Within-treatment changes in a novel addiction treatment program using traditional Amazonian medicine

Berlowitz and colleagues situate their work within a renewed scientific interest in psychedelics and note that ethnobotanical psychoactive substances have long been used in indigenous medical contexts. The introduction outlines ayahuasca (the DMT-containing decoction prepared from Banisteriopsis caapi plus DMT-containing leaves) as an example of a traditionally used medicine with growing evidence for therapeutic potential in conditions including substance use disorders, depression and anxiety. The authors describe the Takiwasi Center in Peru, a nationally accredited residential programme that has combined ayahuasca and other Amazonian plant-based techniques (including dietary retreats and emetic preparations) with Western biomedical and psychotherapeutic approaches since 1992, but they note scientific characterisation of outcomes at the centre has been limited. This study aims to address gaps in the earlier observational evaluations of Takiwasi by examining within-treatment change with higher temporal resolution. Rather than isolating single treatment components, the researchers adopt an observational "black box" approach, using repeated psychological and neuropsychological measurements across defined treatment milestones to characterise the timing and magnitude of patient changes during the residential programme.

The study was an observational repeated-measures cohort conducted at the Takiwasi residential centre in Tarapoto, Peru. Ethical approval was obtained and all participants gave written informed consent. Participation was open to patients admitted to Takiwasi for addiction treatment during April 2014 to August 2015; no patients declined participation and study participation ceased when a patient left treatment. The final analytic sample comprised 36 male inpatients (ages 20–50, mean 29, SD 7). The sample was international (58% South American, 28% European, 14% North American). Commonly reported substances prior to admission were alcohol (83%), cannabis (71%) and cocaine (51%), with 66% reporting poly-drug use. Treatment completion outcomes were reported: 61% completed treatment, 22% exited voluntarily against staff recommendation, 14% were suspended, and 3% abandoned without advising staff. Takiwasi's treatment protocol is a multi-stage, largely inpatient programme typically lasting around 9 months (flexible), comprising initial isolation (about 8–10 days), main treatment (around 7 months) and a reinsertion phase (~2 months). The programme integrates physical detoxification (including emetic and psychoactive plants), group and individual psychotherapy, occupational therapy and community living, spiritual and psychological development through psychoactive plant sessions and dietary retreats, and biomedical evaluation. The centre excludes individuals with a history of psychosis and certain serious medical conditions; psychiatric medications are ceased prior to entry where interactions are possible. Measures included the Addiction Severity Index (ASI) administered at intake, and a clinical battery repeated at treatment milestones: Perceived Stress Scale (PSS-10), Craving Experience Questionnaire—frequency (CEQ-F), Brief Symptom Inventory Global Severity Index (BSI GSI), Spiritual Well-Being Scale (SWBS; Religious and Existential subscales), SF-36v2 Physical and Mental Component Summaries (PCS and MCS), and a single Self-Evaluated Transition (SET) item from SF-36v2. Neuropsychological functioning was assessed with the RBANS in a Spanish-speaking subsample (n = 8) at intake and at a follow-up at least 2 months into treatment. Clinical battery milestones (M1–M5) corresponded to admission, approximately 1 month after each diet, and exit. Sample sizes at M1–M5 were 22, 19, 18, 13 and 9 respectively; average days in treatment at these milestones were 3, 110, 169, 245 and 309 days. Analyses were performed in R. Intake ASI and some clinical battery scores were compared with available normative samples using one-sample t-tests. Within-treatment change on the clinical battery was modelled with mixed-effects models (fixed effect: treatment milestone as a categorical predictor; random effect: patient intercept only) to make use of all available repeated measures and account for within-subject dependence. Addition of total treatment time (days) was tested but not retained based on AIC. Paired-samples t-tests (Hedges' corrected effect sizes) compared RBANS intake and follow-up scores. A logistic regression examined predictors of early dropout (defined as <30 days in treatment) using demographics and ASI intake scores.

Intake profile: On intake the Takiwasi sample showed a high severity of addiction. ASI composite comparisons with available normative inpatient samples indicated elevations in drug-related and family problem domains relative to mainstream centres; the extracted text does not clearly report all specific t-statistics for every ASI domain. ASI psychiatric cutoff comparisons and elevated intake BSI GSI scores suggested psychiatric comorbidity was common. Medical problems were, on average, less severe than a Canadian readmission sample and comparable to US inpatient averages. On the intake clinical battery (n = 22 for most measures, n = 27 for CEQ-F), perceived stress and craving frequency were high, and mental and emotional health were low. Religious well-being (RWB) was possibly lower than US mental health patients (t(21) = -2.21, p = 0.038, d = -0.47, 95% CI for d (-1.37, 0.43)). Existential well‑being (EWB) was not different from US mental health patients (t(21) = -0.41, p = 0.686). CEQ-F scores were significantly higher than an Australian alcohol outpatient sample (t(26) = 4.01, p < 0.001, d = 0.77, 95% CI (-0.05, 1.59)). Cognitive functioning at intake was described as around the "low average" range. Within-treatment changes: Mixed-effects models comparing each milestone against M1 showed statistically significant and clinically positive changes across the repeated measures. For measures where higher scores indicate worse outcomes (for example perceived stress, GSI, craving), model estimates were significant (all p < 0.001 except GSI at M3 where p < 0.01), indicating reductions over time. For measures where higher scores indicate better outcomes (for example SF-36 MCS, SWBS EWB), estimates indicated improvements. The authors report particularly strong increases in mental and emotional health and existential well-being, and large reductions in perceived stress, psychiatric symptoms and craving. The most pronounced changes occurred early in treatment and were generally maintained thereafter, though later-stage analyses had limited statistical power. The SET single-item ratings (n = 22 at M1; n = 28 across M2–M5 with 59 administrations) showed that after intake no patients in treatment rated their health as "somewhat worse" or "much worse," and only two rated it "about the same," with the rest indicating improvement. Neuropsychological functioning: In the RBANS subsample (n = 8), mean scores increased across all indexes from intake to follow-up. Paired t-tests found significant improvement for the Total Scale (t(7) = 3.37, p = 0.012, Hedges' g = 0.55, 95% CI for g (0.17, 0.92)) and for Delayed Memory (t(7) = 2.73, p = 0.029, g = 0.74, CI (0.08, 1.40)). Other index comparisons are reported in the paper's tables but were not all statistically significant. Dropout analysis: Early treatment dropouts (n = 6; defined as <30 days in treatment) did not differ from the rest of the sample (n = 30) in nationality, religion or ASI intake scores according to logistic regression. However, age was a significant difference: early dropouts were younger (mean 22 years, SD 2) than the rest (mean 30 years, SD 7), t(26) = -5.24, p < 0.001, Cohen's d = -1.24, 95% CI for d (-2.17, -0.30).

Berlowitz and colleagues interpret their findings as indicating that patients admitted to Takiwasi present with relatively severe addiction profiles and psychiatric comorbidity, yet tend to show clinically meaningful improvements across multiple domains during residential treatment. The authors emphasise that improvements—large reductions in perceived stress, psychiatric symptoms and craving, increases in mental and emotional health and in existential well‑being—appeared relatively early in the treatment course and were generally maintained thereafter. Neuropsychological data provided additional evidence for improvement, notably in delayed memory. The authors situate their results alongside prior observational reports from Takiwasi and the wider literature on indigenous psychoactive plant sacraments and classical hallucinogens, noting consistency with earlier findings of posttreatment improvements. They highlight a divergence between existential and religious dimensions of spiritual well‑being: existential well‑being showed stronger predicted increases, whereas religious well‑being was more variable, a pattern the authors suggest may reflect baseline differences in religiosity and differing patient responses to spiritual content in plant ceremonies. The authors explicitly acknowledge limitations arising from the observational, "black box" design, which precludes causal attribution to specific treatment components. They note other plausible contributors to observed change including enforced abstinence, removal from daily life, development of community within the centre, and the passage of time. They also discuss safety considerations: while ayahuasca has an acceptable pharmacological safety profile, many Amazonian techniques used at Takiwasi remain scientifically understudied; the centre's exclusion criteria, biomedical evaluations and practitioner expertise likely increase safety, particularly for international patients where drug–medicine interactions might be more likely. The authors caution that the admission protocol selects for motivated patients, which limits generalisability, and that follow-up studies are required to determine whether within-treatment changes translate into longer-term recovery in community settings. As implications, the authors call for further clinical research to assess effectiveness and longitudinal outcomes, and for qualitative work to provide converging evidence on mechanisms, the nature of treatment components (for example the role of dietary retreats versus ayahuasca ceremonies), and concepts of health and safety in this treatment context. They also note the potential relevance of gut–brain axis mechanisms given the focus on emetic and plant-based treatments.

The authors conclude that Takiwasi represents a distinctive, long-standing example of a treatment approach that integrates traditional Amazonian medicine with Western clinical practice, and that the observed within-treatment improvements are consistent with growing evidence for therapeutic effects of indigenous psychoactive plant sacraments and classical hallucinogens. They stress that further clinical trials, longitudinal follow-up and qualitative studies are needed to evaluate treatment effectiveness, delineate mechanisms, and clarify safety and appropriateness for different patient populations.