Effect of nature-based health interventions for individuals diagnosed with anxiety, depression and/or experiencing stress—a systematic review and meta-analysis

Abstract

Objectives

The use of natural environments and nature activities as elements in the treatment and rehabilitation of mental health challenges is gaining international attention. The objective of the present review was to summarise the knowledge on the effects of nature-based health interventions (NBHIs) targeting individuals diagnosed with anxiety, depression and/or experiencing stress.

Design

Systematic review and meta-analyses. The quality and certainty of evidence were assessed using the SIGN and GRADE.

Data sources

Searches were performed in Embase, MEDLINE, PsycINFO, CINAHL, Cochrane and Web of Science.

Eligibility criteria

(1) NBHIs, (2) Individuals with a diagnosis of mild to moderate anxiety, depression and/or experiencing stress, (3) Age of participating individuals: 18–84 years, (4) Study designs: randomised controlled trials, cohort studies, case-control studies and case-series studies and (5) Publication date: 2000–2024.

Data extraction and synthesis

Screening, quality appraisal and certainty of evidence, assessed using SIGN and GRADE, were performed by two independent reviewers, except title screening. Meta-analyses were performed using random-effect models.

Results

Nineteen articles were included, of which 14 were included in the meta-analyses. The articles showed substantial variation in design, interventions, settings and risk of bias, limiting the certainty of evidence according to GRADE. Participating in NBHIs led to a small to large effect in mental health with standardised mean changes of −0.80 (95% CI= (−1.56; −0.04)), −0.87 (95% CI= (−1.18; −0.56)), −0.32 (95% CI= (−0.74; 0.09)) and 0.58 (95% CI= (0.39; 0.77)) for anxiety, depression and stress scores and overall mental health scores, respectively.

Conclusions

This is the first systematic review examining the effect of NBHIs exclusively on individuals diagnosed with anxiety, depression and/or experiencing stress. Our findings suggest small to large improvements after participating in NBHIs. However, methodological limitations to the included articles necessitate cautious interpretation.

PROSPERO registration number

CRD42024516270.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• A rigorous quality appraisal of bias and evidence of confidence was conducted for transparency, using SIGN and GRADE, respectively.

• The systematic review employed a structured approach across six databases, using tabulation for data organisation and narrative analysis for contextual insights, with a rigorous appraisal of bias and evidence confidence for transparency.

• A limitation was the initial title screening performed by one reviewer, though reliability was improved with two reviewers handling abstract screening, full-text reviews, data extraction and quality appraisal.

• The review identified 45 diverse outcome measures, reflecting the complex ways nature influences human health, with a focus on quantitative measures related to mental health, including patient-reported outcome measures.

• Excluding qualitative articles was a limitation, as their inclusion could have enriched the analysis by offering deeper insights into context, processes and subjective experiences.

Introduction

The prevalence and burden of mental health disorders, particularly anxiety, depression and stress, have risen significantly worldwide.^{1 2} Consequently, these conditions are among the most substantial contributors to global disability and health loss.¹ This burden reveals a global shortage of adequate resources to address a worldwide challenge in a sustainable manner, highlighting the need for improved structural mental health strategies.

Mental health challenges cannot be addressed solely through clinical treatment or pharmacotherapy, as these approaches often put less focus on the social and environmental factors that contribute to mental well-being.^{3 4} Furthermore, up to 30% of individuals treated for depression experience treatment-resistant depression, calling for alternative approaches.^{5 6} Emerging evidence suggests that exposure to natural environments and engagement in nature-based activities can serve as a valuable adjunct in treatment and rehabilitation for mental health disorders.^{7 8} Nature-based health interventions (NBHIs), such as green exercise, nature walks, therapeutic horticulture and water activities like sea swimming, are increasingly recognised for their ability to reduce psychological distress, improve mood and foster a sense of relatedness to nature, thereby enhancing mental resilience and well-being.9,12

Recent systematic reviews highlight the positive effects of nature exposure on mental health and well-being. However, previous reviews included healthy student populations without diagnosed anxiety or depression. One review reported significant improvements in physical activity and mental well-being through ‘nature prescriptions’.¹³ Another study found that NBHIs significantly enhanced well-being.¹⁴ In contrast, a third review concluded that evidence supporting nature-based therapies for anxiety, depression and stress remains limited.¹⁵ However, the effect of NBHIs found in these reviews may be diluted by including healthy individuals with limited symptoms. In contrast to these limitations, the current systematic review will only include studies on individuals diagnosed with anxiety, depression and/or experiencing stress.

Objective

The objective of this systematic review and meta-analysis was to identify, assess and summarise the available knowledge on the effect of NBHIs in individuals with a diagnosis of mild to moderate anxiety and/or depression and/or experiencing mild to severe stress.

Study selection and analysis

This systematic review followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses 2020 Checklist and was registered in PROSPERO before initiation (CRD42024516270).

Search strategy and inclusion criteria

A search was conducted on 2 January 2024 in six databases: Embase, MEDLINE, PsycINFO, CINAHL, Cochrane and Web of Science. We included articles focusing on (1) NBHIs, (2) Individuals with a diagnosis of mild to moderate anxiety, depression and/or experiencing stress, (3) Age of participating individuals: 18–84 years, (4) Study designs: randomised controlled trials (RCTs), cohort, case-control and case-series studies and (5) Publication date: 2000–2024. The full list of inclusion and exclusion criteria and search string developed using the PICOS approach is detailed in the online supplemental file S1.

Initial title screening was performed by one reviewer (NHJ, CL or DVP) using the systematic review tool Covidence. Abstract screening, full-text reviews, data extraction and quality appraisal were conducted independently by two reviewers (NHJ and CL).¹⁶ Conflicts were resolved through discussion, with a third reviewer (DVP) involved in cases of disagreement.

Intervention and target group

NBHIs were defined as structured health interventions with a clearly stated treatment goal. The NBHIs should take place in natural environments, facilitated under the guidance of trained healthcare professionals, and are required to take place as part of a programme over multiple sessions.

The target group comprised individuals diagnosed with mild to moderate anxiety and/or depression, and/or experiencing mild to severe stress. Details on the specific definitions can be found in the online supplemental file S1. Articles both with and without a control group were included. The control group could consist of individuals who either did not undergo any intervention (passive controls) or of individuals who participated in a different type of intervention than NBHI (active controls).

Outcome measures

Any mental health-related outcome measure, including symptom score scales related to anxiety, depression and/or stress, well-being and/or quality of life (referred to as ‘overall mental health’), nature connectedness, sick leave/return to work and health consumption and/or sleep quality, were included. The included scales had to be validated and administered at least preintervention and postintervention.

Data extraction and risk of bias assessment

Data was extracted by two reviewers (NHJ and CL), and data incorporated in the meta-analysis were verified for accuracy by a third reviewer (SDS).

To evaluate the methodological quality, the SIGN (Scottish Intercollegiate Guidelines Network) checklists were employed to assess the risk of bias in cohort studies and RCTs.¹⁷ Each study was reviewed independently by two researchers (NHJ and CL), with discrepancies resolved through discussion with a third reviewer (DVP). Based on predefined criteria developed by the authors, a summary score was derived, and the results from the risk of bias assessments are presented in tables and figures and used in a Grading of Recommendations Assessment, Development and Evaluation (GRADE) evaluation.¹⁸ This GRADE evaluation was used to rate the overall certainty of evidence for relevant outcomes. This combined approach ensured a robust and transparent evaluation of the evidence base.

Data synthesis

An overall synthesis was conducted using tabulation and narrative analysis to comprehensively describe the characteristics of the included articles.

Two meta-analyses were initially intended for each outcome category (anxiety, depression, stress and overall mental health). One meta-analysis was planned to compare the effects of intervention versus standard care solely within RCTs. The second analysis aimed to combine cohort studies without a control group with only the intervention arm from RCTs, thereby estimating the participation effect independently of a control group. Only meta-analyses with data from at least three articles were included. Given that no outcome category included more than two RCTs, only meta-analyses combining single-arm cohorts with the intervention arms from RCTs were performed.

The standardised mean change using raw score standardisation was computed, defined as the difference between mean postintervention and preintervention scores, divided by the SD of the preintervention score. If a study reported multiple postintervention scores, the score measured closest to the end of the intervention was used to compute the mean change. Studies that did not report means, SD or standard errors either before or after the intervention were excluded from the meta-analyses. Reported standard errors were converted to SD. All means, SD and standard errors were initially extracted by author SDS and subsequently reviewed by author CL.

Statistical analysis

R V.4.4.1 was used for all statistical analyses.¹⁹ The effect sizes were computed using the metafor package V.4.6.0.^{20 21} Random-effects models were fitted independently for articles reporting on anxiety, depression and stress outcome measures, as well as overall mental health outcome measures. The study ID was used as a random effect. Small study bias was assessed using Egger’s regression test and visualised through funnel plots. Heterogeneity was assessed using Cochran’s Q-test and the I² statistic. Effect sizes between 0.2 and 0.5 were categorised as small, those between 0.5 and 0.8 as moderate, and those above 0.8 as large.

Trial registration

The study was registered in PROSPERO (number: CRD42024516270). No study protocol was registered.

Patient and public involvement

Patients and/or the public were not involved in the design and conduct of the research or the writing process of the manuscript.

Findings

In total, 8602 articles were identified from the databases, figure 1. Of these, 3151 were duplicates. We identified 10 relevant articles through citation searching. After screening, we included 19 original articles, table 1.

Figure 1. Flow diagram of included articles. *Initial screening was undertaken by one person: https://training.cochrane.org/handbook/current/chapter-04/section-4-6 (afsnit 4.6.4). **Second screening was undertaken by two people working independently to determine whether each article met the eligibility criteria. ***Articles may be excluded for several reasons. Total number, therefore, exceeds 59.

Table 1. List of included articles (n=19) (SE online supplemental file S3 for characteristics).

Author	Year	Country	Target group	Study quality
Burlingham et al30	2022	England	Anxiety, depression	Low
Corazon et al31	2018	Denmark	Stress	Low
Djernis et al22	2021	Denmark	Stress	Moderate
Grahn et al32	2017	Sweden	Depression, stress	Low
Hyvönen et al23	2023	Finland	Depression	Moderate
Keenan et al24	2021	England	Anxiety, depression	Low
Kim et al33	2009	Korea	Depression	Low
Korpela et al34	2016	Finland	Depression	Low
Maund et al35	2019	England	Anxiety, depression	Low
Paakkolanvaara et al36	2023	Finland	Depression	Low
Sahlin et al37	2015	Sweden	Anxiety, depression, stress	Low
Stigsdotter et al25	2018	Denmark	Stress	Moderate
Van den Berg et al38	2021	The Netherlands	Stress	Low
Vujcic et al26	2017	Serbia	Anxiety, depression, stress	Low
Vujcic Trkulja et al27	2021	Serbia	Anxiety, depression, stress	Low
Wahrborg et al39	2014	Sweden	Depression, stress	Low
Walter et al28	2023	USA	Depression	Low
Willert et al40	2014	Denmark	Stress	Low
Yeon et al29	2022	Chorea	Depression	Low

Description of included articles

Among the 19 articles, eight were RCTs22,29 and 11 were cohort studies,30,40 table 1. A total of 1901 participants were included. Of these, 1013 participants were included as controls (678 matched controls in one article).³⁹ Most of the participants were females, with a variation between studies from 50–96%.

In six of the articles, the target group was mixed and included both anxiety and/or depression (n=3),^{24 30 35} or anxiety, depression and/or stress (n=3).^{26 27 37} None of the articles had individuals with anxiety as a single target group.

Fourteen articles targeted individuals with depression; in six of these articles, the target group was solely participants with depression,^{23 28 29 33 34 36} and in the other eight studies, the target group was mixed and included both anxiety and/or depression (n=3)^{24 30 35} or depression and/or stress (n=2)^{32 39} or anxiety, depression and/or stress (n=3).^{26 27 37}

Ten articles targeted individuals with stress-related symptoms; in half of these articles, the target group was mixed and included both depression and/or stress (n=2)^{32 39} or anxiety, depression and/or stress (n=3),^{26 27 37} and in the other half, the target group was stress (n=5).^{22 25 31 38 40}

In 18 articles, green spaces were part of the intervention (online supplemental file S2). The green spaces included therapy gardens (n=6),^{22 25 29 31 32 39} green areas like urban parks, forests, lakesides and other natural settings (n=9),2324 28 34,38 40 and botanical gardens (n=3).^{26 27 33} Blue spaces were part of the intervention in two articles.^{28 30}

The interventions lasted from 1 to 24 weeks and consisted of a wide range of sessions in each week (0–5 sessions) and in total (5–96 sessions). In the NBHIs that lasted for 1 week (n=3),^{22 24 40} the intervention included sessions on five consecutive days. A full set of the included raw data can be found in online supplemental file 2.

Outcome measures

Forty-five different outcome measures were found across the 19 articles. These were grouped into four categories (symptom scales, overall mental health scales, nature and therapeutic scales and other health outcomes) (online supplemental file S4).

Symptom scales

Anxiety-related outcome measures

Two different outcome measures from three articles^{30 35 37} showed a large improvement of anxiety after participation in an NHBI (standardised mean change= −0.8, 95% CI= (−1.56; −0.04), p=0.039), figure 2. There was evidence of large heterogeneity (I²=84.59%, p<0.001), but no evidence of small-study bias (Egger’s test p=0.855) (online supplemental file S5).

Figure 2. Meta-analysis of the effect of nature-based health interventions on symptoms of anxiety.

In one RCT, NHBI didn’t appear superior to the control group with regards to anxiety.²⁶

Depression-related outcome measures

Five different outcome measures from six articles2328,30 34 37 showed a large decrease in depression symptoms after participating in an NHBI (standardised mean change= −0.87, 95% CI= (−1.18; −0.56)), figure 3. There was evidence of large heterogeneity (I²=73.53%, p=0.002), but no evidence of small study (Egger’s test p=0.507) (online supplemental file S5).

Figure 3. Meta-analysis of the effect of nature-based health interventions on symptoms of depression.

Two out of four RCTs reported superior effects of an NBHI compared with controls,^{29 33} while the remaining two showed no statistically significant differences.^{23 26}

Stress-related outcome measures

One outcome measure from three articles^{22 35 40} showed a small, non-significant improvement in perceived stress after participating in an NBHI (standardised mean change= −0.32, 95% CI= (−0.74; 0.09)), figure 4. There was no evidence of heterogeneity (I²=41.30%, p=0.196) or small-study bias (Egger’s test p=0.121) (online supplemental file S5).

Figure 4. Meta-analysis of the effect of nature-based health interventions on symptoms of stress.

One out of three RCTs indicated a superior effect on perceived stress of an NBHI compared with the control group,²⁶ while the other two found no difference between participants in the NBHI and the control group.^{22 40}

Overall mental health scales

Overall mental health was assessed through 18 different outcome measures (online supplemental file S3).

Six different outcome measures from seven articles23,2534 35 37 38 showed a moderate improvement in overall mental health after participation in an NBHI (standardised mean change= 0.58, 95% CI= (0.39; 0.77)), figure 5. There was no evidence of heterogeneity (I²=0.00%) or small-study bias (Egger’s test p=0.575) (online supplemental file S5).

Figure 5. Meta-analysis of the effect of nature-based health interventions on overall mental health.

Three of five RCTs reported superior effects of NHBIs compared with controls,^{23 24 29} while two showed no significant differences.^{25 38}

Nature and therapeutic scales

Among the articles examining nature and therapeutic scales, two articles examined the Connectedness to Nature Scale (CNS)^{22 24} (online supplemental file S4). Keenan et al pointed out results indicating an increase in CNS at post and follow-up in the intervention group compared with a control group.²⁴ Djernis et al found that CNS for mindfulness in a natural outdoor setting significantly exceeded the control, at follow-up.²²

Other health outcomes

Among the articles examining other health outcomes, two articles focused on healthcare consumption and sick leave status (online supplemental file S4).^{31 39} Corazon et al found no difference between nature-based therapy and validated cognitive behavioural therapy in the effect on number of contacts with a general practitioner or long-term sick leave after treatment, although both programmes were efficacious.³¹ Währborg et al investigated the effect of a nature-assisted rehabilitation programme in a group of patients with stress and/or depression and found no difference in effect on sick leave status and healthcare consumption when compared with a matched population-based reference cohort (treatment as usual).³⁹

Study quality

The quality of the articles ranged from low to moderate, with most articles (83%) rated as low quality and 17% rated as moderate quality (online supplemental file S6 and S7). For the RCTs, the primary reason for downgrading quality was the lack of blinding. Also, insufficient or unclear reporting of concealment methods, the presence of differences in treatments under investigation and whether subjects were analysed in the groups to which they were randomly allocated were some of the noticeable potential sources of bias. The primary limitations of the multi-cohort studies included missing or unclear information regarding participation rate and unclear information about full comparisons between groups. Additionally, the absence of blinding raised concerns about its potential impact on the self-reported psychometric outcomes. For single-cohort studies, the two main sources of bias relate to potential confounding, which impacted most studies to a high degree, and the knowledge of exposure status could have influenced the assessment of outcome.

GRADE assessments

A prominent challenge in many of the included articles was unclear or missing descriptions of the study methods (assessed as unclear risk of bias). Risk of bias led to downgrading the certainty of evidence across all study types, particularly cohort studies. Additionally, substantial heterogeneity in study design was observed among the RCTs, reflecting inherent variation in results caused by underlying differences between the included populations and interventions, necessitating a further downgrade in the certainty of evidence. In addition, marked imprecision was noted in the RCTs. When the CI was overly wide or the total information size was insufficient, the certainty of evidence was downgraded. In this analysis, the evidence was downgraded due to imprecision. All GRADE evaluations rated the certainty of evidence as very low, indicating that the true effect could differ substantially from the stated effect estimates (online supplemental file S8).

Discussion

This is the first systematic review and meta-analysis on the effect of NBHIs for individuals diagnosed with anxiety, depression and/or experiencing stress. Across 19 included articles, we found large reductions in anxiety and depression symptoms, a small non-significant reduction in stress symptoms and a moderate improvement in overall mental health after participating in an NBHI. However, methodological limitations and substantial heterogeneity resulted in very low certainty of evidence that the improvement could be causally attributed to the NBHIs.

Several prior quantitative systematic reviews on NBHIs have been conducted, all of them including healthy target groups or a mix of mental health target groups and chronic physical diseases like cardiovascular disease and cancer.915 41,43 Paredes-Céspedes et al assessed the effects of nature exposure therapies on stress, depression and anxiety levels through a systematic review of eight RCTs.¹⁵ While some evidence indicated significant reductions in stress, no significant effects were observed in anxiety and depression outcomes. However, in contrast to the present systematic review, they were unable to directly evaluate the effects of NBHIs on the specific target group, as the study encompassed both healthy and unhealthy populations.¹⁵ Likewise, Coventry et al, who investigated nature-based outdoor activities for mental and physical health in adults, also included participants with or without mental and/or physical health problems. In a meta-analysis of RCTs, they showed that NBHIs may be effective for improving depressive mood −0.64 (95% CI: 1.05 to −0.23) and reducing anxiety −0.94 (95% CI: 0.94 to −0.01).⁹ Based on the inclusion of both healthy and unhealthy populations of these reviews, it remains unclear whether the interventions might have provided greater benefits to these distinct subgroups compared with the target groups included in the present systematic review.

In contrast, Picton et al investigated experiences of outdoor nature-based therapeutic recreation programmes for persons with a mental illness through a qualitative systematic review.^{44 45} They highlighted the positive perceptions of outdoor nature-based therapeutic recreation among individuals with mental illness, emphasising its benefits for mental health, social inclusion and psychological well-being.⁴⁴

Other systematic reviews have often included studies with interventions such as ‘shinrin-yoku’ (forest bathing), where single-session interventions are common.⁴³ Instead, we only included articles with interventions incorporating multiple sessions led by a healthcare provider. This ensured that participants were exposed to nature repeatedly, which is likely to enhance the long-term impact of mental health interventions.⁴⁶

As in the present systematic review, all prior systematic reviews suggest that current evidence provides limited support for NBHIs as a primary intervention for mental health challenges based on quality assessment.

Strengths and limitations

This systematic review, performed across six different databases, used a structured approach. Tabulation organised data for easy comparison, while narrative analysis provided contextual insights. Furthermore, the rigorous appraisal of the risk of bias, together with the overall assessment of the confidence in the evidence, heightened the transparency of the systematic approach.

A limitation to the selection of articles was that the initial title screening was performed by one reviewer, as multiple reviewers could have enhanced the reliability and rigour of the selection process. However, it strengthened the forward process that two reviewers performed the abstract screening, full-text reviews, data extraction and quality appraisal.

The review identified 45 distinct outcome measures. This diversity may reflect the multifaceted and not yet fully understood ways in which nature influences human health. We opted to include any quantitative outcome measure related to mental health, recognising that this broad approach would encompass a range of patient-reported outcome measures (PROMs). This decision was made with the understanding that PROMs may demonstrate limited correlation with physiological indicators and other objective outcomes.⁴⁷ Still, these outcome measures are often the ones most valued by patients and their families.⁴⁸ However, the systematic approach constrained the number of eligible articles, leading to the inclusion of only a limited number of articles with many different PROMs in the meta-analyses.

An important limitation was the considerable heterogeneity in the meta-analyses for anxiety, depression and stress. Given the small number of studies included in the meta-analyses, it was not possible to quantify the sources of this heterogeneity, but at least for the analysis of depression, there appears to be some degree of clustering based on the different outcomes measured, which may explain some of this heterogeneity. Likewise, the heterogeneity could also stem from the wide range of psychological conditions and inclusion criteria among study populations, as well as from substantial differences in the interventions themselves—ranging from nature-based activities to approaches more focused on cognitive therapy in a variety of different natural environments. As a consequence, the findings of the meta-analyses should be interpreted cautiously.

A further limitation of the present review was the exclusion of qualitative articles, as their inclusion could have provided valuable insights and a deeper understanding of the context, processes and subjective experiences. This integration could have enriched the overall analysis and enhanced the comprehensiveness of the conclusions.

Implications and future research

While the integration of natural environments and activities into mental health interventions like NBHIs is an emerging and promising field, further research is needed to optimise interventions and establish clear guidelines for mental health conditions. Rigorous, well-designed studies will be essential to inform ‘best practices’ for NBHIs, ultimately supporting a more holistic and effective approach for individuals diagnosed with anxiety, depression and/or experiencing stress.

Although the mechanisms behind the therapeutic effects of NBHIs are hypothesised to involve reductions in rumination, increases in mood and ability to be present in the moment, lower levels of cortisol and reduced symptoms of anxiety, depression and stress, more research is needed to confirm these pathways, especially in clinical populations.^{49 50} Furthermore, the impact of specific types of nature settings, as well as the therapeutic adjustment of the activities to achieve the most effective NBHIs, is warranted. This research gap is especially pressing for individuals diagnosed with mild to moderate anxiety, depression and/or experiencing stress, as targeted interventions could offer a sustainable and cost-effective accessible complement to traditional mental healthcare approaches and may help decrease healthcare consumption and avoid sick leave.

Furthermore, a pertinent question is how to evaluate the quality of studies involving complex interventions such as NBHIs. Complex interventions comprise multiple interacting components, which may influence their design and implementation, potentially leading to lower quality ratings. For instance, conducting a blinded study – one of the standard criteria for higher methodological rigour—may not be feasible in such contexts. Over the past decade, there has been a substantial increase in research focused on NBHIs. Despite this growth, many reviews and individual studies consistently conclude that the quality of the existing research remains low. One possible explanation is that the field is still in its exploratory phase and, as a result, has yet to adopt the rigorous methodological standards typically applied in medical research. NBHIs, however, show promise as adjunctive strategies for individuals diagnosed with anxiety, depression and/or experiencing stress.

Conclusion and clinical implications

Collectively, this systematic review and meta-analysis suggest that participating in an NBHI may have positive effects on individuals diagnosed with anxiety, depression and/or experiencing stress as well as on overall mental health. However, the methodological quality of the present literature does not allow for conclusively confirming or denying that NBHIs should be included in evidence-based treatments of individuals diagnosed with anxiety, depression and/or experiencing stress.

Still, NBHIs hold significant potential as complementary approaches to enhance mental health. While not intended to replace conventional therapeutic methods, NBHIs may serve as valuable adjuncts, addressing gaps in traditional treatment paradigms and offering innovative solutions for individuals suffering from anxiety, depression and/or stress. By integrating NBHIs into existing mental health interventions, practitioners may better address the multifaceted nature of mental health challenges, promoting resilience, improving accessibility and fostering holistic well-being. Further research is essential to fully understand NBHI’s mechanisms of change and efficacy, optimise their implementation and explore their role within broader mental health systems.

Data availability statement

All data relevant to the study are included in the article or uploaded as supplementary information.