	7.7

Note: No. = number.

In terms of employment status, region of birth, religious affiliation, level of mental health comorbidity, substances of concern, and baseline status of all measures of potential outcomes, the 1-year follow-up group (n = 52) did not differ significantly from three other subgroups at baseline: those who left the program before any ayahuasca (n = 47) (i.e., early dropouts), those still in treatment or not yet scheduled for their 1-year follow-up (n = 21), and those unable to be contacted for the scheduled follow-up interview at the time the present sample was drawn (n = 17). These similarities notwithstanding, the early dropout group was significantly lower at baseline than the other three subgroups in identified motivation (F = 6.407, df = 3, p < .001) (Table 2), and slightly younger (3–5 years younger on average) (F = 2.979, df = 3, p < .05) (data not shown).

Table 2.

Comparison of 1-year follow-up sample to other baseline subgroups on the Treatment Entry Questionnaire (TEQ)

Variable	Early dropoutb (n = 44) M (SD)	Completed 1-year follow-up (n = 52) M (SD)	Unable to contact for 1-year follow-up (n = 17) M (SD)	Not yet scheduled for 1-year follow-upc (n = 21) M (SD)	p
Identifieda	24.1 (4.77)	26.5 (2.87)	26.5 (2.58)	27.6 (0.98)	<.001
Introjected	9.1 (4.15)	10.1 (3.95)	10.2 (3.73)	11.0 (3.11)	n.s.
External	11.5 (8.04)	8.6 (5.62)	8.4 (5.98)	10.2 (5.82)	n.s.

Notes: n.s. = Not significant.

^aThe Identified domain score does not include responses to Q4 of the TEQ because of a data collection error;

^bthree participants in the early dropout group did not complete the TEQ in its entirety;

^cthis category includes the small number of clients still in treatment (n = 9) and clients not yet scheduled for their 1-year follow-up (n = 12). The one client who had been discharged and returned to treatment is also excluded here as he is now included among the 52 who completed the 1-year follow-up.

With respect to program retention, the mean length of stay was 27.9 weeks (SD = 12.9 weeks). Seven participants (13.5%) stayed less than 10 weeks; 9 (17.3%) stayed between 10 and 19 weeks; 8 (15.4%) stayed between 20 and 29 weeks; 21 (40.4%) stayed between 30 and 39 weeks; and 7 (13.5%) stayed 40 weeks or more. Given the highly variable and voluntary length of stay it is no surprise that there was significant variation in the frequency of participation in various components of the program (Table 3 and Supplemental Figures 1a to 1f).

Table 3.

Descriptive statistics on participation in the various components of the Takiwasi program

Variable	n	Min.	Max.	Mdn	M	SD
Ayahuasca sessions	52	2	29	14.5	13.8	7.4
Dietas	52	0	5	2	2.3	1.5
Purga sessions	52	7	36	21	20.6	7.3
Purgahuasca sessionsa	52	0	5	3	2.7	1.6
Intake of plantas de contenciónb	52	3	23	13	11.9	5.1
Therapeutic sessions	52	2	62	31	29.2	13.9

Notes: Min. = minimum; max. = maximum.

^aPurgahuasca sessions are a variation of ayahuasca sessions and therefore could be included as an ayahuasca session;

^bplantas de contención or “containment plants” are low-concentration decoctions of medicinal plants taken outside the context of the dieta.

Table 4 shows the significant improvements from baseline to follow-up at 1-year post-discharge for the various composite scores of the ASI and the BDI and BAI. With respect to the WHOQOL-BREF, significant improvements were noted for the measure's two single items assessing overall quality of life and satisfaction with health, as well as the psychological health subscale. No difference was found for the physical health, social relationships, and environment subscales of the WHOQOL-BREF, the latter containing items concerning financial resources, perceived safety and security, home environment, and participation in and opportunities for recreation and leisure. There was notable variation in size and direction of the change in the various subscales of the WHOQOL-SRPB (spirituality) measure.

Table 4.

Pre–post changes in the outcome domains of addiction severity, depression and anxiety, quality of life, and spirituality

Variable	M at baseline (SD)	M at 12 months (SD)	M change (SD)	df a	t (p)	Cohen's d
ASI composite scores
Medical	0.33 (0.29)	0.21 (0.32)	-0.12 (0.40)	51	2.11 (.04)	0.29
Employment	0.62 (0.33)	0.50 (0.29)	-0.12 (0.29)	51	2.99 (<.01)	0.41
Alcohol	0.28 (0.28)	0.08 (0.14)	-0.19 (0.26)	50	5.39 (<.01)	0.76
Drugs	0.24 (0.15)	0.05 (0.07)	-0.19 (0.16)	51	8.74 (<.01)	1.22
Legal	0.15 (0.19)	0.04 (0.15)	-0.10 (0.23)	50	3.38 (<.01)	0.45
Family	0.41 (0.22)	0.18 (0.20)	-0.23 (0.24)	51	6.82 (<.01)	0.95
Psychiatric	0.44 (0.22)	0.32 (0.22)	-0.12 (0.32)	51	2.68 (.01)	0.37
Beck Depression and Anxiety
Depression	17.7 (10.0)	11.1 (11.2)	-6.6 (12.5)	46	3.64 (<.01)	0.53
Anxiety	18.8 (12.0)	13.7 (11.6)	-5.0 (15.5)	46	2.23 (.03)	0.33
WHOQOL-BREF
Overall quality of life	2.6 (1.2)	3.2 (0.98)	0.60 (1.4)	46	2.96 (<.01)	0.43
Satisfaction with health	2.3 (1.1)	3.5 (1.1)	1.2 (1.5)	46	5.62 (<.01)	0.82
Physical	22.7 (5.4)	23.9 (4.7)	1.2 (6.4)	46	1.32 (.19)	0.19
Psychological	17.6 (4.4)	19.6 (4.4)	2.0 (4.9)	46	2.85 (<.01)	0.42
Social relationships	8.1 (2.5)	8.6 (2.5)	0.5 (3.1)	46	1.05 (.30)	0.15
Environment	26.7 (5.8)	26.0 (4.5)	-0.7 (6.0)	46	0.81 (.42)	0.12
WHOQOL-SRPB
Connection	13.1 (4.3)	13.8 (3.4)	0.7 (4.2)	46	1.07 (.29)	0.16
Meaning	15.8 (2.9)	14.4 (3.7)	-1.4 (3.7)	46	2.58 (.01)	0.38
Experience	15.5 (3.2)	14.7 (3.5)	-0.83 (3.2)	46	1.79 (.08)	0.26
Peak experience	12.1 (3.5)	13.2 (3.3)	1.1 (3.7)	46	1.99 (.05)	0.29
Strength	13.8 (3.0)	13.9 (3.3)	0.1 (3.3)	46	0.13 (.90)	0.02
Peace	10.3 (3.4)	12.2 (3.3)	1.9 (3.5)	46	3.77 (<.01)	0.55
Hope	14.6 (2.9)	13.5 (2.9)	-1.1 (3.4)	46	2.07 (.04)	0.31
Faith	12.8 (4.7)	13.6 (3.8)	0.7 (4.3)	46	1.17 (.25)	0.17

Notes: ASI = Addiction Severity Index; WHOQOL-BREF = World Health Organization Quality of Life Brief Version; WHOQOL-SRPB = World Health Organization Quality of Life Spiritual Religious and Personal Belief scale.

^aDegrees of freedom vary slightly because of missing data at follow-up for a small number of cases.

Figure 2a shows individual variation in 1-year post-discharge outcomes for the ASI Composite Alcohol Score according to the number of weeks in treatment. By plotting baseline scores along the x-axis and follow-up scores along the y-axis, cases that fall below the diagonal are those reporting a positive change, and those above the line a negative change. Interestingly, although many severe cases at baseline improved substantially, some cases with low severity at baseline had higher-severity scores at follow-up. In addition, a small number in the program for less than 10 weeks improved markedly (e.g., patients 35 and 64), whereas others with longer stays did not show any improvement (e.g., patient 40 after 20–29 weeks) or were marginally more severe (e.g., patient 54 after 40 weeks of treatment).

Figure 2a.

Scatterplot of Addiction Severity Index (ASI) Composite Alcohol Scores at baseline and follow-up by number of weeks in treatment

Figure 2b shows individual variation in 1-year post-discharge outcomes for the ASI Composite Drug Score and, as above, according to the number of weeks in treatment. Clearly there is more consistency in improvements on drug severity compared with alcohol severity. However, the individual variability of 1-year outcomes in relation to length of treatment is again quite evident. Interesting examples include patient 61, who scored more severe on drug use (but not alcohol use) after more than 40 weeks of treatment, compared with patients 4 or 29, who improved markedly after the same treatment duration; and patient 25, who showed significant improvement after 30–39 weeks of treatment; and patient 65, who also showed considerable improvement, but after only 10 weeks or less. Supplemental Figures 2a–2c demonstrate the variation in outcomes in the ASI Composite Psychiatric and Legal Scores and the BAI.

Figure 2b.

Scatterplot of Addiction Severity Index (ASI) Composite Drug Scores at baseline and follow-up by number of weeks in treatment

Correlations between length of stay in the program and change scores on all outcome measures (i.e., composite indices of the ASI, BDI, BAI, WHOQOL-BREF, and the WHOQOL-SRPB) failed to reach statistical significance (Supplemental Table 3a).

Participants were asked to rate on a four-point scale the importance of the different activities within the Takiwasi program. Correlational analyses found statistically significant associations between selected outcome measures and these participant ratings (Supplemental Table 3b). Notably, significant correlations were found between change on five subscales of the WHOQOL-SRPB (Connection, Meaning, Peak Experience, Strength, and Faith) and the importance ratings for several of the Takiwasi program activities, including those derived from traditional Amazonian practices.

Table 5 shows the descriptive statistics for the ratings of importance indicating that all program activities were viewed quite positively and with little variation across the categories. Supplemental Figures 3a–3e show the distribution of the importance ratings for each of the program activities. With respect to ayahuasca specifically, participants answered two questions on a five-point scale specific to their experiences, one related to spiritual significance and the other in relation to changes in personal well-being and life satisfaction. The majority of participants rated ayahuasca as a highly significant spiritual experience, with 20 participants (38%) rating it as the “the most significant spiritual experience in my life.” Similarly positive ratings were given for the significance of ayahuasca sessions for changes in well-being and satisfaction with life. A small number of participants felt unable to make the rating at the time of their interview (Supplemental Figures 4a–4b), and these cases were excluded from the mean scores shown in Table 5.

Table 5.

Participant ratings of the various activities within the Takiwasi program

Variable	n	Min.	Max.	Mdn	M	SD
Importance ratings for components of Takiwasi
Therapy/integration work	48	2.3	4.0	3.7	3.5	0.5
Community life	48	1.0	4.0	3.0	3.2	0.7
Ritual elements	48	2.5	4.0	4.0	3.6	0.5
Dieta/plants	48	2.0	4.0	4.0	3.7	0.5
Ayahuasca	48	2.0	4.0	4.0	3.8	0.4
Significance of ayahuasca
Spiritual significance	44	2	5	4.0	4.3	0.8
Significance for change in wellbeing and satisfaction with life	38	4	5	5.0	4.8	0.4

Notes: Min. = minimum; max. = maximum.

Table 6 shows the descriptive statistics across the eight items of the Client Satisfaction Questionnaire as well as the overall summary score. The majority of participants reported a positive experience with respect to the overall quality and helpfulness of the service (Q1). Ratings were lower for questions concerning satisfaction with the kind of service provided (Q2); whether they would recommend the program to others (Q4); satisfaction with the kind of help received (Q5); and whether they would return for more help (Q8).

Table 6.

Descriptive statistics for each item on the Client Satisfaction Questionnaire (CSQ) and the overall score

CSQa	n	Min.	Max.	Mdn	M	SD
Q1 How would you rate the quality of the service you have received?	47	2	4	3	3.4	0.6
Q2 Did you get the kind of service you wanted?	47	1	3	2	1.7	0.6
Q3 To what extent did the program meet your needs?	47	1	4	3	2.9	0.9
Q4 If a friend were in need of similar help, would you recommend the program to him or her?	47	1	4	1	1.4	0.7
Q5 How satisfied are you with the amount of help you have received?	47	1	3	1	1.5	0.6
Q6 Have the services you received helped you deal more effectively with your problems?	47	1	4	4	3.5	0.7
Q7 In an overall general sense, how satisfied are you with the services you have received?	47	1	4	4	3.5	0.7
Q8 If you were to seek help again, would you come back to the service?	47	1	4	2	1.6	0.7
CSQ total score	47	18.0	32.0	27.0	27.0	4.0

Notes: Min. = minimum; max. = maximum.

^aItems are scored on a Likert scale from 1 (low satisfaction) to 4 (high satisfaction), with different descriptors for each response point.

Discussion

The size and positive direction of improvements at 1-year follow-up for various dimensions of addiction severity, depression, anxiety, and quality of life suggest that the multifactorial Takiwasi treatment program is a useful model for delivering an ayahuasca-assisted intervention for people with substance use disorders. The ATOP findings presented here are consistent with results of a previous observational study of the ceremonial use of ayahuasca for addiction in a nonclinical context (Argento et al., 2019; Thomas et al., 2013). Although our quantitative outcome measures were not assessed during treatment, our results are consistent with other studies focused on Takiwasi clients during treatment or at the point of discharge (Berlowitz et al., 2019; Giovannetti et al., 2020; O'Shaughnessy, 2017). These ATOP outcome data contribute to a growing literature on the benefits of ayahuasca-related practices (Maia et al., 2023; Rush et al., 2022), but the lack of a comparison group and other limitations inherent to observational studies (e.g., self-selection bias; small sample size) preclude conclusions about causal associations and mediators and moderators. Another limitation is the lack of data on the small number of individuals who were lost to follow-up and the significant proportion of clients who left the program before any use of ayahuasca (47 of 138, or 34.1%) and were, therefore, not eligible for follow-up. Such high dropout rates are common in long-term TC programs (Malivert et al., 2012) and are influenced by many personal and institutional factors (Prangley et al., 2018). This potential for selection bias presents a challenge for any longitudinal study of substance use treatment. We found the substantive early dropout group to be somewhat younger and lower on self-identified motivation for treatment. These factors are of utmost relevance in interpreting our findings considering that baseline motivation is known to be a significant indicator for treatment retention and beneficial outcomes (e.g., Wild et al., 2016). This is also particularly salient in the TC context, where baseline motivation is predictive of treatment retention and time in program, the latter being the greatest predictor of beneficial outcomes at follow-up (De Leon, 2019; Debaere et al., 2014; Zhang et al., 2003). The TEQ scores suggest that the ATOP participant sample can be characterized as being highly motivated for this very particular model of substance use treatment. This prompts questions about who is best suited for this kind of multi-modal ayahuasca-assisted treatment and the relationship between treatment retention and outcomes among participants who may be more skeptical toward the use of Amazonian plant medicine in general, and ayahuasca in particular.

Notwithstanding the limitations to this observational outcome evaluation, we enumerate some important contributions to future research on ayahuasca and other psychedelic-assisted treatments for substance use and other mental disorders. First, the evident variability in the demographic characteristics, the degree and nature of psychiatric comorbidity, previous treatment histories, and motivation to change—as well individual outcomes achieved—point to the need for research focused on the most appropriate populations for whom these innovative yet complex interventions will be best suited. Sample sizes sufficient for multivariate analyses will be needed if the goal is to identify the strength of the relationship between outcomes and participant characteristics, length of stay, and participation in different treatment components while controlling for potential moderators and mediators. That being said, the magnitude and variance in the changes in several well-known outcome measures at 1 year after discharge will allow for more accurate sample size calculations for future longitudinal outcome research with these same measures. Our graphical results also illustrate how conclusions drawn from statistically significant group differences may mask considerable individual variation in outcomes, and we suggest a similar approach for future research on psychedelic-assisted treatment.

Second, questions remain regarding the sustainability of outcomes over time, which is a common concern in treatment outcome research for substance use health. Equally relevant is the outstanding question of what specific outcomes are of most interest and value for various stakeholders (e.g., abstinence or change in use behaviors, quality of life). To date, studies on psychedelic-assisted therapy for substance use disorders are generally characterized by few dosing sessions (typically 1–3) over a period that ranges from 1 to 12 weeks (e.g., Bogenschutz et al., 2015, 2022; Dakwar et al., 2020), and follow-up for less than a year, with few exceptions (e.g., Johnson et al., 2017; Noorani et al., 2018). Outcomes from these studies are promising yet prompt several questions concerning multiple dosing sessions over a long period and the role of “booster” sessions for maintaining well-being and/or motivation for recovery. In the case of Takiwasi Centre, it is notable that even with up to 23 ceremonies over a 9-month period with weekly psychotherapeutic support, there remain persistent issues related to post-treatment recovery goals such as challenges related to mental well-being, short-term lapse, or full-on relapse (Marcus et al., 2022). Conversely, we also found quite substantive improvements among some individuals with very short program experience, which accords with some clinical trial data that demonstrate rapid-acting and sometimes lasting benefits with just one to three psychedelic sessions (Bogenschutz et al., 2022; Johnson et al., 2017; Palhano-Fontes et al., 2019). Importantly, at 1-year follow-up we also found no change on the social relations and environment subscales of the WHOQOL-BREF, which point to pivotal aspects of participants' lives associated with recovery capital (Cano et al., 2017) and prompt further inquiry into how best to sustain benefits related to substance use severity and mental health following psychedelic-assisted treatment.

Last, this ATOP study highlights the complex and ambitious mission to identify the “active ingredients” or “essential elements” of ayahuasca- and other psychedelic-assisted treatments, a question of great interest to the ATOP research team since the design phase of the study (Rush et al., 2021). Such a pursuit underscores important ontological assumptions of the how and why of therapeutic mechanisms. From a Western scientific perspective, there is considerable focus on research design (the gold standard being the randomized controlled trial); sufficient sample size to facilitate multivariate analyses; and often studying relatively small-scale manipulation of the treatment protocol to get the optimal “recipe” matched to severity and case mix. In contrast to this “technical” model of treatment research, the explanatory models that typically characterize what has come to be known as Traditional Amazonian Medicine look to the ritualized engagement with the plants, such as ayahuasca, which are agential actors in the healing process, not simply a vector for the delivery of psychoactive alkaloids. This complicates the role of each aspect of treatment (psychotherapist, plants, co-living) and how they are considered to interact with each other. Relevant to this perspective is that clients at Takiwasi found it challenging to identify one activity within the Takiwasi program as more important than others. Although there was often a particular emphasis on the spiritual significance of ayahuasca and its importance for promoting well-being through providing insights and physical/emotional purging, qualitative data not reported here point to participants' perceived value of the “total program experience.” A common sentiment was expressed that the work with ayahuasca would not have been as effective without other treatment components (e.g., the diet, communal living, psychotherapy). Similarly, our qualitative data suggest that perceptions of individual treatment components changed considerably during and after the program experience. Given the integrative nature of this kind of treatment, we invite other researchers to step outside the technical research model and explore theoretical and conceptual frameworks from a more holistic starting point. Consistent with an Indigenous understanding of mental wellness, this will entail a bio-psycho-social-spiritual understanding of outcomes (Thunderbird Partnership Foundation, 2020) as well as an acknowledgment of the inherent interconnectedness of virtually all aspects of a given treatment and support protocol. Based on our experience, a positive step in this direction is to incorporate a mixed-methods approach within a complexity-based evaluation design, especially important given the interplay of set and setting (Stame, 2004; Hartogsohn, 2016, 2017). Although the current literature on the use of psychedelics gives attention to the concepts of set and setting as primary for understanding experience and outcomes, it is notable that these are mutable in a long-term and complex health intervention such as the one presented here. The role of therapeutic context resonates with developments in the science of program evaluation, in particular realist evaluation and complex interventions (Pawson & Tilley, 1997; Rog, 2012), and summarized succinctly by the phrase “outcome = intervention + context.” Our experience also points to the need for a broad perspective on the meaning of recovery from substance use and other mental health challenges (American Society of Addiction Medicine, 2013; Kaskutas et al., 2014). In particular, there is a need to emphasize the role of recovery capital (e.g., Yates & Malloch, 2010) and the concomitant use of a variety of outcome measures (Costello et al., 2020; Black et al., 2024).

Conclusion

Participants in the study improved on outcomes related to substance use severity and mental wellness 1 year after discharge. They also reported the experience with ayahuasca as a meaningful spiritual and therapeutic aspect of the treatment process. Even though no causal conclusions can be drawn about the effectiveness of ayahuasca alone because of the observational study design, the lack of a control group, and the complexity of the program, our findings suggest promise for its effectiveness within the Takiwasi multifactorial context for individuals with significant treatment histories, high levels of comorbidity, and internalized motivation for treatment. This contributes to a small but growing literature that calls for studies with more rigorous research designs. Further, the study reveals valuable methodological insights that researchers can benefit from and apply to future study design. This includes the importance of our mixed-methods design and our graphical analysis that highlights individual variation, as well as the measures used in the ATOP evaluation, especially our nuanced measure of treatment motivation and its potential to predict program retention and other determinants of outcomes.

With respect to mixed methods, several questions arise from the quantitative analysis that set the stage for analysis of the qualitative interviews to allow for a fuller understanding of outcomes and potential determinants. We have already found the qualitative data helpful in interpreting the data from the CSQ-8 such that, although participants rated their overall experience as very positive, they would only recommend it to people who seemed fitting for such a demanding and idiosyncratic program as Takiwasi Centre (see Marcus et al., 2022). Further, although we found few statistically significant associations between outcomes and participants' perceived importance of the various activities within the Takiwasi program, in future qualitative analyses we will continue to probe the important question of whether there are specific types of experiences and internal processes that can predict positive change in substance use and well-being. To that end we encourage other researchers studying ayahuasca- and psychedelic-assisted treatments to consider the active “ingredients” of a treatment protocol as internalized processes that may be triggered by program activities rather than a recipe comprising the program elements per se.

Conflict-of-Interest Statement

FM was research coordinator at the Takiwasi Centre between 2008 and 2017 and scientific coordinator of the Ayahuasca Treatment Outcomes Project (ATOP) between its creation in 2013 and May 2022. SGA has been research assistant and in charge of data collection of ATOP at the Takiwasi Centre between 2016 and May 2022. OM was employed as a research assistant at Takiwasi for 6 months in 2018.

Statements related to religious, spiritual, or political values made by persons working at the Takiwasi Centre, or their ideology, do not reflect those of the authors.