Nature-Based Health Interventions for People with Mild to Moderate Anxiety, Depression, and/or Stress: Identifying Target Groups, Professionals, Mechanisms, and Outcomes Through a Delphi Study

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Dear Authors,

Thank you for this interesting manuscript. The methodology combining systematic review, stakeholder engagement, and Delphi consensus represents a thoughtful approach to a complex research question.

Your manuscript reports findings from a two-round Delphi study aimed at identifying key elements for nature-based health interventions (NBHIs) targeting individuals with mild to moderate anxiety, depression, and/or stress. The work addresses an important gap in developing systematic frameworks for NBHIs and makes a valuable contribution to the field. The manuscript would be strengthened by revisions that enhance clarity in operational definitions, improve methodological transparency (particularly regarding the preparatory phase and analytical processes), and develop the theoretical framing more fully. These revisions will help readers better understand the study's foundation and apply the findings in their own contexts.

Clarification regarding target group definition and operationalisation

Line 53-54: You repeatedly reference "mild to moderate anxiety, depression, and/or stress" but never operationalises these terms. Which diagnostic criteria, screening tools, or severity thresholds define "mild to moderate"? Was this made clear to Delphi respondents? They would need clear parameters to evaluate the statements appropriately. Without a clear definition, the target group boundaries remain ambiguous throughout.

You group stress with anxiety and depression as part of the target population, which is pragmatically justified given high comorbidity (lines 422-440). However, this creates an operational definition problem: unlike anxiety and depression, stress lacks standardised diagnostic criteria. The manuscript should explicitly address upfront how 'mild to moderate stress' is identified and measured, particularly since stress is a universal human experience rather than a clinical diagnosis. What distinguishes actionable 'mild to moderate stress' warranting intervention from normal life stress? Additionally, could you briefly explain the shift from 'mild to severe stress' in the systematic review (line 129-130) to 'mild to moderate' in this study.

I would recommend providing explicit operational definitions, preferably in a table, specifying severity criteria for each condition (e.g., PHQ-9 scores 5-14 for mild-moderate depression; GAD-7 scores 5-14 anxiety; DASS-21 Stress subscale scores or PSS scores for stress).

Clarification regarding the Delphi panel

Lines 155-169: The decision to limit the Delphi panel to researchers, whilst understandable given the focus on target population expertise, raises questions about perspectives on implementation feasibility and delivery practicalities. This is particularly relevant for Delphi items directly addressing delivery aspects (interdisciplinary teams, professional backgrounds, competencies).

Notably, practitioners were meaningfully involved in the preparatory stakeholder dialogues (lines 136-154), which informed the Delphi questionnaire. However, I would recommend that you better articulate how practitioner input from the preparatory phase specifically shaped the Delphi items. Also, is it possible that the near-consensus on delivery-related items (e.g., interdisciplinary teams 69%, group composition 46%) could have benefited from practitioner representation in the Delphi rounds.

Lines 580-593: In the limitations section you acknowledge the homogeneous researcher panel but I recommend you discuss more explicitly how this may have shaped specific findings. For instance, does the panel composition help explain the lack of consensus on delivery-oriented items (group composition, team structure) versus strong consensus on theoretical mechanisms? Which findings should readers view as most robust given this limitation, and which might benefit from future validation with practice-based stakeholders?

Findings from the preparatory phase

It is not clear to me why the preparatory phase has not been included in this manuscript. If there is a valid reason, I recommend it is justified more clearly. Based on the information available to me, I recommend including evidence from the preparatory phase as I believe it is the foundation for the work you have presented.

Lines 110-154: You describes preparatory phase methods (systematic review and stakeholder dialogues) but the actual findings from this work are not presented. Presenting these findings would strengthen the manuscript considerably: readers would be able to assess what informed the Delphi items and evaluate how effectively the preparatory work established the foundation for expert evaluation

While reading the manuscript I would have appreciated information regarding the following:

• What specific findings from each source informed which Delphi items? The systematic review identified 45 outcome measures and variation in effects—how did this translate into the 21 Delphi items? Which stakeholder themes became which questions?
• How were the stakeholder dialogue meetings analysed? Lines 141-154 mention qualitative analysis but you do not specify the method (thematic/framework/content analysis) or present any themes or findings.
• How did the research team synthesise different inputs? What process integrated quantitative review evidence with qualitative stakeholder perspectives? When sources conflicted, how were decisions made?
• What were the item selection criteria? Twenty-eight stakeholders (including substantial practitioner representation) participated in dialogues—what did they contribute? Which suggestions were incorporated versus excluded, and why?

Without this information regarding the preparatory findings, readers cannot evaluate whether the Delphi items adequately represent the evidence base and stakeholder perspectives, whether practitioner input from stakeholder dialogues was meaningfully incorporated despite the researcher-only Delphi panel, or to what extend the study is evidence-based and practice-informed.

Line 91: The Nature Impact project requires proper citation or website reference for broader programme context.

I recommend adding a table (or supplementary appendix if space is an issue) presenting key themes/findings from stakeholder dialogues (with analysis method specified); synthesis of systematic review findings relevant to the four Delphi domains; explicit traceability mapping (e.g., table format) showing how preparatory sources informed specific Delphi items; decision-making criteria for item inclusion/exclusion.

This transparency would allow readers to assess the foundation of the Delphi study for themselves rather than accepting it was "robust" on assertion alone.

Clarification regarding analytical transparency

Lines 213-225: Whilst Reflexive Thematic Analysis is cited, the analysis lacks transparency and reflexivity:

• Who conducted the initial coding? Was it a single researcher or multiple coders?
• Were themes developed inductively, deductively, or through a hybrid approach?
• How was consensus achieved within the research team?
• Where is the reflexivity? No discussion of researcher positionality, assumptions, or how these may have shaped interpretation
• No illustrative quotes are provided to support identified themes

The integration of quantitative consensus data with qualitative theme development is methodologically interesting but requires clearer explanation. How did the two analytical streams inform each other?

I recommend adding a reflexivity statement addressing researcher backgrounds and potential biases. Also, include representative quotes (adequately anonymised) to illustrate each theme, and provide a clearer account of the analytical process, potentially with a diagram showing how quantitative and qualitative strands were integrated.

Clarification regarding Consensus Criteria and Interpretation

Lines 186-190: You have employed dual consensus criteria (mean ≥5.0 AND ≥70% selecting categories 5-7), citing consistency with Delphi methodological conventions. However, Delphi consensus criteria are notoriously variable across studies, and the justification for this specific approach requires strengthening. I recommend more consideration and clarification regarding these points:

• "Near consensus" at 69% (Line 246-249) appears arbitrary. Is 69% meaningfully different from 70%?
• Items with high means but lower percentage agreement (or vice versa) warrant more nuanced discussion
• The shift in prioritisation between rounds (especially "participation in everyday life" dropping from rank 1 to rank 11) deserves deeper theoretical engagement

Lines 276-284: The interpretation that ranking forced reconsideration is plausible, but alternative explanations exist: Respondents may have interpreted "mechanisms of change" differently in a ranking vs. rating context; the distinction between outcomes and mechanisms may have been unclear; forced ranking may have revealed hierarchy rather than changing minds

I recommend that you discuss alternative interpretations and acknowledge the complexity of comparing rating and ranking data. Consider whether the two approaches measure the same construct.

Clarification regarding specification of healthcare resources and treatment context

Line 56: You mention "substantial demands on healthcare resources" but never specify what types of demands: financial capacity, staff capacity, infrastructure, or service provision capacity? This would help understanding how NBHIs might alleviate pressure.

Line 58: "Existing treatment options" remains vague. Which specific treatments are available but inadequate? Cognitive behavioural therapy? Pharmacotherapy? Stepped care models? Understanding what NBHIs complement or replace is essential.

Lines 76-78: Clarification needed regarding whether NBHIs are envisioned as preventative interventions, alternative treatments, complementary add-ons to traditional care, or rehabilitation tools.

I recommend adding an explicit discussion of how NBHIs fit within existing treatment pathways and what specific gaps they address.

Justification for pre-selection of nature connectedness

Lines 194-196: Pre-selecting nature connectedness as an outcome introduces bias into the Delphi process. Whilst theoretically justified, this decision removes one outcome from democratic deliberation. It may have influenced respondents' prioritisation of the remaining outcomes and represents a theoretical commitment made before expert consultation

I recommend that you More explicitly acknowledge this as a limitation and justify the decision with reference to the preparatory work.

Stakeholder Exclusion

Lines 147-150: The manuscript acknowledges that stakeholder dialogues "did not directly include representatives from volunteer organisations" and that this "may have constrained the diversity of perspectives." This is presented somewhat passively. Given that many NBHIs operate in the voluntary sector, this exclusion is consequential.

Lines 147-150: The decision to anchor within formal healthcare was influenced by stakeholder dialogues that excluded voluntary sector voices—a circular justification.

Could you more critically examine how this exclusion may have biased the findings toward medicalised, professionalised models of NBHIs.

Mechanisms of Change: Theoretical Integration

Your manuscript identifies three core mechanisms (nature interaction, social community, physical activity)

Lines 347-374: Theme 3 acknowledges interdependence but lacks theoretical synthesis. Could you consider if/how these mechanisms interact; are they sequential, parallel, or synergistic? Do different mechanisms matter more for anxiety vs. depression vs. stress? How do these relate to the theoretical frameworks mentioned (Stress Reduction Theory, Attention Restoration Theory, Lines 66-71)?

I suggest that you engage more explicitly with existing theoretical models.

 

 

Line 356: Delete 'being' (…making strict ranking being inadequate to capture their interdependence)

Line 359-362: “Mechanisms like body awareness or mindfulness, while potentially beneficial, were described as distressing for some individuals, particularly those with trauma histories or heightened internal stress.” Without additional context I don't follow the logic here.

 

This manuscript addresses an important gap in NBHI research by systematically identifying key considerations for intervention development. The study employs a sound multi-phase methodology and achieves consensus on several core elements, including mechanisms (nature interaction, social community, physical activity), outcome priorities (mental wellbeing, quality of life), and professional competencies. The finding that legitimate variation exists in delivery models and group composition is valuable, appropriately reflecting the context-dependent nature of NBHIs.

The manuscript has notable strengths: the preparatory phase integrating systematic review with stakeholder engagement provides a solid foundation, and the combination of Delphi consensus with qualitative thematic analysis adds interpretive depth to the findings. Positioning NBHIs within the MRC complex interventions framework is a useful theoretical contribution.

The manuscript would benefit from enhanced transparency in three areas: operational definitions (particularly for 'mild to moderate' presentations), fuller reporting of the preparatory phase findings, and more explicit discussion of analytical processes including researcher reflexivity. These revisions would strengthen the manuscript's utility for practitioners, policymakers, and researchers seeking to design and evaluate context-adapted NBHIs.

With these revisions, this work has the potential to make a meaningful contribution to evidence-based NBHI development in Denmark and provide useful guidance for similar efforts internationally.

Good luck with the publication of your work.

Author Response

Clarification regarding target group definition and operationalisation

Comments 1

Line 53-54: You repeatedly reference "mild to moderate anxiety, depression, and/or stress" but never operationalises these terms. Which diagnostic criteria, screening tools, or severity thresholds define "mild to moderate"? Was this made clear to Delphi respondents? They would need clear parameters to evaluate the statements appropriately. Without a clear definition, the target group boundaries remain ambiguous throughout.

Response 1

Thank you for noticing the emphasis on a target group of people with mild to moderate anxiety, depression, and/or stress. Diagnostic criteria for mild to moderate anxiety, depression, and/or stress were not included, as this Delphi study draws on expert perspectives grounded in municipal practice, where service provision is necessarily pragmatic and not based on fixed diagnostic categories. The Delphi respondents have extensive experience with the target group and therefore based their assessments on functional needs and real-world practice rather than formal diagnoses. Furthermore, they were thoroughly informed about the context.

 

Comments 2

You group stress with anxiety and depression as part of the target population, which is pragmatically justified given high comorbidity (lines 422-440). However, this creates an operational definition problem: unlike anxiety and depression, stress lacks standardised diagnostic criteria. The manuscript should explicitly address upfront how 'mild to moderate stress' is identified and measured, particularly since stress is a universal human experience rather than a clinical diagnosis. What distinguishes actionable 'mild to moderate stress' warranting intervention from normal life stress? Additionally, could you briefly explain the shift from 'mild to severe stress' in the systematic review (line 129-130) to 'mild to moderate' in this study.

I would recommend providing explicit operational definitions, preferably in a table, specifying severity criteria for each condition (e.g., PHQ-9 scores 5-14 for mild-moderate depression; GAD-7 scores 5-14 anxiety; DASS-21 Stress subscale scores or PSS scores for stress).

Response 2

Thank you for the interesting question. As mentioned above, diagnostic criteria for mild to moderate anxiety, depression, and/or stress were not included. The study seeks to elicit consensus among experts based on their practice-based knowledge. Given the municipal context, where interventions are delivered pragmatically and not according to fixed diagnostic categories, the experts drew on their extensive experience with the target group rather than formal diagnoses. This is in contrast to the systematic review, where diagnostic criteria was operationalised. However, in the systematic review, it was not possible to differentiate between low, moderate, and severe stress; therefore, we chose to include the full spectrum in order to capture studies reporting stress levels (PSS >16).

 

Clarification regarding the Delphi panel

Comments 3

Lines 155-169: The decision to limit the Delphi panel to researchers, whilst understandable given the focus on target population expertise, raises questions about perspectives on implementation feasibility and delivery practicalities. This is particularly relevant for Delphi items directly addressing delivery aspects (interdisciplinary teams, professional backgrounds, competencies).

Response 3

We agree with the reviewer that restricting the Delphi panel to researchers may have limited perspectives on implementation feasibility and delivery practicalities, particularly for items addressing interdisciplinary teams and professional competencies. This issue has now been addressed more explicitly in the Limitations section.

We clarify that while the researcher-only panel strengthened conceptual coherence and expertise related to the target population, delivery-oriented findings may reflect research-based perspectives rather than practice-based experience. Importantly, however, practice-based stakeholders were not excluded from the broader research process. As illustrated in Figure 1, they were involved in the preparatory phase and are re-engaged after the Delphi process to co-create and refine the final programme theory.

 

Comments 4

Notably, practitioners were meaningfully involved in the preparatory stakeholder dialogues (lines 136-154), which informed the Delphi questionnaire. However, I would recommend that you better articulate how practitioner input from the preparatory phase specifically shaped the Delphi items. Also, is it possible that the near-consensus on delivery-related items (e.g., interdisciplinary teams 69%, group composition 46%) could have benefited from practitioner representation in the Delphi rounds.

Response 4

We appreciate the reviewer’s focus on practitioner involvement in the preparatory phase. Practitioner perspectives from the stakeholder dialogue meetings directly informed the formulation of several Delphi items, particularly those related to target group definition, delivery context, and competency requirements. This linkage has now been made more explicit in the revised manuscript and is documented through an item-level traceability mapping in Supplementary Appendix 1 (see also response to Comment 8).

 

Comments 5

Lines 580-593: In the limitations section you acknowledge the homogeneous researcher panel but I recommend you discuss more explicitly how this may have shaped specific findings. For instance, does the panel composition help explain the lack of consensus on delivery-oriented items (group composition, team structure) versus strong consensus on theoretical mechanisms? Which findings should readers view as most robust given this limitation, and which might benefit from future validation with practice-based stakeholders?

Response 5

We thank the reviewer for this valuable observation. We have now expanded the limitations section to explicitly address how the homogeneous researcher panel may have shaped the findings. As requested, we clarify that the panel’s strong clinical and theoretical orientation likely contributed to the high consensus on conceptual mechanisms, while the lack of practice-based perspectives may explain the absence of consensus on delivery-oriented topics such as group composition and team structure.

We also specify which findings may be considered most robust (mechanisms and outcomes) and which require further validation (delivery models and professional roles). Finally, we highlight the need for future work involving practitioners, service users, and nature-based facilitators to complement the researcher perspective.

 

Findings from the preparatory phase

Comments 6

It is not clear to me why the preparatory phase has not been included in this manuscript. If there is a valid reason, I recommend it is justified more clearly. Based on the information available to me, I recommend including evidence from the preparatory phase as I believe it is the foundation for the work you have presented.

Response 6

We thank the reviewer for this observation. The initial decision not to include detailed preparatory findings in the main manuscript was pragmatic, reflecting space limitations and the fact that the systematic review has been published elsewhere.

In response to this and related comments, we have now expanded the manuscript to include a clearer summary of key findings from the preparatory phase and clarified how these informed the Delphi process. Further details are provided in response to Comment 8 and in Supplementary Appendix 1.

 

Comments 7

Lines 110-154: You describes preparatory phase methods (systematic review and stakeholder dialogues) but the actual findings from this work are not presented. Presenting these findings would strengthen the manuscript considerably: readers would be able to assess what informed the Delphi items and evaluate how effectively the preparatory work established the foundation for expert evaluation

Response 7

We agree that presenting findings from the preparatory phase strengthens the manuscript and improves transparency. The manuscript has been revised to more explicitly summarise key findings from both the systematic literature review and the stakeholder dialogue meetings, including how these informed the formulation of Delphi items.

To support traceability, we have added Supplementary Appendix 1, which maps preparatory sources to specific Delphi items. A fuller explanation of the integration and decision-making process is provided in response to Comment 8.

 

Comments 8

While reading the manuscript I would have appreciated information regarding the following:

• What specific findings from each source informed which Delphi items? The systematic review identified 45 outcome measures and variation in effects—how did this translate into the 21 Delphi items? Which stakeholder themes became which questions?
• How were the stakeholder dialogue meetings analysed? Lines 141-154 mention qualitative analysis but you do not specify the method (thematic/framework/content analysis) or present any themes or findings.
• How did the research team synthesise different inputs? What process integrated quantitative review evidence with qualitative stakeholder perspectives? When sources conflicted, how were decisions made?
• What were the item selection criteria? Twenty-eight stakeholders (including substantial practitioner representation) participated in dialogues—what did they contribute? Which suggestions were incorporated versus excluded, and why?

Without this information regarding the preparatory findings, readers cannot evaluate whether the Delphi items adequately represent the evidence base and stakeholder perspectives, whether practitioner input from stakeholder dialogues was meaningfully incorporated despite the researcher-only Delphi panel, or to what extend the study is evidence-based and practice-informed.

Response 8

We thank the reviewer for this important comment. We agree that clearer presentation of the preparatory phase strengthens the manuscript and improves transparency regarding how the Delphi items were developed.

In response, we have revised the manuscript to more explicitly summarise key findings from both preparatory sources and clarify their analytical treatment. The systematic literature review is already published in BMJ Open and thus we had only presented limited details of the study: Jessen NH, Løvschall C, Skejø SD, Madsen LSS, Corazon SS, Maribo T, et al. Effect of nature-based health interventions for individuals diagnosed with anxiety, depression and/or experiencing stress-a systematic review and meta-analysis. BMJ Open. 2025;15(7):e098598. doi: 10.1136/bmjopen-2024-098598. Now more details have been added to the methods section.

We have also clarified how the stakeholder dialogue meetings were analysed and how their findings contributed practice-based perspectives on implementation, contextual constraints, and feasibility. These insights were synthesised with review findings and selectively incorporated into the Delphi questionnaire based on alignment with the study aim and programme theory development.

To enhance transparency, we have added supplementary material (Appendix 1) documenting traceability between preparatory sources and individual Delphi items. Together, these revisions clarify how the preparatory phase established a robust, evidence-informed and practice-informed foundation for the Delphi process.

 

Comments 9

Line 91: The Nature Impact project requires proper citation or website reference for broader programme context.

I recommend adding a table (or supplementary appendix if space is an issue) presenting key themes/findings from stakeholder dialogues (with analysis method specified); synthesis of systematic review findings relevant to the four Delphi domains; explicit traceability mapping (e.g., table format) showing how preparatory sources informed specific Delphi items; decision-making criteria for item inclusion/exclusion.

This transparency would allow readers to assess the foundation of the Delphi study for themselves rather than accepting it was "robust" on assertion alone.

Response 9

We thank the reviewer for this constructive suggestion. We have now added a reference to the Nature Impact project to provide appropriate contextual grounding for the broader programme.

To strengthen transparency and traceability between the preparatory phase and the Delphi process, we have included Supplementary Material (Appendix 1). This appendix specifies how each Delphi item was informed by preparatory sources, including the systematic literature review and stakeholder dialogue meetings. The appendix also clarifies the analytical basis of the stakeholder dialogues and outlines the decision-making criteria used for item inclusion and exclusion.

Together, these additions provide a clearer account of how the preparatory work informed the Delphi questionnaire and address concerns regarding traceability and methodological transparency. For related clarifications, please also see our response to Comment 8.

 

Comments 10

Clarification regarding analytical transparency

Lines 213-225: Whilst Reflexive Thematic Analysis is cited, the analysis lacks transparency and reflexivity:

• Who conducted the initial coding? Was it a single researcher or multiple coders?
• Were themes developed inductively, deductively, or through a hybrid approach?
• How was consensus achieved within the research team?
• Where is the reflexivity? No discussion of researcher positionality, assumptions, or how these may have shaped interpretation
• No illustrative quotes are provided to support identified themes

 

Response 10

Thank you for this detailed and helpful comment. We have revised the Analysis section to improve transparency and reflexivity page 7-8. The manuscript now clarifies that initial coding of open-ended responses was conducted by the first author. Codes were developed abductively as the four predefined domains of the Delphi study: target group, professionals, mechanisms and outcomes, which served as an overarching analytical frame rather than a deductive coding scheme. Data within each domain was analysed inductively.

We have further specified how analytical rigor was supported through iterative discussions within the multidisciplinary research team, where emerging codes and themes were critically examined and refined, rather than through formal intercoder agreement. A reflexivity statement has also been added (page 8) to explicitly acknowledge researcher positionality and how disciplinary backgrounds may have shaped interpretation.

Finally, we clarify how qualitative themes were grounded in the data and, where appropriate, supported by illustrative excerpts in the Results section (page 11-12). We believe these revisions address concerns regarding analytical transparency and reflexivity.

 

Comments 11

The integration of quantitative consensus data with qualitative theme development is methodologically interesting but requires clearer explanation. How did the two analytical streams inform each other?

I recommend adding a reflexivity statement addressing researcher backgrounds and potential biases. Also, include representative quotes (adequately anonymised) to illustrate each theme, and provide a clearer account of the analytical process, potentially with a diagram showing how quantitative and qualitative strands were integrated.¨

Response 11

Thank you for this constructive comment. We have revised the Analysis section to clarify the integration between the quantitative and qualitative analytical streams (page 7-8). The manuscript now specifies that the quantitative analysis was conducted first to identify patterns of consensus and prioritisation, which subsequently informed the qualitative thematic analysis. Open-ended responses were then used to elaborate and nuance areas of agreement and divergence identified in the quantitative results. This revision clarifies how the two analytical approaches were combined in a sequential and complementary manner.

We have now added a reflexivity statement to the Methods section acknowledging the researchers’ disciplinary backgrounds and potential influences on study design and interpretation, as well as the steps taken to support reflexive and rigorous analysis.

Representative quotes have been added to illustrate each theme, in the results section.

 

Clarification regarding Consensus Criteria and Interpretation

Comments 12

Lines 186-190: You have employed dual consensus criteria (mean ≥5.0 AND ≥70% selecting categories 5-7), citing consistency with Delphi methodological conventions. However, Delphi consensus criteria are notoriously variable across studies, and the justification for this specific approach requires strengthening. I recommend more consideration and clarification regarding these points:

• "Near consensus" at 69% (Line 246-249) appears arbitrary. Is 69% meaningfully different from 70%?
• Items with high means but lower percentage agreement (or vice versa) warrant more nuanced discussion
• The shift in prioritisation between rounds (especially "participation in everyday life" dropping from rank 1 to rank 11) deserves deeper theoretical engagement

Response 12

We appreciate the reviewer’s comments regarding the use of dual consensus criteria. We agree that consensus thresholds in Delphi studies are variable and should be interpreted with caution. In the revised manuscript in the methods section, we clarify that consensus is treated as a continuum rather than a binary threshold, and that distinctions such as “near consensus” are used to support nuanced interpretation rather than to imply substantive categorical differences. Items showing divergence between mean scores and percentage agreement are therefore interpreted as reflecting heterogeneity in expert perspectives and are further contextualised through qualitative findings.

 

Comments 13

Lines 276-284: The interpretation that ranking forced reconsideration is plausible, but alternative explanations exist: Respondents may have interpreted "mechanisms of change" differently in a ranking vs. rating context; the distinction between outcomes and mechanisms may have been unclear; forced ranking may have revealed hierarchy rather than changing minds

I recommend that you discuss alternative interpretations and acknowledge the complexity of comparing rating and ranking data. Consider whether the two approaches measure the same construct.

Response 13

Thank you for this insightful comment. While our initial reading suggested that the ranking task may have prompted respondents to reconsider their earlier ratings, we agree that alternative explanations are equally plausible. We have therefore revised the Discussion to explicitly acknowledge the methodological complexity of comparing rating and ranking data.

The revised text now clarifies that Likert-scale ratings and forced-ranking tasks do not necessarily measure the same construct. Ratings allow multiple mechanisms to be endorsed as important in parallel, whereas ranking requires respondents to differentiate between mechanisms even when they are conceptually overlapping. We also address the possibility that some items - such as participation in everyday life - may have been interpreted as outcomes rather than mechanisms, which could have influenced their positioning in the ranking exercise. These revisions are intended to temper interpretation and more accurately reflect the limitations of the ranking procedure.

 

Comments 14

Clarification regarding specification of healthcare resources and treatment context

Line 56: You mention "substantial demands on healthcare resources" but never specify what types of demands: financial capacity, staff capacity, infrastructure, or service provision capacity? This would help understanding how NBHIs might alleviate pressure.

Response 14

We appreciate your comment and have elaborated on a specification of healthcare resources and treatment context in the revised manuscript in the Introduction section.

 

Comments 15

Line 58: "Existing treatment options" remains vague. Which specific treatments are available but inadequate? Cognitive behavioural therapy? Pharmacotherapy? Stepped care models? Understanding what NBHIs complement or replace is essential.

Response 15

Thank you for your insightful thoughts and suggestion to specify the introduction regarding existing treatment options. It is beyond the scope of the paper to discuss existing treatment options; however, we agree to elaborate slightly with an example and have added a sentence regarding ‘difficult-to-treat depression’ in the Introduction section.

 

Comments 16

Lines 76-78: Clarification needed regarding whether NBHIs are envisioned as preventative interventions, alternative treatments, complementary add-ons to traditional care, or rehabilitation tools.

I recommend adding an explicit discussion of how NBHIs fit within existing treatment pathways and what specific gaps they address.

Response 16

We thank the reviewer for raising this important and relevant point. A broader discussion of whether NBHIs should be envisioned as preventive interventions, alternative treatments, complementary add-ons to traditional care, or rehabilitation tools is indeed highly pertinent. However, such conceptual positioning lies beyond the specific scope and aims of the present article, which aims to identify target group, professionals, mechanisms, and outcomes of NBHIs for people with mild to moderate anxiety, depression, and/or stress .

The broader discussion will instead be addressed as part of the overarching Nature Impact project, where the role and positioning of NBHIs across prevention, treatment, and rehabilitation will be examined in greater depth.

 

Justification for pre-selection of nature connectedness

Comments 17

Lines 194-196: Pre-selecting nature connectedness as an outcome introduces bias into the Delphi process. Whilst theoretically justified, this decision removes one outcome from democratic deliberation. It may have influenced respondents' prioritisation of the remaining outcomes and represents a theoretical commitment made before expert consultation

I recommend that you More explicitly acknowledge this as a limitation and justify the decision with reference to the preparatory work.

Response 17

Thank you for this valuable point. We have now expanded the limitations section to explicitly acknowledge the potential bias introduced by pre-selecting nature connectedness and have clarified that this decision was grounded in the preparatory phase, including both the systematic literature review and focus group findings. This revision strengthens transparency around the methodological choice and clarifies its implications for outcome prioritisation.

 

Stakeholder Exclusion

Comments 18

Lines 147-150: The manuscript acknowledges that stakeholder dialogues "did not directly include representatives from volunteer organisations" and that this "may have constrained the diversity of perspectives." This is presented somewhat passively. Given that many NBHIs operate in the voluntary sector, this exclusion is consequential.

Response 18

Thank you for this reflective comment. We have revised the limitations section to acknowledge this more explicitly and to clarify how this may have shaped the conceptual framing toward more formalised, healthcare-embedded models of NBHIs. While this focus was intentional due to the intervention’s planned anchoring within the Danish healthcare system, we recognise that excluding voluntary-sector perspectives may have constrained the diversity of organisational logics and practices considered at an early stage. This limitation is now articulated more critically in the manuscript.

 

Comments 19

Lines 147-150: The decision to anchor within formal healthcare was influenced by stakeholder dialogues that excluded voluntary sector voices—a circular justification.

Could you more critically examine how this exclusion may have biased the findings toward medicalised, professionalised models of NBHIs.

Response 19

Thank you for highlighting this point. We have now expanded the limitations section to explicitly acknowledge how excluding voluntary-sector perspectives may have shaped the findings toward more professionalised models of NBHIs.

 

Mechanisms of Change: Theoretical Integration

Your manuscript identifies three core mechanisms (nature interaction, social community, physical activity)

Comments 20

Lines 347-374: Theme 3 acknowledges interdependence but lacks theoretical synthesis. Could you consider if/how these mechanisms interact; are they sequential, parallel, or synergistic? Do different mechanisms matter more for anxiety vs. depression vs. stress? How do these relate to the theoretical frameworks mentioned (Stress Reduction Theory, Attention Restoration Theory, Lines 66-71)?

I suggest that you engage more explicitly with existing theoretical models.

 Response 20

Thank you for this thoughtful suggestion. We agree that the therapeutic mechanisms identified in the study may vary in salience across anxiety, depression, and stress. But the Delphi process did not elicit diagnostic differentiation from participants, nor was it designed to resolve condition-specific mechanisms. Introducing such distinctions into the manuscript would therefore go beyond the scope of the data and risk overstating the level of granularity supported by the findings.

Instead, we have clarified in the Discussion that the mechanisms should be understood as interdependent, synergistic processes operating across conditions, and we acknowledge that their relevance may differ among individuals. We believe this preserves the integrity of the data while signalling the potential value of future research into diagnosis-specific pathways.

 

Comments 21

Line 356: Delete 'being' (…making strict ranking being inadequate to capture their interdependence)

Response 21

Thank you. The sentence has been revised as suggested.

 

Comments 22

Line 359-362: “Mechanisms like body awareness or mindfulness, while potentially beneficial, were described as distressing for some individuals, particularly those with trauma histories or heightened internal stress.” Without additional context I don't follow the logic here.

Response 22

Thank you for noting the need for further clarity. We have revised this sentence to more explicitly convey what participants reported regarding body awareness and mindfulness practices. Several experts noted that these mechanisms may heighten internal focus in ways that some individuals, particularly those with trauma histories or elevated internal stress, experience as overwhelming. The revised text clarifies this without extending beyond the data.

Reviewer 2 Report

Comments and Suggestions for Authors

This manuscript presents a well-structured Delphi study. It is rich in detail, clearly written, and offers strong theoretical grounding to existing nature–health literature. However, following issues need to be addressed before publications:

- Title is too long.

Abstract:

- There should be at least one sentence related to the background of the study before stating aim. Furthermore, abstract contains too much information about the method and results which should be reduced in terms of word, but not in terms of richness of information.

Introduction: This introduction is well-written.

Methods:

- While the authors described thoroughly the process, how they generated items is not described. Moreover, how they determined the number of initial items? Were these items fully explored what they intended to do?

- Is there any impact of attrition in the results?

Discussion: The first paragraph should be a good summary of the main findings, maximum three core findings. However, this section some repetitions of the results. These repetitions should be avoided.

Conclusion section is also unnecessarily longer.

Author Response

Comments 1

- Title is too long.

Response 1

Thank you for the suggestion to shorten the title.

According to the ‘Instructions for Authors’ section there are no limitations on the length of the title. However, the IJERPH suggests the title to be concise, specific and relevant, and that it should not include abbreviated or short forms of the title, such as a running title or head.

We therefore respectfully propose retaining the originally suggested title.

 

Abstract:

Comments 2

- There should be at least one sentence related to the background of the study before stating aim. Furthermore, abstract contains too much information about the method and results which should be reduced in terms of word, but not in terms of richness of information.

Response 2

Thank you for your comments and suggestions. In the revised version of the manuscript, we have expanded the background to include a sentence before stating the aim of the study. Furthermore, we have shortened the methods and results.

 

Introduction: This introduction is well-written.

Methods:

Comments 3

- While the authors described thoroughly the process, how they generated items is not described. Moreover, how they determined the number of initial items? Were these items fully explored what they intended to do?

- Is there any impact of attrition in the results?

Response 3

We appreciate the question. Regarding item generation, we have clarified in the Methods section how the preparatory phase - comprising a systematic literature review and stakeholder dialogue meetings - directly informed the development and selection of Delphi items. Item content and scope were guided by identified evidence gaps, practice-based themes, and the four predefined Delphi domains (target group, professionals, mechanisms, outcomes). This process is now described more explicitly.

With respect to attrition, the drop from Round 1 to Round 2 involved only two participants (13 to 11), which we consider a relatively small and expected level of attrition in Delphi studies. While it is theoretically possible that these participants held perspectives at the margins of the distribution, we consider the risk of systematic attrition bias to be limited, given the overall diversity of professional backgrounds represented among the remaining respondents.

 

Comments 4

Discussion: The first paragraph should be a good summary of the main findings, maximum three core findings. However, this section some repetitions of the results. These repetitions should be avoided.

Response 4

We are grateful for the constructive suggestion to provide a short summary of the main findings as the first paragraph. We agree and have revised the discussion accordingly.

 

Comments 5

Conclusion section is also unnecessarily longer.

Response 5

Thank you for your thoughtful suggestion to shorten the conclusion. We agree that the Conclusion section was long, and therefore we have shortened this in the revised manuscript.

Reviewer 3 Report

Comments and Suggestions for Authors

With the rise in attention on nature-based health interventions there has been an increase in research about its efficacy with much of the research either focusing on a particular aspect of the interventions or systematic reviews attempting to get a broader picture. This study focuses on four primary areas of the nature-based interventions: the target group, the professionals involved in their delivery, the mechanisms by which the intervention is delivered and the outcomes of the intervention. The study allows for a basic look at the importance of each of these components as determined by health care professionals (all from Denmark?).

As mentioned by the authors in the Discussion section, the study is limited by the homogenous nature of the expert panel and its focus to healthcare settings. It would be helpful if this was more explicit right up front in the abstract - specifically the focus on healthcare settings. 

One major strength of the article is that is was very clearly presented - literature review, methods and results. It has a high level of readability and clear articulated results. This makes the article both relatable and useful. I'm not sure that the outcomes are considerably new to the field but the method of study brings a unique perspective on what is already known - nature-based health interventions can be quite complex.

In terms of specific edits, I don't have a lot but they are as follows...

1. As mentioned above, it would be helpful if it was more explicit from the start that this is predominantly focused on healthcare settings.
2. Results - there is a header on line 300 that is followed by headers on lines 308, 326, 347, & 375 that visually appear as if they are all on the same level but actually the subsequent headers are components under the first one (line 300). It reads awkwardly.
3. In text citations - there are a few places where there seems to be some mixed citation styles popping up - specifically on lines 402, 454, 466, 487 & 514

Author Response

Comments 1

1. As mentioned above, it would be helpful if it was more explicit from the start that this is predominantly focused on healthcare settings.

Response 1

We appreciate this valuable and insightful suggestion. Therefore, we have revised the abstract accordingly to be more explicit from the start that the study is predominantly focused on healthcare settings.

 

Comments 2

1. Results - there is a header on line 300 that is followed by headers on lines 308, 326, 347, & 375 that visually appear as if they are all on the same level but actually the subsequent headers are components under the first one (line 300). It reads awkwardly.

Response 2

Thank you for your relevant comment. In the revised manuscript, we have corrected this, so it is clear for the reader that the first header is different from the following subsequent headers.

 

Comments 3

1. In text citations - there are a few places where there seems to be some mixed citation styles popping up - specifically on lines 402, 454, 466, 487 & 514

Response 3

Thank you for the comment. Unfortunately, we cannot find the suggested mixed citation style on the mentioned lines.