The positive impact of nature on human health and happiness has been the subject of empirical study in various academic fields, with researchers coming to the same conclusion: spending time in natural settings improves people’s emotional and physical well-being (22). Trees, parks, woods, community gardens, grassy verges, and green roofs are all examples of urban green infrastructure that provide a wide variety of experiences to city dwellers (23). Experiencing nature has sound effects on people’s social lives, mental health, and moods and speeds up the body’s healing process (24). People’s attention is renewed, mental tiredness is alleviated, and tension is relieved thanks to nature’s natural eco-environmental factors, which also restore focus (25). Based on this understanding, nature-based therapies encourage individuals to engage in healthy behaviors by engaging in natural activities like gardening, farming, exercise, or interacting with animals (26). Parks’ positive effects on health are conditional on several factors. Larger parks, parks with more green space, and parks with vistas of nature have all been linked to increased happiness among their visitors. It is impossible to overstate how important parks are to human health, and the advantages of visiting parks for health are greatly influenced by the particular park features that visitors like (27). The value of multiple ecological services and functions to people’s health and urban resilience, particularly in the face of public health emergencies, has been demonstrated by recent research on green infrastructure (28). Changes in brain activity in the prefrontal cortex that play a crucial role in emotional regulation have been linked to time spent in nature (29). The increasing expense of chronic diseases has prompted a greater focus on preventative measures, such as prescriptions for spending time in parks that use nature to improve health (30).

The natural and emotional attachment of humans to other living species, defined by biologist Edward O. Wilson, is biophilia. Innate traits are inherited and fundamental to human nature. According to Wilson, biophilia is more accurately described as a “complex of learning norms” refined by thousands of years of evolution and interactions between humans and their environments (31). The process of psychological restoration involves replenishing exhausted resources that have been used up due to adjusting to regular demands, such as stresses and strenuous activities (32). Indoor plants and street trees show how exciting stuff frequently happens in manufactured settings. Allotment (or communal) gardens and urban parks are also created, developed, controlled, and maintained, but they include natural characteristics and provide chances to participate in and follow biological processes. Studies have shown that people may have a genuine “nature experience” even while looking at natural scenes or landscapes from within a structure or a moving vehicle, in a still picture, a moving video, or a virtual reality environment (33).

Social prescribing has been widely used in the geriatric population to combat isolation, raise rates of physical activity, and enhance general health (34). Culturally and scientifically, human relationships with the natural world are beneficial. In addition, participation and focused therapies involving interaction with nature have grown popular due to the growing relevance of social prescribing in healthcare. The use of nature-based social prescriptions by healthcare professionals is a promising new frontier in assisting people of all ages in fostering a sense of belonging in their communities and the natural world. However, healthcare professionals are also facing challenges regarding the effective prescription of nature for health and well-being (35). The psychophysiological stress recovery theory (PSRT) and Kaplan’s attention restoration theory (KART) are the two main schools of thought discussing what constitutes a therapeutic environment. The PSRT, grounded on the biophilia theory, postulates that people have an inbuilt affinity for natural surroundings because of the benefits these settings provided to their ancestors (2). There are advantages to both the individual and the group when people work on improving their health as part of a community. Two separate studies find that social support under social prescription significantly improves both psychological health and the ability to control chronic medical illnesses (36). Numerous demographic, observational, empirical, and intervention studies have shown the positive health effects of spending time in natural settings, including woods, urban parks, local green spaces, and rural parks. Psychological and physiological health improved after exposure to green spaces (37, 38). One such option is “green exercise prescription,” which involves both movement and time spent in the great outdoors. Contact with nature, such as blue and green areas, indoor plants, and gardening, may attenuate or buffer the adverse health effects, reducing or eliminating them (39). Exposure to more green space under nature-based social prescription has been connected with better mood, perceived overall health, and more significant physical activity (40).

Human health depends crucially on the natural world. Pollination of food crops, water purification, flood prevention, and climatic stability are just a few of nature’s many documented health advantages. Clarifying how nature encourages physical exercise for the various mental and physical health advantages is an essential but challenging study area, especially in highly crowded cities with limited and declining access to natural environments (30). Reducing negative emotions and heightening pleasant ones are all part of the psychological, emotional, and cognitive shifts that occur in nature after a stressful event has ended (31). Park-goers’ physical and mental health may benefit from the variety of nature-, movement-, and community-based recreational activities in a semi-natural setting that satisfies their material and immaterial, solitary, and sociable requirements (41). People have known that spending time in natural outdoor settings is good for mental and physical health (42). Parks, canals, woods, and recreational areas are all examples of natural outdoor habitats, often called green and blue environments or green and blue spaces, since they have aspects of nature such as flora and water (43). Green exercise has been linked to better stress recovery, more remarkable attentional restoration, and a decrease in negative emotions, all of which may aid obese persons in adequately coping with these challenges (42, 44). The emerging subject of lifestyle psychiatry is based on the idea that making positive adjustments to one’s daily habits might help treat mental health problems (45). A stronger connection with nature is good for everyone’s health. However, there is an increasing need for nature-based therapies for mental health outcomes. The effect size of NBSPs has not been thoroughly evaluated by systematic reviews and meta-analyses. Furthermore, while social prescribing is increasingly being used as a non-pharmacological intervention for mental health, there is a lack of research specifically focusing on NBSP’s impact (46). The current study addresses these research gaps by synthesizing the literature on NBSPs and comprehensively evaluating their effect size in improving mental health outcomes.

Three independent researchers (MY, MS, and ZS) performed the literature search. Literature searches were carried out through eight databases (PubMed, PsychINFO, Web of Science, Scopus, CNKI, Embase, CINAHL, and Cochrane Library). To find pertinent studies, keywords and medical subject headings (MeSH) were combined. The search terms were used (“nature-based social prescriptio*” AND “nature-based interventio*” AND “green car*” AND “ecotherap*” AND “Ecopsycholog*” AND “green exercis*” AND “outdoor therap*” AND “horticultural therap*” AND “green space*” AND “natural environment*” AND “outdoor environment*” AND “natural settin*”) AND (“mental healt*” AND “mental disorder*” AND “mental illness*” AND “depression*” AND “anxiety*” AND “stress*” AND “well-bein*” AND “mental wellness*” OR “mental health outcomes” AND “psychological healt*” AND “emotional healt*” AND “mood disorder*”).

Researcher’s separately extracted data from the chosen papers using a pre-established form after the articles had been selected. Any differences in opinions between the authors were resolved through group discussions. All investigators were involved in evaluating and discussing the final selection of articles. To ensure the accuracy and integrity of the meta-analysis, the data and references from each included study were thoroughly crosschecked by ZAS to prevent duplication.

The inclusion criteria: (i) the effectiveness of NBSPs for enhancing mental health was examined in studies that were published in English; (ii) studies evaluated the impact of nature-based social prescription, ecotherapy, green therapy, green exercise, and activities performed or related to green spaces or green prescriptions on mental health outcomes; (iii) studies that include a comparison group (control or alternative intervention); (iv) studies that use standardized measures of mental health outcomes such as depression, anxiety, stress, or overall psychological well-being; and (v) studies published between “2017 to 2022” were used.

Exclusion criteria: the studies were excluded if (i) do not report original data (e.g., review articles, editorials, and letters); (ii) do not focus on nature-based interventions for mental health; (iii) do not include a comparison group (control or alternative intervention); (iv) do not use standardized measures of mental health outcomes; and (v) animal assisted-based therapy studies. The titles and abstracts of the initial search results were scrutinized for eligibility. The whole text of the included studies was read to evaluate them further. Discussion and agreement among the reviewers were used to settle any disagreements regarding the eligibility of studies.

The collected data were analyzed by using STATA 16.0. Publication bias, summary measures, and forest plot and meta-regression were performed. In addition, study characteristics were also assessed.

The data analysis evaluates the research outcomes based on the meta-analysis conducted. A sample of 28 articles based on mental health and interventions or social prescriptions was taken for the data collection. The study characteristics of all 28 articles were considered in the meta-analysis (see Table 1). From 2020, 12 articles were taken, six articles were taken from the year 2019, five articles were taken from the year 2021, three were from 2022, one was from 2017, and one was from 2018. Regarding the data collection, nine articles were considered experimental studies, seven articles were considered controlled studies, three of the articles conducted a pilot study, and seven articles were surveyed. At the same time, there were others with repeated measure design and treatment. In the context of the country of research, four articles were conducted in China, four were based in Iran, and five were conducted in the United States, two from Italy, the other three from the United Kingdom, and two from Sweden. In addition, the studies were conducted in South Korea, Canada, Kenya, Australia, Germany, Serbia, and France, and 18 countries study (see Table 1). Regarding gender, 20 studies were male and female, while eight were female. The study characteristics also involved the psychiatric condition of the patients, where most of the studies were based on anxiety, depression, and stress. The interventions included a mixture of natural prescriptions, social prescriptions, and cognitive-behavioral group therapy (CBGT). The studies were also characterized based on the age group, and most were conducted on youngsters.

The I² index calculates the degree of absolute heterogeneity by multiplying by 100 and dividing the discrepancy between the degrees of freedom and the findings of the Q test by the Q value itself. The I² index for the provided data set is indicated to be 95% CI. It shows that the heterogeneity is considerable in the provided data set. The value confirms that the articles used in the meta-analysis are heterogeneous. It shows that the outcomes and models used in the studies are not similar and have substantial differences (see Table 2). It rejects the assumption that the studies are homogenous and that a fixed effect meta-analysis would be a suitable measure to indicate the summary findings. Hence, random effect meta-analysis is considered more suitable for the current study.

Figure 2 indicates the forest plot and the summary measures for the interventions. It shows that there are very few studies that have larger effect sizes. These measures tend to have strong effects, while most studies indicated small effect sizes with the two-point drop. The I² calculation (CI 95%) also shows higher heterogeneity.

Figure 3 indicates the effect sizes of the model for the country. The figure shows that very few studies indicated substantial effect sizes. Most of the studies within the meta-analysis are shown to have effect sizes below two points or four points. The I² calculation (CI 95%) also shows the heterogeneity of the data.

Figure 4 represents the forest plot for the type of data collection. The figure shows that most studies have smaller effect sizes while very few have larger and stronger ones. It shows that most of the studies fall below the point of 2.

Figure 5 shows the forest plot for the studies based on the year of their publication. It shows that most studies have smaller effect sizes falling below two points, whereas very few have effect sizes of less than four points. The I² calculation (CI 95%) also shows higher heterogeneity in the data set.

Figure 6 shows the forest plot for the data set grouped in gender. The studies provided showed that a few of them tend to have larger effect sizes. In contrast, most studies have tiny effect sizes, falling below the two points. The I² calculation (CI 95%) also shows the heterogeneity in the studies considered for the meta-analysis.

There is a high probability that certain studies that should have been included in the study were not included due to the sample bias as the studies with non-significant are not published, which can be assumed in the case of the present study. Hence, it could have affected the accuracy and generalizability of the findings (51). In addition, the exclusion of non-significant results could have affected the precision of the effect size, followed by the absence of smaller and non-significant studies that led to a wider confidence interval and less precise estimates (52). Therefore, it is critical to analyze the impact of publication bias on the study. Various techniques have been used previously, such as funnel plots, Egger’s test, and the trim-and-fill method. To assess the publication bias, funnel plots that seem symmetrical were constructed and presented in Figure 7.

Since the funnel plot seems asymmetrical, it indicates a positive sign of no publication bias in the present study. However, it is also essential to state that outliers on both sides of the plot indicate that certain studies have sample sizes that are too small, and few have a larger sample size, leading to outliers in the data. Meanwhile, these outliers can also be attributed to factors other than publication bias, such as the study’s design and quality. Therefore, these factors could have influenced the results of the meta-analysis, but there is no significant publication bias evident in the study.

Table 2 demonstrates that nature-based social prescriptions have a moderate overall effect size (SMD = 0.2285, 95% CI [0.1690, 0.2880]) on mental health outcomes across both genders. Among females, the impact appears to be weaker (SMD = 0.0583, 95% CI [−0.1011, 0.2178]), while among males, it is moderate (SMD = 0.1987, 95% CI [0.1374, 0.2600]), falling between the effect sizes observed for the combined group and females. These results underscore the gender-specific differences in the effectiveness of nature-based social prescriptions on mental health outcomes. Furthermore, the substantial variability in effect sizes among studies for both genders, as indicated by the heterogeneity statistics, highlights the importance of considering potential moderators or subgroup differences in future research and interventions in this domain.

Table 3 presents the meta-analysis results by year, where it is evident that studies published in 2017 and 2018 did not affect mental health outcomes, as the SMD is not statistically significant and the confidence interval contains zero. However, studies published in 2019, 2020, and 2021 showed a positive effect on mental health outcomes, with SMDs ranging from −0.03 to 0.85. There was also moderate to high heterogeneity among the studies in these subgroups, as indicated by the tau^2 and tau values. The subgroup analysis for 2022 is based on only two studies, which showed no significant effect on mental health outcomes.

Table 4 presents the meta-analysis, including four subgroups based on marital status: married and single, married, single, and married, single, divorced, and widowed. The subgroup with the most significant number of studies (K = 15) is married and single, with an SMD of 0.174 and a 95% CI ranging from −0.3150 to 0.6631. The tau^2 value of 0.8649 and tau value of 0.93 suggest moderate to high heterogeneity across the studies in this subgroup. Meanwhile, referring to the subgroup married, single, divorced, and widowed have (K = 3) and an SMD of 0.2876 and a 95% CI ranging from −0.5374 to 1.1127. The tau^2 value of 0.5008 and tau value of 0.7077 suggest moderate heterogeneity across the studies in this subgroup. The third subgroup, single, has an SMD of 0.831 and a 95% CI ranging from −0.4733 to 2.1354; this subgroup only includes four studies. In addition, the tau^2 value of 1.7093 and tau value of 1.3074 suggest high heterogeneity across the studies. Lastly, the subgroup with the lowest heterogeneity is married to an SMD of 0.3822 and a 95% CI ranging from −0.0209 to 0.7853. The tau^2 value of 0.1725 and tau value of 0.4153 suggest low heterogeneity across the five studies in this subgroup.

Table 5 presents the meta-analysis results of subgroups by data collection. The survey subgroup includes seven studies and reports a small positive effect size (SMD = 0.1931) with a wide confidence interval including negative and positive effects and zero. The subgroup (data collection) has a high degree of heterogeneity, as indicated by a significant tau^2 value and a tau value of 1.2357. Meanwhile, the experimental study subgroup includes nine studies. It reports a moderate positive effect size (SMD = 0.4726) with a confidence interval that does include zero. In the third group subgroup treatment, only one study reports a moderate positive effect size (SMD = 0.6041). Therefore, based on the above results, it is evident that only the treatment group shows a statistically significant influence of the intervention on the participants’ mental health, where the intervention of green and blue nature-based social prescription may have a positive effect on mental health outcomes.

Table 6 presents the data analysis of meta-analysis by country, where the 18-countries subgroup is a standalone study that reports a small to moderate effect size (SMD = 0.3474) with a 95% confidence interval ranging from 0.2254 to 0.4694; France has SMD 0.836 with a 95% confidence interval 0.6043 to 1.0696; Canada (Ontario) has SMD 0.6041 with a 95% confidence interval 0.1663 to 1.0419, indicating that a statistically significant impact of interventions of green and blue nature-based social prescription on mental health outcomes in those countries. It also implies that intervention positively affects mental health in that studies included those studies. However, referring to other groups such as Italy, Iran, South Korea, the United States, the United Kingdom (London), Sweden, China, the United States (Chicago), the United Kingdom, Germany, and Serbia have shown moderate to higher size effects. Therefore, it can be determined that green and blue nature-based social prescriptions could be associated with improved mental health outcomes within 18 countries, including Ontario and France. Still, the strength of the association varies across different subgroups and countries. However, no other country was found to have a statistically significant association.

According to the Table 7 meta-analysis of interventions statistics, CBT (SMD −0.0035; 95% CI: [−0.5090; 0.5020]; Tau^2: 0.1011; Tau: 0.318), digital intervention (SMD −0.3654; 95% CI: [−0.5258; 1.2566]; Tau^2: 0.2976, Tau: 0.5455), music intervention (SMD −2.1281; 95% CI: [−0.4659; 4.7221]; Tau^2: 3.4046; Tau:1.8452), and psychological interventions (SMD −0.8529; 95% CI: [0.3051; 1.4007]; Tau^2: 0.1224; Tau: 0.3499) do not significantly impact. The other interventions (social belongingness, communication training, blue intervention, nature-based education, cognitive behavior group therapy, social prescribing coordinator, self-help intervention, participatory, organizational intervention, inpatient services, brief diet, internet-based intervention, prenatal intervention, yoga and meditation, ergonomics training program, yoga nidra intervention, and storytelling) highlighted above are significant.

Table 8 shows the meta-analysis of the mental health condition. It shows that anxiety (SMD: 0.3955; 95% CI: [0.0901; 0.7010]; Tau^2: 0.1906; Tau: 0.4365), quality of life (SMD: 1.2846; 95% CI: [−1.3431; 3.9123]; Tau^2: 3.5679; Tau: 1.8889), and depression (SMD: 0.8407; 95% CI: [−0.4988; 2.1803]; Tau^2: 2.2055; Tau: 1.4851) does not have a significant impact while anxiety and depression (SMD: 0.3474; 95% CI: [0.2254; 0.4694]) anxiety, stress, and depression(SMD: 0.0008; 95% CI: [−0.2963; 0.2980]), Alzheimer (SMD: 0.0825; 95% CI: [−0.7075; 0.8726]), depression and anxiety (SMD: 0; 95% CI: [−0.2079; 0.2079]), burnout (SMD: 0.0171; 95% CI: [−0.4657; 0.5000]), insomnia (SMD: −0.2417; 95% CI: [−0.5333; 0.0500]), and mental well-being (SMD: −1.6043; 95% CI: [−1.9222; −1.2865]) have significant impacts.