The impacts of Payment for Ecosystem Services programmes on health and wellbeing: a systematic review protocol

Background / Rationale

Ecosystems and human health are intrinsically interlinked¹. Nature provides services and goods that are fundamental for health, such as food, clean air and water, as well as more complex services that ensure longer-term population health, like preventing floods, land degradation, collectively referred to as “ecosystem services (ES)”^1–3. Conversely, environmental degradation can have myriad diverse effects on human health, increasing the risk of food insecurity, acute and chronic diseases, and harm the mental wellbeing of populations^1,4–6.

From an economics perspective, the consequences of environmental deterioration reflect not only an overuse and exploitation of natural resources, but also a historical failure to properly put an economic value on the positive externalities that people are receiving from ES^1,7. In response, policymakers and governments have increasingly been promoting market-based policies, programmes, and plans to offset human actions that have a negative impact on nature and address environmental degradation^8,9, putting a monetary value to ES that were previously unpriced, and, in doing so, valuing the benefits derived from ES^7,10. Among these interventions, Payment for Ecosystem Services (PES) programmes have gained popularity, and are being implemented worldwide for their potential to improve environmental conservation, enhance sustainable land use practices, and contribute to poverty alleviation^11–15, with an estimated $36-42 billion spent on them in 2018^16,17.

While various definitions of PES exist, we will adopt Wunder’s 2015 definition¹⁴: “voluntarily transactions, between service users and service providers that are conditional on agreed rules of natural resource management for generating offsite services”¹⁴. In practice, PES often involves transferring money from those who ultimately benefit from ecosystem services (buyers/users) to the people who maintain the provision of these services (PES providers/sellers/participants). PES programmes can be theorised as complex interventions¹⁸ and their effects are likely to vary depending on the characteristics of the PES programme itself, the context in which it is implemented, and which people receive it^19–22.

The main components and characteristics of PES are related to: type of ecosystem service that the programme wants to conserve, payment structures, financing schemes, contract length, whom to pay (individuals, communities), type of payment (cash or in-kind), conditions for receiving the payment, amount of the payment, timing of payments, scale of the PES programme (local, national, regional), actors involved in the programme (public or private sector), type of compliance monitoring, and types of sanctions for non-compliance^{10,14,19,23–27}. Additionally, the source of funding for programmes should be considered which could involve public money or other sources such as carbon offsets and similar schemes^28,29.

From a purely environmental perspective, PES programmes seem promising as they may have improved forest cover and reduced deforestation rates, although the quality of evidence is low³⁰. However, the health implications of PES remain unclear. While improving ecosystems is itself likely to be beneficial for longer-term human health, it is possible that the health of vulnerable providers of ecosystem services could be adversely impacted. For example, Blundo-Canto et al. discuss in their review, focused on the impacts of PES programmes on livelihoods, that cases in which access rules and land-use restrictions led to segregation between powerful groups and less influential farmers, consequently decreasing community integration³⁰. Le et al. also discuss that PES programmes may also aggravate conflicts (especially when marked social disparities already exist) and highlight that land use changes due to deforestation and poorly defined land use rights have resulted in precarious and vulnerable livelihoods for people that rely on forests³¹.

Furthermore, evidence suggests that the impacts of PES programmes on health and wellbeing may vary by population subgroup, including socioeconomic position, disabilities, gender, level of education, ethnicity, and governance structure of the communities), as well as the characteristics of the payment schemes such as transaction costs, types of schemes, and payment mechanisms^{10,20,25,32,33}. For instance, Brownson et al. showed that PES providers who advocate for reforestation, instead of only receiving cash payments, experience improved health and wellbeing³⁴. A meta-analysis of PES in developing countries indicated that PES programmes have a positive impact on livelihoods (total income, total consumption expenditure and/or gross value of household assets), especially when PES payments are given partly in the form of communal infrastructure, the amount of payment is higher (cash or in-kind) or the transaction costs are lower³⁵. Hejnowicz et al. highlight the importance of how the programmes are constructed and design to achieve effectiveness, efficiency, and equity trade-offs³⁶.

No previous systematic reviews of PES have assessed impacts on health or subjective wellbeing in a comprehensive manner, with existing reviews limited to specific population groups or health outcomes, such as children with malaria and diarrhoea disease^30,37,38. In this review we will assess health (encompassing physical and mental health) and subjective wellbeing^39,40, with outcomes including but not limited to happiness, life satisfaction, quality of life, respiratory diseases, health-related behaviours, depression, anxiety, suicide, maternal and child mortality/morbidity, malnutrition, among others^12,34,41.

The aim of this systematic review is therefore to assess the impacts of PES programmes on health and subjective wellbeing of both PES participants (individuals and communities) who are incentivised to protect their environments and on the wider community providing the ecosystem. We anticipate this evidence will help understand the potential implications of market-based policies of this type.

Logic model

Logic models can increase the transparency and utility of systematic reviews, highlighting aspects of complex problems that might otherwise be overlooked^42–44. The following model was developed based on scoping the relevant literature and feedback from community leaders that are participating in a Payment for Ecosystem programme in Ecuador (Socio Bosque and Socio Manglar), to illustrate the hypothesised causal relationships between PES programmes and health and subjective wellbeing outcomes. While it is well-documented that PES programmes primarily affect environmental outcomes through changes in environment-related behaviours^30,37, our logic model focuses on the effects of this programme on health and subjective wellbeing. We plan to refine the logic model in light of the findings from the systematic review.

As illustrated in Figure 1, PES programmes are introduced with the primary goal of environmental conservation. However, a number of intermediate outcomes and effects on health and subjective wellbeing are likely to arise due to mechanisms related to increase economic material resources and positive environmental externalities. We hypothesise two groups of mechanisms through which PES programmes can affect health and subjective wellbeing: the behavioural mechanism, material resource mechanism, and the environmental mechanism through spillover effects or positive externalities of the programme.

Figure 1.Logic Model illustrating how Payment for Ecosystem Services (PES) programmes might affect health and subjective wellbeing.

This logic model outlines the potential pathways through which PES interventions may impact individual and community health and subjective wellbeing outcomes. It includes behavioural, material, and environmental mechanisms, as well as moderating factors such as participant characteristics and policy context. The model links PES design elements (e.g., type of payment, contract length, recipient) with short-term and long-term health and subjective wellbeing outcomes.

1. Behavioural mechanism

The behavioural mechanism operates through influencing the actions and decisions of individuals and communities in the context of environmental conservation⁴⁵, with four potential pathways to health and subjective wellbeing. The first pathway is related to changes in environmental behaviours, such as reducing harmful environmental practices for the conservation of the environment and its ES³⁰. This is connected to health and subjective wellbeing through the services that these ecosystems provide. For example, forest conservation improves air quality through the provisioning service that acts via carbon sequestration in the air, resulting in cleaner air, which could affect health by reducing chronic respiratory diseases such as asthma or vector-borne diseases e.g., malaria. Another example could be related to the impact that conservation activities may have on the perceptions of the importance of caring for nature and the acquisition of new environmentally friendly skills. These impacts may lead to changes in the use of less harmful pesticides that can have an impact on diseases like cancer, or an improvement in self-esteem and social capital through networking within the community members. In contrast, replacing other uses of the land may provide lower economic gains⁴⁶, and therefore can negatively impact people's income that could have an adverse consequence on their health and wellbeing, e.g., less income reduces access to food and may impact on the food security of individuals and communities. Another harmful effect can be linked to the sense of obligation that PES programmes can have, which could generate a negative psychological impact on PES providers because they may feel they are losing control of decisions about their lands and community.

The second pathway is related to the impacts PES can have on the composition of the communities. PPI informed us that in some places the programme has increased rates of out-migration within the population in their communities. These changes can lead to loss of social support structure that will have an impact on mental health outcomes.

The third pathway incorporates PPI’s perspectives regarding how policy related activities such as monitoring activities in mangroves or native forests is perceived as a dangerous activity due to several threats of violence that participants of the programme are receiving from people that want to do some environmental harmful activities such as logging. This pathway can have a potential effect on all-cause of mortality and an increase in violence rates.

The fourth pathway is related to changes in health-related behaviours, which are also linked to the material resource mechanism. Extensive research has shown the association between low income and poorer health and subjective wellbeing outcomes, such as an increase in cardiovascular diseases, malnutrition (undernutrition/obesity), anxiety, depression, limited access to healthcare, reduced life satisfaction, and lower happiness among others^47–53.

Conversely, studies have also indicated that policies incorporating financial incentives can lead to positive behavioural changes, particularly in health-related behaviours such as healthier eating habits or increases in physical activity^54,55. In the case of PES programmes, these health-related behaviours could, for example, be related to an increase in physical activity due to conservation activities, thereby reducing cardiovascular diseases, as well as changes in dietary habits linked to an increase in individual/family income that may impact health outcomes related to malnutrition. Furthermore, this mechanism may shepherd a reduction in stress and anxiety through the feeling of calmness or happiness due to contact with nature⁵⁶. However, is important to highlight that from the leader’s perspectives some conservation activities such as monitoring activities could be dangerous and may put at risk their lives and health.

2. Material resources mechanism

This mechanism may affect health as a result of increasing the income of PES providers and/or affecting health-related behaviours. The latter was already described in the paragraph above. In this section we will only describe how the material resource can be a mechanism itself. For example, payments made to individuals can increase the acquisition of food that could have an impact on maternal and child health outcomes such as undernutrition or obesity, as well as providing a material resource would increase providers' income, alleviating situations such as the financial stress of not having a fixed monthly income, which would conduct to an impact on mental health. Payments at the individual or community level could expand the use of healthcare services through improvements in the community infrastructure. However, poor management of the payments or a lack of consensus of how the payments should be distributed can lead to an increase in conflicts and violence within communities or can lead to an increase in alcohol consumption due to the stress that technicians and managers of the payments may experience.

Additionally, it is important to consider that material resources may also cause some other negative effects on health and subjective wellbeing. For example, having more economic capital allows the purchase of more productive assets, which can involve either expanding or intensifying agricultural production from existing land or renting more land. This can lead to the diminishing of other ES, producing a negative impact on health/subjective wellbeing outcomes associated with the provision of these services.

3. Environmental mechanism

The indirect environmental mechanism of PES interventions is related to the conservation of ecosystem services¹⁴. This mechanism could bring indirect benefits or spillover effects on health and subjective wellbeing for both individuals participating in the programme and those who are not, by safeguarding ecosystems that provide regulating, cultural, provisioning, and supporting services. This can potentially reduce diseases associated with the loss of these services²³. For instance, a reduction in water and air pollution may impact respiratory and infectious diseases within these communities.

Finally, we highlight two essential points regarding possible variations in health and subjective wellbeing outcomes. Firstly, the importance of considering that health outcomes could vary over time. For example, diseases like cancer may require a long-term perspective to observe any changes as the progression from initial exposure to detectable cancer can span a lengthy period. In contrast short-term outcomes have shorter causal pathways, such as a decrease in maternal and child mortality rates as a result of the provision of healthcare services, that could be observed following a shorter lag period. This will allow us to infer that, although changes in health outcomes like cancer may not be directly observed, it does not mean these changes do not occur, but rather that the time needed to evaluate them might need to be longer. Secondly, it is expected that the effects of PES on health and subjective wellbeing are likely to be heterogenous, and may be influenced by various moderating factors such as participant characteristics (gender, age, socioeconomic position, ethnicity, level of education, forest management governance, disabilities), contextual factors (socioeconomic context, political context, previous ecosystem conditions, geographic location, land ownership, community governance structures), and the specific characteristics of the programme³⁰. For instance, many PES programmes require landowners who wish to participate voluntarily in the programme, but often the people living within areas that provide ES do not have legal ownership of these lands, leaving them excluded from the programmes. This can cause health inequalities by reducing access to material resources for the poorest areas, leading to the health of those in greatest need falling further behind.

Aim / Research questions of the review

The systematic review aims to evaluate the positive and negative impacts of PES interventions on health and subjective wellbeing outcomes for both recipients and the wider community. Drawing upon the logic model, the review will be framed around the following research questions:

Patient and Public Involvement

Leaders of three Ecuadorian communities (Quilotoa, Puerto del Morro and Kichwa Rukullakta) involved in a Payment for Ecosystem Programme (Socio Bosque and Socio Manglar) were aided the development of the logic model. They contributed via online meetings and provided feedback following their experiences and community to identify the most suitable primary health outcomes. This information was used to refine the logic model presented in this protocol. Leaders did not participate elsewhere in the study design.

Eligibility criteria

The PECOS criteria are described as following and are summarized in Table A1.

Population

The population of interest includes individuals, communities and/or households living in areas that are providing ecosystem services and those residing in an area where the PES programme is active but did not directly take part in the programme. We will include studies of programmes in all countries.

Outcomes for both adults and children will be included in the review and communities will be defined as a group of individuals who share elements of residence that characterise a population. It could include geographical proximity and living in areas that are providing ecosystem services. The communities can be at local, regional or national level.

Exposure

The exposure (intervention) of interest is defined by Wunder. S as “voluntary transactions, between service users and service providers that are conditional on agreed rules of natural resource management for generating offsite services”¹⁴. PES programmes can vary according to their components e.g., type of ecosystem service that the programme wants to conserve, payment structures, financing schemes, contract length, whom to pay, payment duration, type of payment, conditions to receive the payment, amount of the payment, time to pay, scale of the PES scheme, who the actors involved in the programme are, type of compliance monitoring compliance, and types of sanctions for non-compliance^{10,14,19,23–27}. Payments can be directed to individuals, households, communities, or organizations and may be contingent on meeting specific environmental commitments, such as maintaining a certain level of forest cover, or they may be issued in advance of the PES programme. We will not exclude programmes based on any variation in the type of PES programme implemented.

However, we will not be assessing the impacts of the money raising element on people who pay for PES e.g. money being collected from payers for carbon offsets.

Comparison

The comparison group will consist of other similar areas or people where the intervention did not occur, were exposed to a different intervention, or were exposed to differing forms of PES implementation, including: differing intensity, e.g. lower doses of payments; payment duration, e.g. long-term duration; type of payment (cash or in-kind); who received the payment (e.g. individuals or communities); who is the funder/ implementer (private versus public for example); or status of land (private land or state-owned/ protected land).

Those residing in an area where the PES programme is active but did not directly take part in the programme.

Studies that include a temporal (e.g. before-after) comparison will also be included.

Outcome(s) and prioritisation

We will include studies that assess the impact of PES programmes on health and subjective wellbeing outcomes as defined below. Outcome measures will include any measure which can be interpreted as a direct measure of health or subjective wellbeing.

Studies which only report health service use outcomes will be included. There will be no minimum follow-up period to assess health effects. In the case where synthesis across more than one study is possible, the outcomes from the most similar time point across the studies will be prioritised. For the primary outcomes this protocol will be looking for the outcomes most likely to impact health and subjective wellbeing and are likely to have data available^{1,6,45,56–66}. Specific outcomes could include any of the following: self-reported health measures, maternal and child health outcomes, external causes of morbidity and mortality e.g. violence, alcohol and drug harms, mental health outcomes, healthcare use, all-cause of mortality. All other outcomes will be considered secondary outcomes e.g., infectious diseases, asthma, malnutrition, cancer.

Study design and characteristics

Exclusion criteria

1. Incorrect population:

2. Incorrect intervention scheme:

3. Incorrect outcome:

4. Incorrect study type:

5. Incorrect language:

Information sources

The following databases will be systematically searched: EconLit, Medline, CAB Abstracts® and Global Health® via Web of Science, GreenFILE, International Initiative for Impact Evaluation (3ie).

In addition, grey literature from World Health Organisation (WHO), Food and Agriculture Organisation (FAO), and World Bank will be searched to identify publication using the same keywords as were used in the database searches.

Search strategy

The terms and search strategy are listed in the Table A2, where we illustrate a search strategy example in CAB Abstracts® and Global Health® via Web of Science. The strategy and set of terms used will be modified as needed for each database.

Reference lists and citations of the included studies as well as relevant systematic reviews identified during the search will also be reviewed to identity additional pertinent studies. The search strategy will not be restricted by factors such as age, gender, language, or race.

We will conduct the search strategy from 1990 until the present period. This date is justified since the first implementation of PES as a policy tool was in 1990, as noted in previous systematic reviews³⁰.

Study records

Selection of studies

Covidence will be used to manage the search results. MCalderon, as the lead reviewer, along with a second reviewer, will independently assess each study at all stages. Initially, the screening will focus on titles and abstracts, with only those studies meeting the inclusion criteria moving forward to full-text review. Studies whose titles and abstracts clearly do not fit the inclusion criteria will be excluded at this stage. During the full-text screening phase only studies which fully meet the inclusion criteria will be selected to progress to data extraction. Where there is disagreement or ambiguity about inclusion the full reference will be obtained to allow further scrutiny of the eligibility of the study. The reasons for exclusion will be documented in Covidence and reported according to the PRISMA guidelines⁶⁷.

In cases where included articles report duplicate data, the decision on which data to include will be based on the study design (prioritising lower risk of bias), sample size, directness in relation to the review question and ability to standardise effect estimates for the synthesis.

Conflicts in any part of the study selection and selection process will be discussed with a third independent reviewer or expert. The review team will hold regular meetings to discuss and interpret the evidence.

Data extraction and data management

Endnote X9 will be used as a reference library for the review. During the search and screening Covidence will be used to manage records and Excel spreadsheets will be used for data extraction. Data extraction will be performed by the lead reviewer (MCalderon) and independently checked by a second reviewer. To ensure consistency across reviewers, prior to beginning we will pilot the data extraction process with all who are involved in the data extraction process, extracting data from ~3 diverse studies and comparing results. We will resolve disagreements within the data extracted by discussion and in consultation with a third member of the research team when necessary.

Data extraction process

Data extraction will be carried out by the lead reviewer, MCalderon, and independently verified by a second reviewer. Considering the heterogeneity of the study designs and outcomes included within the scope of this review, both quantitative and qualitative data will be extracted using a convergent segregate approach. This approach involves separately synthesising quantitative and qualitative data leading to generate separate evidence, which will be then integrated⁶⁸.

Data will be extracted on an Excel spreadsheet. Any disagreements between reviewers will be resolved through discussion or by consulting a third reviewer. If relevant data appear to be collected but are insufficiently reported for data extraction, the study investigators will be contacted to obtain the necessary information.

We will extract specific details about populations, study methods, PES components, and outcome data relevant to the review objective. For the qualitative studies we will extract will specifically details about population, context, geographical location, study design, methods, PES components and self-reported health perceptions of interest participants’ reported health and wellbeing-related experiences of PES relevant to the review objective, as well as any potential mechanism through PES may impact on health and wellbeing.

Data extracted will include all the data items described in detail in Table A3.

Risk of Bias Assessment in individual studies

Two reviewers will independently assess risk of bias for all studies with disagreements in assessments resolved through discussion with the assistance of third reviewer. Based on the initial scoping of the available literature, most of the data will be from a broad range of studies, with a large range of potential biases. The following tools will be used:

Quantitative studies

For randomised studies, we will use the Cochrane Risk of Bias 2 tool , conducting assessments at the outcome level when required⁶⁹. The tool assesses the following domains bias arising from the randomisation process, bias to deviations from intended interventions, bias to missing outcome data, bias in measurement of the outcome, bias in selection of the reported result, with each domain receiving a judgement of low, some concerns and high risk of bias. An overall rating of bias will be determined by the worst risk of bias in any of the domains. However, if a study is judged to have “some concerns” about risk of bias for multiple domains, it might be judged as at high risk of bias overall.

For non-randomised quantitative studies, we will use the Cochrane Risk of Bias in Non-randomised Studies - of Interventions tool (ROBINS-I)⁷⁰. The tool assesses the following domains bias due to confounding, bias in selection of participants into the study, bias in classification of interventions, bias due to deviations from intended interventions, bias due to missing data, bias in measurement of outcomes, bias in selection of the reported result, with each domain receiving a judgement of low, moderate, serious, critical risk of bias. An overall rating of bias will be determined by if the study is comparable to a well performed randomised trial the study will be judged to be at low risk of bias for all domains. If the study provides sound evidence for a nonrandomised study but cannot be considered comparable to a well performed randomised trial, then it will be considered moderate risk of bias. However, if the study has some important problems, the study will be judged to be at serious risk of bias in at least one domain, but not at critical risk of bias in any domain. If the study has at least one domain at critical risk of bias then the study will be judged to be at critical risk of bias. Finally, if there is no clear indication that the study is at serious or critical risk of bias and there is a lack of information in one or more key domains of bias, the study will be considered as no information. We provide a detailed explanation of biases in each domain in Table A4.

Qualitative studies

Qualitative studies, including mixed-methods studies reporting qualitative data, will be assessed using the Critical Appraisal Skill Programme, which is publicly available for use in educational and research settings⁷¹ and includes ten questions divided into three sections: validity, results, and relevance. The signalling questions have three possible answers: yes, no, can’t tell⁷²

Where necessary, authors of the included studies will be contacted to request clarification or additional details related to their published results. This process aims to enhance data completeness for the review and does not involve the collection of new or original data from participants.

Any disagreements that arise between the reviewers will be resolved through discussion, with the assistance of a third reviewer. The results of critical appraisal will be reported in narrative form and in a table.

We will not exclude studies on the basis of their risk of bias but instead incorporate assessments into the synthesis process.

Data synthesis and integration

According to the research objectives and due to the high heterogeneity expected in the studies, quantitative and qualitative data will be collected and analysed independently and then integrated during the synthesis stage. The extracted data will include specific details about population, PES components, contextual factors, and RoB assessment.

To summarise key information from each study we will begin tabulating the data according to the objectives of the research as: impacts of PES programmes on self-reported health, maternal and child mortality and morbidity, mental health, all-cause of mortality and healthcare use according to the programme characteristics, participant characteristics, contextual factors, and the potential mechanisms of how PES programmes may affect health and subjective wellbeing.

For the quantitative data we will display estimates within a forest plot, when possible, to harmonise effect measures across the studies. We will perform an effect direction plot and calculate P values for effect direction, followed by a Harvest plot to study differences in effect across population subgroups.

If two or more studies report the same outcome and are sufficiently homogenous, we will perform meta-analyses. Given the likely variability among studies, we will use random-effects models for meta-analysis. When meta-analysis is not possible or appropriate, we will consider the Synthesis Without Meta-analysis (SWiM) framework instead^44,73. To make the best use of the available data we will use appropriate methods, such as descriptive statistics as well as harvest plots/effect direction plots/vote counting. A funnel plot will be generated to assess publication bias if there are 10 or more studies included in a meta-analysis.

Subgroup analyses will be conducted where there are sufficient data to investigate. We will stratify meta-analyses and conduct meta-regression by study design, RoB, characteristics of the PES programme, participant characteristics (gender, socioeconomic position, disabilities, ethnicity, level of education and age, forest management governance), and contextual characteristics (socioeconomic context, political context, previous ecosystem conditions, geographic location, land ownership, community governance structures) where data allow.

For the qualitative data we will conduct a framework analysis paying attention to contradictory data and making comparisons in findings across key characteristics related to our logic model such as the intervention characteristics, population subgroups, contextual factors, among others⁷⁴.

Integration of quantitative evidence and qualitative evidence

The results from each synthesis method included in this review will then be compared and organised into a cohesive argument to generate an overall configured analysis. This section will present a narrative summary that reflects the configured analysis of both quantitative and qualitative evidence. It will also compare the findings addressing the following questions: What are the impacts of PES programmes on health and subjective wellbeing? Does the impact of PES programmes on health and subjective wellbeing outcomes differ according to PES characteristics, participant characteristics, contextual characteristics? What are the potential mechanisms through which PES programmes may affect health and subjective wellbeing?⁶⁸.

Results of data synthesis will be mapped against our initial logic model, to refine the theory of change and to assess the credibility of the assumed causal pathways.

Subgroup analysis and investigation of heterogeneity

We will use sub-group analysis to explore heterogeneity by different treatment sub-groups. We will undertake the moderator analysis by the following groups of variables described according to our logic model:

Meta-bias(es)

Funnel plots will be generated if we find ten or more studies of the same outcome to detect potential reporting biases. When testing asymmetry in funnel plots (small study effects), we will investigate whether the relationship between a measure of study size and the estimated intervention effect is asymmetrical⁷⁵. For the qualitative and SWiM analysis, tables will be used to describe the synthesis of methods, limitations and subgroups of the review⁴⁴. We will consider conducting meta-regression when ten or more studies are included within a meta-analysis.

Before deciding which studies to include in the final synthesis, a sensitivity analysis will be performed to assess how the reported effects vary based on different study characteristics. We will accomplish this by reporting separately the low/moderate risk ROBINS-I studies. For the randomised studies we will report separately low and unclear risk studies. We will consider conducting meta-regression when ten or more studies are included within a meta-analysis.

Confidence in cumulative evidence

In this review, the certainty (previously referred to as quality) of evidence will be evaluated using Grading of Recommendations Assessment, Development, and Evaluation (GRADE)⁷⁶. We will present the information in a 'Summary of findings' (SoF) table to understand the characteristics of the populations and intervention included in the review and review the outcome data from the available studies in a structured way. Using GRADE, we will reflect the extent to which we have confidence, or our level of certainty, that the estimates of effect are correct. For GRADE the certainty of evidence will be described as high, moderate, low, and very-low confidence, and the domains assessed will be risk of bias, inconsistency, indirectness, imprecision, and publication of bias.

Results of the certainty assessment for each outcome will be presented alongside other key information for included outcomes in a SoF table, as specified in the Cochrane Handbook. The outcomes intended to be included in the SoF table are:

It is acknowledged that these planned outcomes may change during analysis due to data availability. However, it will be noted if there are critical outcomes for which there is a lack of available evidence.

Registration number – PROSPERO

CRD42024591264

Dissemination

We will disseminate the evidence for this systematic review to the academic community and policymakers through publication in a high-impact peer reviewed journal, relevant conferences and policy briefings.