Therapist-Guided Forest Therapy in Adolescents: Operational Feasibility and Anxiety-Symptom Signals from a Two-Classroom Quasi-Experimental Pilot Study
Round 1
Reviewer 1 Report
Comments and Suggestions for Authors
This study is a quasi-experimental pilot study. The study assigned one high-school classroom (n = 24) to four weekly therapist-guided forest therapy sessions and a parallel classroom (n = 28) to usual activities. The intervention classroom showed directionally larger reductions on the Spence Children’s Anxiety Scale (SCAS) at four follow-ups over eleven weeks, while acute mood improvements emerged most clearly during the third session. The study is an original one and within the scope of the journal. It is an important study that could contribute to current literature after minor revisions. My comments are as follows;
- In introduction, the hypothesis of the study should also be clarified.
- The sample is 94% female (47/52). It limits the generalizability to male and gender-diverse adolescents. This issue should be included in limitations.
- The SCAS-C (Child version) was developed for children and young adolescents. Using it with 17–18-year-olds near the end of secondary school might be problematic. Authors should justify its use, consider the possibility of ceiling effects
- The CNS and nature-exposure questionnaire were administered only after the intervention. Without baseline data, the authors cannot claim the intervention changed nature connectedness, yet they treat these as exploratory outcomes. Revise your statements on this issue accordingly.
- While Benjamini–Hochberg correction was applied to POMS-A domain comparisons, the numerous exploratory analyses (threshold analyses, acute–longitudinal correlations, academic outcomes, nature questionnaire items) were not systematically corrected. Corrections should also be conducted to oother relevant tests. .
- The discussion repeatedly uses phrases like “associated with” and “produced” in ways that imply causation, despite the quasi-experimental, two-classroom design. Revise these statements.
- The authors should refrain to propose attentional restoration, emotion regulation, and stress recovery mechanisms, as they did not measure physiological, attentional, or emotion-regulation mechanisms.
- Limitations should be expanded based on above suggestions.
Author Response
Please see the attachment.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for Authors
The MS aims to present a study on the effectiveness of a nature-based intervention to reduce anxiety symptoms in adolescent students.
It looks easy to read, clear and informative. However, the initial enthusiasm declined, because I noticed a few issues which limits a lot the strength of the study I am listing below.
- You did not considered students with anxiety, but whole classes including students with nuanced/no sign of anxiety which makes impossible to draw conclusions on reduction of anxiety problems
- With 17-18 yrs old students I would have considered GAD-7 as an assessment instrument
- Very frequently there is a comorbidity with other diseases such as depression or learning disabilities. I would have included a larger set of assessment instruments to better understand the profiles and changes over time
- The sample size is limited to two whole classes, which reduces generalization
- Social aspects (perform the activities with the classmates) and the presence of the school psychologist are important variables to isolate. An active control group should have been included, performing something else during the whole mornings. Again, sessions led by a teacher (for instance) should have been included to assess the additional benefits of a psychologist as a trainee
Moreover
- The Discussion should have benefitted of a larger perspective including a reflection on the length and kind of the intervention: why not other form of nature intervention, such as gardening, mindfulness in the green, green exercise?
- In the tables and figures I would have reported the time, e.g. after one month etc. instead of the number for a more immediate understanding
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 3 Report
Comments and Suggestions for Authors
Overall this article is well organized with clear methodological practices. One of the greatest strengths of this paper is that the authors are very transparent about the limitations of this pilot study. It is clear throughout the entire paper that this is an exploratory study and the authors don't try to "over claim" any of the results.
Pilot/exploratory studies can be limited in their impact, however the fact that there is limited research related to the impact of nature on mental health issues specifically related to adolescence is important. As simply as demonstrating the possibility of implementing these sessions in a regular school curriculum is important.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 4 Report
Comments and Suggestions for Authors
Brief summary
This school-based quasi-experimental pilot study examines whether four therapist-guided forest therapy sessions are associated with changes in anxiety symptoms and acute mood among late adolescents. The repeated follow-up design, use of validated symptom measures, and attention to feasibility are strengths, but the one-classroom-per-condition design, lack of randomization, small sample, and extensive exploratory analyses substantially limit causal interpretation.
Please see my comments below.
Introduction
The rationale for focusing on adolescents is well developed, and the manuscript appropriately identifies the lack of rigorous adolescent studies and the frequent use of passive comparators in the existing literature. However, the specific clinical contribution should be stated more precisely. This study is best positioned as a feasibility and signal-detection study rather than an evaluation of treatment efficacy.
The term "forest therapy" should be defined with greater precision and consistently distinguished from forest bathing, nature exposure, outdoor activity, and therapist-guided psychotherapy in nature. Because the intervention includes therapist guidance, multisensory attention, slow walking, peer-group interaction, and time away from normal classroom activities, the active ingredient is not necessarily forest exposure alone.
The Introduction could provide a stronger mental health framework for why the intervention might affect anxiety specifically. The proposed mechanisms of attentional restoration, emotion regulation, stress recovery, and co-regulation are plausible, but it would help to distinguish which are hypothesized to influence acute mood versus persistent anxiety symptoms.
The authors appropriately distinguish acute affective outcomes from longer-term anxiety trajectories. This is a conceptual strength and should perhaps be emphasized even more clearly as one of the principal contributions of the study.
Methods
The most important methodological limitation is that there is only one intervention classroom and one control classroom. Classroom is therefore completely confounded with treatment condition. Statistical adjustment at the student level cannot separate the intervention from pre-existing classroom climate, peer relationships, teacher effects, expectancy, or other classroom-specific influences. This should remain central to the interpretation of all between-group findings.
The use of one-sided hypothesis tests is difficult to justify in a study that was not preregistered. The authors acknowledge this and report two-sided analyses alongside them, but the one-sided findings risk receiving disproportionate emphasis, particularly because several two-sided comparisons do not reach conventional significance. I would recommend prioritizing the two-sided estimates, confidence intervals, and effect sizes in the main text, with directional one-sided analyses clearly relegated to exploratory sensitivity analyses.
The primary analytical sample excludes two intervention students who attended fewer than three sessions. Although the authors provide an all-assigned sensitivity analysis, the primary analysis is effectively per protocol. Given the quasi-experimental classroom-level allocation, I would favor presenting the all-assigned analysis as the principal analysis and the attendance-restricted analysis as secondary.
The statistical analysis is extensive relative to the sample size. ANCOVA, Mann-Whitney tests, Welch tests, repeated-measures mixed models, GEE, responder analyses, threshold analyses, moderation analyses, acute-longitudinal correlations, academic outcomes, and nature-related outcomes are all conducted in a sample of approximately 52 students. This raises a substantial risk of unstable estimates and chance findings. The manuscript would benefit from a clearer hierarchy distinguishing one primary outcome and a small number of secondary outcomes from the many exploratory analyses.
More information is needed regarding intervention fidelity. The school psychotherapist and psychologist received protocol training, and each session followed a structured format. However, the manuscript does not report whether fidelity was formally assessed, whether a checklist was used, or whether the same therapeutic prompts and duration were maintained across sessions. This is particularly important if the intervention is intended to be reproduced in other schools.
Results
The central SCAS findings should be presented more cautiously. None of the baseline-adjusted two-sided comparisons in the primary analytical sample crossed p < .05, with the strongest result at SCAS4 being coefficient = -5.30, 95% CI -10.69 to 0.09, p = .054. The manuscript generally acknowledges this, but language such as "more favorable anxiety-symptom trajectory" should consistently be framed as preliminary rather than indicative of efficacy.
The all-assigned sensitivity analysis yields a nominally significant SCAS4 result, whereas the attendance-restricted primary analysis does not. The authors should discuss this discrepancy explicitly rather than simply describing the overall pattern as preserved. With such a small sample, modest changes in analytic inclusion can materially affect statistical significance.
The acute POMS findings are interesting, particularly the large change observed during the third session. However, because POMS-A was not collected in the control classroom, these changes cannot be attributed specifically to forest therapy. Repeated testing, relief from classroom activities, social interaction, therapist attention, weather, and simple regression toward the mean remain plausible explanations.
The exploratory findings regarding school conduct and nature connectedness deserve careful presentation. The intervention classroom had significantly lower adjusted conduct scores and significantly lower post-intervention nature-connectedness scores than controls. Although these outcomes were exploratory and lack baseline nature-connectedness data, they should not be minimized simply because they do not align with the hypothesized direction. They may indicate baseline differences or other classroom-level confounding, which actually reinforces the main design limitation.
Discussion
The Discussion appropriately states that the results are preliminary signals rather than evidence of efficacy and explicitly acknowledges that the study cannot separate the intervention from classroom-level effects. This caution should be maintained throughout the Abstract, Discussion, and Conclusion, particularly when discussing possible benefits for students with higher baseline anxiety.
The manuscript proposes several mechanisms, including multisensory attention, therapist-guided pacing, attentional restoration, emotion regulation, and group co-regulation. These are plausible hypotheses, but none were directly measured. The authors should avoid implying mechanistic support and instead present these as candidate pathways for future mediation studies.
From a school mental health and medical education perspective, the feasibility findings are potentially useful. However, the manuscript should discuss what level of therapist training is required, whether this intervention could realistically be implemented by school mental health professionals, and how safeguarding, identification of significant anxiety, referral pathways, and management of students who deteriorate would be handled in routine practice.
The limitations section is appropriate and already identifies the major concerns, including the quasi-experimental design, one classroom per condition, small and predominantly female sample, lack of preregistration, passive control, expectancy effects, possible contamination, weather confounding, and multiple exploratory analyses. I would add that the lack of blinded outcome assessment and exclusive reliance on student self-report for the principal mental health outcome may further increase expectancy and reporting effects.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report
Comments and Suggestions for Authors
I appreciate the efforts of the AA in revising. However in my opinion this study is not publishable per se for all the issues I listed in my previous revision with the term 'reject'. As the AA themselves recognize it should be considered as a pilot. A MS including this preliminary feasibility trial and at least a regular trial with selected clinical populations and all the variables controlled for could reach the publication standards.
Author Response
|
Comment 1: I appreciate the efforts of the AA in revising. However in my opinion this study is not publishable per se for all the issues I listed in my previous revision with the term 'reject'. As the AA themselves recognize it should be considered as a pilot. A MS including this preliminary feasibility trial and at least a regular trial with selected clinical populations and all the variables controlled for could reach the publication standards. |
|
Response 1: Thank you for reconsidering the revised manuscript. We understand and respect the Reviewer’s position. We agree with its central substantive premise: this two-classroom, non-clinically selected pilot cannot establish clinical efficacy and is not a substitute for a prospectively designed trial in a clinically selected population. The revised manuscript now states this explicitly from the title and Abstract through the Methods, Discussion, Limitations, Future Research, and Conclusions. As detailed in our previous point-by-point response, the participants were two intact classrooms and were not selected by anxiety symptoms, diagnosis, or treatment status; SCAS was used only as a continuous symptom indicator; classroom and condition were perfectly confounded; no principal SCAS comparison survived multiplicity correction; the passive comparator could not isolate forest exposure, professional guidance, peer interaction, movement, expectancy, or time away from lessons; and the absence of diagnostic, comorbidity, concurrent-treatment, and complementary late-adolescent assessment data is explicitly acknowledged. The manuscript therefore makes no causal, therapeutic, or clinical-efficacy claim. Its deliberately limited contribution is to document operational delivery in an ordinary school setting, report descriptive and uncertainty-qualified signals—including discordant findings—and provide design information for future research. We respectfully differ only on whether those pilot contributions must be combined in the same manuscript with a new clinical trial in order to be reportable. Adding a ‘regular trial with selected clinical populations and all the variables controlled for’ is not a revision that can be performed on the present dataset. It would require a new prospectively approved protocol, prespecified clinical eligibility and outcome measures, an a priori sample-size calculation, new recruitment and consent, adequate randomization or replicated cluster allocation, an active comparator, and prospective measurement of comorbidity, concurrent care, fidelity, and other relevant covariates. Retrospectively treating such a future study as part of the present pilot would be methodologically inappropriate, and such a study would independently require prospective ethics approval. Section 4.6 already distinguishes universal school-based prevention research from clinically targeted intervention and specifies these requirements for future studies. Because the Round 2 comment does not identify a remaining correctable error in the revised manuscript or a new analysis that can validly be conducted with the existing data, we have not added further analyses or altered the scientific content in response to this comment. We hope the study can be assessed according to its explicitly delimited aim—as an operational-feasibility and descriptive signal-detection pilot—rather than as evidence of efficacy in a clinical population, which it does not claim to provide. |
Reviewer 4 Report
Comments and Suggestions for Authors
The authors adressed all my previous comments. I have no further comments.
Author Response
|
Comment 1: The authors adressed all my previous comments. I have no further comments. |
|
Response 1: We sincerely thank the Reviewer for confirming that all previous comments have been satisfactorily addressed. We are grateful for the Reviewer’s careful and constructive feedback, which substantially strengthened the manuscript. As no further comment was raised, no additional change to the manuscript or Supplementary Materials was required in response to this review round. |
