A protocol for analysing the effects on health and greenhouse gas emissions of implemented climate change mitigation actions

Background: It is crucial to understand the benefits to human health from decarbonisation to galvanise action among decision makers. Most of our existing evidence comes from modelling studies and little is known about the extent to which the health co-benefits of climate change mitigation actions are realised upon implementation. We aim to analyse evidence from mitigation actions that have been implemented across a range of sectors and scales, to identify those that can improve and sustain health, while accelerating progress towards a zero-carbon economy. Objectives: To understand the implementation process of actions and the role of key actors; explain the contextual elements influencing these actions; summarise what effects, both positive and negative, planned and unplanned they may have on emissions of greenhouse gases and health; and to summarise environmental, social, or economic co-benefits. Data: We will review evidence collected through partnership with existing data holders and an open call for evidence. We will also conduct a hand search of reference lists from systematic reviews and websites of organisations relevant to climate change mitigation. Screening: Screening will be done by two reviewers according to a pre-defined inclusion and exclusion criteria. Analysis: We will identify gaps where implementation or evaluation of implementation of mitigation actions is lacking. We will synthesise the findings to describe how actions were implemented and how they achieved results in different contexts, identifying potential barriers and facilitators to their design, implementation, and uptake. We will also synthesise their effect on health outcomes and other co-benefits. Quantitative synthesis will depend on the heterogeneity of outcomes and metrics. Conclusions: Findings will be used to identify lessons that can be learned from successful and unsuccessful mitigation actions, to make inferences on replicability, scalability, and transferability and will contribute to the development of frameworks that can be used by policy makers.


Introduction
It is of great importance to identify and communicate scientific evidence that could support national and sub-national policy makers to take actions towards a post-carbon society, in which there is net zero emissions of carbon dioxide and short-lived climate pollutants. In addition to the benefits of decarbonisation to the environment, there could also potentially be significant benefits to human health 1,2 . The Pathfinder Initiative aims to synthesise evidence on the development and implementation of actions across a range of sectors that improve and sustain health while accelerating progress towards a zero-carbon economy. This study will fill a number of knowledge gaps that are impeding progress -namely which climate change mitigation actions will have the largest benefits (and will have the least trade-offs and inequities) for health in particular contexts, and what additional environmental, social, or economic benefits such actions might have. We will also discuss the successes or challenges of implementation, the pathways of change which demonstrate how actions lead to health and other co-benefits in addition to greenhouse gas reductions, and which actions should be employed for effective scale-up in particular contexts. The Pathfinder Initiative seeks to gather and analyse examples where climate change mitigation actions have been implemented in practice and their impacts on health assessed. This protocol describes the process by which these examples will be identified and analysed.

Aim, objectives, and research questions
The overall aim of the research is to review the evidence from examples where mitigation actions have been implemented in practice, in different contexts, and have had an impact on a human health outcomes or exposures. The objectives are: to understand the implementation process of these actions and roles of the key actors; explain the contextual elements influencing these actions; and summarise the impacts these actions might have, both positive and negative, planned and unplanned, on human health. The intention is also to summarise what additional effects they may have on greenhouse gas emissions and other environmental, social, or economic outcomes. We will also seek to identify mechanisms to explain the success or failure of these actions and their implementation, and identify plausible links between actions, context, and outcomes. This study will address the following questions:

Implementation
· What are the implemented actions and how closely did they correspond to what was intended?
· Who implemented these actions?
· Who are the beneficiaries of these actions (planned and unplanned)?
· Why were these actions implemented?
· Where were those actions implemented?
· How was implementation achieved?
· What are the spatial and timescales of the implemented actions?
· What costs were associated with the implementation?
· What is the potential for scaling up implementation?
Pathways · What are the pathways of impact on health?
· What is the response of stakeholders (local, national)?
Context · What contextual elements (barriers and facilitators) influenced the design, implementation, trade-offs and spill-overs, or the rate and scale of uptake of these actions?
Outcomes and impact · What is the magnitude of the benefits of these actions on health outcomes?
· What is the magnitude of the benefits, trade-offs, or spill-over effects on other environmental, social, or economic outcomes?
· How are the benefits, and potentially the costs, distributed? Is there equitable distribution across groups, societies and regions?
· What are the timeframes of the benefits achieved or expected to be achieved?

Amendments from Version 1
We have made changes to the introduction to clarify that will discuss the successes or challenges of implementation, the pathways of change which demonstrate how actions lead to health and other co-benefits in addition to greenhouse gas reductions.
We have added text to the "Design" section to clarify that this is not an exhaustive effort that would allow for mapping of the gaps in the evidence base, however, the effort is extensive and allows for identification of where to target data collection and where evidence may be scarce or lacking.
We have fixed the error in Table 4, where there was a mismatch between the system category and the actions listed for each. This only required moving the categories up to include the last row of actions from the category above.
We have made two changes to Table 4, in the "actions" column. 1) change from "switch to low-carbon diets" to "switch to low GHG emission healthy diets", 2) From "replace fossil fuel energy with clean energy" to "Replace fossil fuel energy with clean or renewable energy", 3) From "switch to renewable energy sources" to "switch energy generation to renewable sources".
Under the sub-section "measured impact" we made a minor change to the text where we removed the word "modelled" and clarified that this means we are not focusing on studies based on data that are assumed.

Design
To answer the research questions outlined above, we will analyse and summarise process and outcome evaluations of mitigation actions that have been implemented in practice. We will also analyse examples where formal evaluations have not taken place or are not explicitly described as such, but where sufficient information is provided to answer the research questions.
We will conduct a thematic synthesis of examples describing the implementation, pathways to impact, and context of implemented mitigation actions. We will describe how these actions achieved their effects in different contexts (including unintended effects and spill overs), or why they did not. This will allow us to develop theories of change that incorporate potential barriers and facilitators of design, implementation, and receipt of these actions, relating to characteristics of participants and contexts.
We will also conduct a narrative synthesis of examples describing outcome evaluations, aimed at summarising measured outcomes and their effects. Where data permits and depending on the heterogeneity of outcome evaluations found, we will conduct a meta-analysis of experimental and quasi-experimental studies of the effectiveness of mitigation actions on health, and any environmental, social, or economic co-benefits to health.
This work will allow us to understand which mitigation actions have been implemented in practice and across which sectors and regions and identify where implementation or adequate evaluation of implementation may be lacking. Although this is not an exhaustive effort to allow for mapping of the gaps in the evidence base, the effort is extensive and allows for identification of where to target data collection and where evidence may be scarce or lacking.

Data sources
The main source for data will be through a review of evidence, which will be obtained in the following ways. An open call for evidence will be circulated through networks of the above collaborators and distributed to other international actors including major funders of climate action (e.g., the Green Climate Fund, Regional Development banks, bilateral donors, national and sub-national governments), UN agencies (including WHO and UNDP), the Climate Ambition Alliance, non-governmental organisations (NGOs) and the private sector (through organisations such as the World Business Council on Sustainable Development).
The Pathfinder Initiative also comprises of the Lancet Pathfinder Commission with membership from all major global regions and sectors involved with climate mitigation that provides scientific guidance and oversight. Commissioner networks will be used to further circulate the call. A Comment has been published in The Lancet outlining the Pathfinder Commission and its call for evidence to encourage submissions from readership of The Lancet 3 .
We will utilise LinkedIn and Twitter in particular for our call for evidence in order to make use of professional networking platforms.
Drawing on the systematic search for a related study. Our second source of evidence will be the collection of systematic reviews identified as part of the umbrella review developed by the Pathfinder Initiative. The umbrella review will metasynthesise scientific evidence (both modelled and observed in implementation studies) on the solutions/actions that have been synthesised in published systematic reviews (a protocol for this study will be published separately). For the purposes of the analyses described in this protocol, we will review the reference lists and data extracted for the umbrella review to identify any relevant original studies that could serve as further examples meeting our inclusion and exclusion criteria.
Hand search. We will hand search the literature in several ways. First, we will search websites of organisations and climate change projects that are known to the Pathfinder Initiative team, a non-exhaustive list of sources is provided in Table 1.
Second, we will examine the reference lists of included documents to identify other potentially relevant documents for inclusion. Finally, we expect that the open call for evidence and our hand search will yield reports that summarise multiple implemented mitigation actions, in addition to reports of

Second stage of document collection
In this stage we will identify gaps in terms of system transitions, as listed in Table 2, as well as world regions, from which we find insufficient evidence and attempt to fill these gaps with a second round of targeted collection. We will ask collaborators, funders of climate change related actions, and members of the Lancet Pathfinder Commission to make suggestions or referrals to relevant local institutions, groups, or communities and take a snowballing approach to identify and purposively target entities within the sectors and regions from which the first stage of document collection yielded insufficient evidence.

Selection criteria
The criteria that we will base our selection of implementation examples on were agreed through discussions among all teams of the Pathfinder Initiative and are as follows.

Type of document
We will include all types of documents that are submitted to us. This could include published or grey literature, policy documents, internal reports, etc. Year We will restrict inclusion to examples that describe interventions or actions of climate change mitigation (as defined below), that were implemented from 2000 onwards to keep a relevant scope of actions.

Language
We will also include reports written in all the languages that are within the capacity of our research team, which includes English, Arabic, French, Spanish, Portuguese, Dutch, Italian, German, Russian, and standard Chinese (Mandarin).

Implementation stage
We will restrict inclusion to actions that have been fully or partly implemented, including those that are currently ongoing or those that were abandoned due to failures in implementation. implemented mitigation actions that are individually described.
Actions discussed in such summary or comprehensive reports will be extracted and traced back to their source for more information.

Actors
We will include actions implemented by any public sector, civil society, or non-profit actors. Actions that are implemented by individuals or for-profit firms that represent a very large part of a specific sector will be included only if an assessment of impact was conducted by an independent evaluator.

Implementation scale
We will restrict inclusion to actions that can be adopted by policy-makers and applied on a collective level. This includes plans adopted by national, regional or governmental bodies; international or sectoral agreements; community-led projects; built environments; technologies; governance arrangements; regulatory changes; fiscal mechanisms; mass media interventions; indigenous approaches. We will not include examples of individual or household-level choices which are not described as driven by specific collective or institutional-level actions.

Health outcomes
We will restrict inclusion to examples of climate change mitigation actions that have demonstrated an observed or measured impact on health outcomes directly, or outcomes that have an evidence-based pathway to health or quantifiable associations to health outcomes. This includes risk factors for health as identified from the Global Burden of Disease 5 ; reduced exposures to climate change impacts (e.g., from combined adaptation and mitigation actions such as reduced heat exposure, or reduced impacts of climate disasters); and improved socioeconomic determinants of health that have direct links to health. A preliminary list of health or health-related outcomes is shown in Table 5. We will also include examples of mitigation actions where no health outcomes are measured only if these are actions implemented within the healthcare sector specifically. In our view, the delivery of healthcare at reduced carbon intensity is a relevant health-related outcome in itself. Figure 1 demonstrates how inclusion or exclusion of examples based on the definition of mitigation actions and health or health-related outcomes will be decided.

Measured impact
We will restrict inclusion to examples that report on measured or observed health outcomes (as defined in the sections above), not those that are solely forecast or based on data that are assumed.

Screening
We operationalised the criteria described above into an inclusion and exclusion worksheet ( Table 6) through pilot screening of 20 documents that were already known to the research team. Four reviewers (RG, SH, SCGD, SW) screened the documents in pairs and met to discuss the screening process to ensure consistency in applying the criteria. This resulted in refining the screening criteria and developing the inclusion and exclusion worksheet.
Actions that achieve climate change mitigation Based on the Intergovernmental Panel on Climate Change (IPCC) definition, climate change mitigation is defined as actions, or interventions, that reduce the rate of climate change. Climate change mitigation is achieved by limiting or preventing greenhouse gas emissions and short-lived climate pollutants and by enhancing activities that remove these from the atmosphere.
We will include examples where climate change mitigation has been demonstrated through an observed or measured reduction of greenhouse gas emissions and short-lived climate pollutants (as listed in Table 3). We will also include examples of implemented mitigation actions that do not directly measure the reduction in greenhouse gas emissions when there is sufficient evidence for the type of action that mitigation can be assumed. For example, reports that describe a switch from the use of fossil fuels to the use of solar panels as an energy source. Table 4 shows a preliminary list of relevant mitigation actions adapted from the IPCC 2014 Synthesis Report 4 that will guide the inclusion and exclusion of implemented mitigation action examples.
We will include all actions where climate change mitigation was achieved, whether as a primary objective, or occurring as one of multiple benefits (whether planned or unplanned) to the implemented action (for example, adaptation or health actions that result in multiple benefits including climate change mitigation). The documents used in the pilot process were not of the typical format of published literature, whereby a clear title and abstract can be used for screening. The reviewers relied on executive summaries for screening, and in the case that executive summaries did not provide enough detail to enable a decision, the full document was assessed. This is expected to be the case for the majority of the full set of documents.
Screening of the full set of documents will be done by two reviewers. One reviewer (SH) will screen all the available  Unemployment 10-12 Female education and female participation in the workforce 13 documents. Documents will then be divided among the rest of the team for a second screening, blinded to the assessment of the first reviewer. The reviewers will meet to discuss and resolve differences in opinion, and the principal investigators (PI) will make the final decision in case of disagreement.
Screening will be conducted using the screening worksheet on Microsoft Excel (Version 16). Each document will be entered as a row and assessed against each of the criteria presented in the columns (Table 6).

Data extraction
We will extract data to a worksheet on Microsoft Excel (Version 16) on study characteristics (study location, timing of mitigation action, timing of evaluation); organisational and individual participant characteristics; and characteristics of study design and methods (research aims and objectives, sampling, data collection and analysis). We will use NVivo 12 to code data on implementation, pathways, context, and outcomes.
The Medical Research Council (MRC) guidance for process evaluation of complex interventions 14 will be used to guide the extraction of process evaluation data as follows.

Theory of change:
We will extract data on the theory of change underpinning the implemented action; links to other theories; and the descriptions of how the action is intended to achieve its health, environmental, or socioeconomic outcomes.

Action design or development:
We will extract data on how the mitigation action was designed or developed and by whom.

Action characteristics:
We will extract data on the description of the specific climate change mitigation actions that were used; the components; sector or domain; target; and scale of the action.

Resources:
We will extract data on resources that were required for the implementation of the intervention. These include human, material, and economic resources, including costs of the intervention and costs associated with implementing the intervention including investment, supply, and opportunity costs.

Implementation:
We will extract data on individual or multifaceted strategies that were used to implement these actions and the broader key processes that were used, such as planning, educating, financing, restructuring, managing quality, or attending to policy context. We will extract data on the fidelity, or the quality, of the implemented action in practice compared to what was designed or intended and whether adaptations had to be made; the extent to which the intended population was reached; the strength of the implemented action, i.e., how much of the intended action was actually delivered; and the acceptability of the action to stakeholders or the target population. We will also extract data on the characteristics of the actors, both on the organisational as well as the individual level, including who enacted the mitigation actions, their role and their skills.  * This table shows the possible classification, or labelling, that could be given to each example of an implemented mitigation action, for each of the criteria specified in the columns. ** The criteria that will be used to guide the decision on inclusion or exclusion, as described in the 'selection criteria' section of this protocol *** Outcomes that will be flagged during the screening process to aid in the identification of these examples for analysis of co-benefits, or additional secondary analysis, but will not be used to guide the decision on inclusion or exclusion Context: We will also extract information on contextual factors affecting the design, communication and implementation of actions, trade-offs and spill overs, and the rate and scale of uptake. This will include contextual factors within the organisation (inner setting), as well as the wider cultural, economic, and governance and political context (outer setting).
For outcome evaluations, we will extract data on the research design; the type of impact; which could be either net impact or gross impact; the outcomes measured (health, environmental, economic); and effect sizes. In the case of experimental or quasi-experimental studies (first tier of evidence), we will extract data on the nature of the control group(s); unit of allocation; generation and concealment of allocation; blinding; adjustment/control of clustering and confounding.
One reviewer (SH) will extract the data. The reviewer will pilot data extraction on 15 studies and discuss their extraction coding with the Pathfinder team to ensure quality and consistency in their interpretation, and will meet regularly with the Pathfinder team to discuss their coding and findings.
We expect that the documents obtained for this study will mainly be in the form of reports, administrative documents or internal records, news articles, etc. and as such are written for a specific purpose and audience other than research. In the case that the documents are found to have missing information that precludes quality assessment or synthesis of findings, we will contact the authors of the reports for additional information.

Quality assessment
The quality of each included case will be assessed by two reviewers. One reviewer (SH) will appraise all included documents. Documents will then be divided among the rest of the team for a second appraisal. Differences in opinion will be resolved by discussion and in the case of disagreement the PI will make the final decision.
The quality of each study will be assessed based on the AACODS (Authority, Accuracy, Coverage, Objectivity, Date, Significance) checklist that is designed for the critical appraisal of grey literature 15 . This checklist is used to assess the grey literature based on the following criteria: authority, or who is responsible for the intellectual content; the accuracy of the content, based on whether the document has clear aims and objectives and used valid methodologies that are clearly described; whether limitations in the coverage of the work are clearly stated; the extent to which the work is objective or is biased (i.e., representing opinion); whether the work is clearly dated in terms of when it took place, when it was reported, and key contemporary material were included; and whether the work undertaken is of significance or relevance to the research area.
In addition to this overall assessment, further assessment for each of the process and outcome evaluation data will be undertaken.

Process evaluations
Documents that report on process evaluations (whether explicitly using this terminology or implicitly describing implementation, pathways of change, and context) will be assessed using quality tools for qualitative studies 16,17 . These criteria address reliability in terms of the rigour of sampling, data collection, and data analysis; and usefulness in terms of breadth and depth of findings, and the extent to which stakeholder perspectives were explored (Table 7). Table 7. Quality assessment of qualitative data.

Rigour
Were there clearly stated aims and objectives?
Were steps taken to minimise bias and error/increase rigour in sampling?
Were steps taken to minimise bias and error/increase rigour in data collection?
Were steps taken to minimise bias and error/increase rigour in data analysis?
Were the findings of the study grounded in/supported by data?

Usefulness
Was there good breadth and/or depth achieved in the findings?
Was there an explicit account of a theoretical framework, a theory of change, or a logic model and/or the inclusion of a literature review which outlined a rationale for the intervention?
Was the implementation of mitigation actions adequately described?
Were the perspectives of stakeholders adequately explored?
Was there a clear description of context which includes detail on barriers and facilitators important for interpreting the results?

Outcome evaluation
Quantitative studies may be classified according to the following hierarchy of evidence. Studies will be classified into "Tier 1" if they use methods that control for confounders, such as natural experiments, experimental (cluster randomised controlled trials, stepped-wedge trials, etc.) or quasi-experimental designs (interrupted time series, difference in difference, etc.). The strongest type of evidence will be from studies that report on a control or comparison group that is similar in its characteristics and pre-intervention outcome variables to the intervention group and report on pre-and post-intervention data for all groups recruited into the evaluation and on all outcomes. Studies will be classified into a "Tier 2" of evidence when they use a simple comparator, such as before/after measures, or cross-sectional data to compare affected or unaffected areas. Studies will be classified into "Tier 3" of evidence if they report on data collected after the intervention with no comparator provided. Finally, a "Tier 4" of studies will be those that report a description of the health outcomes achieved rather than providing quantitative estimates. The quality of the documents within each of these tiers of evidence will be appraised using standard Critical Appraisal Skills Program tools.

Synthesis and reporting
We will identify gaps in evidence where we did not find examples of implementation or implementation that have been adequately evaluated against a framework of mitigation actions, health outcomes, and pathways to health outcomes, which is currently under development. The findings will in turn also be used to refine this framework in an iterative process.
We will produce a taxonomy of specific mitigation solutions that have been implemented in practice, group them into broader categories of mitigation actions, and describe how they were implemented, the effects they had on different outcomes, and how they achieved their effects in different contexts. We will identify the potential barriers and facilitators to the design, implementation, and receipt of these actions among the target population. We will compare the types of actions that have been implemented and the context within which implementation took place between examples to identify similarities and differences.
We will conduct a narrative synthesis of the effects of mitigation actions on health outcomes, or any environmental and socioeconomic co-benefits. Once we know the number of implementation examples available to us and the extent of heterogeneity amongst them, we will make a decision on whether and how we can present a pooled effect size. For example, if we find that a sufficient number of examples of similar mitigation actions have measured similar outcomes, we may be able to pool these effects.
We will use graphical synthesis techniques to represent our data. For example, heat maps to show the distribution of evidence across sectors, and harvest plots where quantitative evidence on outcomes is available but not amenable to pooled analysis.
We will use the findings to identify lessons that can be learned from both successful and unsuccessful implementation of mitigation actions and implementation strategies used to deliver these actions, and make inferences on their replicability, scalability, and transferability. The findings from this synthesis will contribute to the development of frameworks and documents that can be used by policy makers and other actors in the field of planetary health. We will produce practical guidelines targeted towards key audiences with a focus on the findings of this study, illustrating worked examples of what actions can be implemented in practice, how they can be implemented, in which contexts, and what effects they might have on health and other co-benefits.

Dissemination
Early findings will be published in an interim Commission report in The Lancet ahead of COP26 in November 2021 and a more detailed overview of the key findings will be published in the full Commission report ahead of COP27 in 2022. Example case studies and policy briefs for specific audiences will also be published online and shared through our partner networks in briefing papers, newsletters and through webinars and presentations to key stakeholder groups. Results will also be shared at key events on decarbonisation and health in 2021 and 2022 including (but not limited to) the Planetary Health Alliance annual meetings, the 2021 WHO Global Conference on Climate Change and Health and the World Health Summit in October 2021.

Study status
We are currently undertaking the first stage of document collection. We are engaging with partners to search their existing databases and have circulated our open call for evidence through our networks of collaborators and other international actors and major funders, and social media, as outlined in the 'Data sources' section. We have also identified several documents from the references of the systematic search for a related study and the hand search, as outlined in the 'Data sources' section. One reviewer (SH) has screened 30 documents that have been identified from these processes so far.

Data availability
No data are associated with this article.

Jose M. Rodriguez-Llanes
European Commission Joint Research Centre, Ispra, Italy This study protocol is extremely welcome, and will be helpful in keeping track and accountable the efforts to decarbonize our planet and give birth to a new world economy. This protocol opens an important path, one of demonstrating the multi-layered benefits of mitigation policies.
The authors are realistic about their prospects for project implementation: selected examples where sufficient evidence is available and have been published will be analyzed via secondary analysis (e.g., meta-analyses) and supplemented with thematic analysis. This is fine. And yet one might question if given the complexity and emergency of the task, should the authors consider running their own integrated statistical analyses using some of these data systems? For example, getting all the evidence together could take considerable time if we think that reports sometimes take years to appear and might be of varying quality. But also understandable, many of these fields are out of reach and use highly specialized data not readily available or even shared. One benefit of integrated analysis is robustness of the findings (given data is jointly analyzed), other is quality control of analytics. By adding this option into their protocol, the authors could answer specific questions, which maybe remain unanswered by others.

Ending paragraph at Design section, the authors wrote:
'This work will allow us to map which mitigation actions have been implemented in practice and across which sectors and regions and identify where implementation or adequate evaluation of implementation is lacking'.
I think this point is contradictory. The intent, if I understand correctly, is to capture examples which can be investigated in-depth, and for which sufficient quality evidence can provide an overall picture of benefits and co-benefits. The sentence above ('map') gives the impression that this will be an exhaustive effort. In section Measured impact, the authors wrote: 'We will restrict inclusion to examples that report on measured or observed health outcomes (as defined in the sections above), not those that are solely forecast or modelled.' I assumed that a great majority of those will be modelled observed data, as health outcomes could be confounded by other interventions in the real world (reason why authors rightly chose to include experimental or quasi-experimental studies). Hence, I am confused by the sentence ('not those that are solely forecast or modelled'). Moreover, in figure 1, the two first situations may include reports with no health outcomes. This seems, at first, to contradict the above statement by the authors.
Overall, a good piece of work, and a clear guide for the ongoing and future data collection and synthesis. The piece has the potential to inspire more scientific evidence and accelerate the scaling of similar initiatives by other groups.
Is the rationale for, and objectives of, the study clearly described? Yes

Are sufficient details of the methods provided to allow replication by others? Yes
Are the datasets clearly presented in a useable and accessible format? Not applicable think that reports sometimes take years to appear and might be of varying quality. But also understandable, many of these fields are out of reach and use highly specialized data not readily available or even shared. One benefit of integrated analysis is robustness of the findings (given data is jointly analyzed), other is quality control of analytics. By adding this option into their protocol, the authors could answer specific questions, which maybe remain unanswered by others." Thank you for your comments you're your suggestion. We do aim to synthesise and integrate the findings from all literature that we find (whether reports or published articles), in an iterative manner. That is, analysis will begin alongside data collection, whereby early findings will feed into how we collect data and which areas we target, and the additional data collected will be added and analysed. Therefore, findings will be synthesised and integrated as we go along, rather than wait until data collection is complete.
"Ending paragraph at Design section, the authors wrote: 'This work will allow us to map which mitigation actions have been implemented in practice and across which sectors and regions and identify where implementation or adequate evaluation of implementation is lacking'. I think this point is contradictory. The intent, if I understand correctly, is to capture examples which can be investigated in-depth, and for which sufficient quality evidence can provide an overall picture of benefits and co-benefits. The sentence above ('map') gives the impression that this will be an exhaustive effort." Thank you for raising this point. We have made a change taking this into account.
" Table 5:  Shall table 5 include wasting, stunting, and other standard, widely used, and measurable nutritional outcomes? In addition to deficiencies in minerals and vitamins?" The table is not intended to be exhaustive. As the title of the table "Preliminary list of health outcomes or risk factors for health to guide the inclusion/exclusion of examples" suggests, t his list is preliminary because it will be updated with any additional outcomes that are identified from the literature, and it was put together only to guide data extraction and not to be prescriptive. We do mention that data extraction and analysis will be iterative.
"In section Measured impact, the authors wrote: 'We will restrict inclusion to examples that report on measured or observed health outcomes (as defined in the sections above), not those that are solely forecast or based on data that are assumed.' I assumed that a great majority of those will be modelled observed data, as health outcomes could be confounded by other interventions in the real world (reason why authors rightly chose to include experimental or quasi-experimental studies). Hence, I am confused by the sentence ('not those that are solely forecast or modelled')." Thank you, we have made a clarification. We do not mean that we will not use analytical models. We only mean that outcomes should be observed, rather than forecast or based on assumed data.
"Moreover, in figure 1, the two first situations may include reports with no health outcomes. This seems, at first, to contradict the above statement by the authors." We do make clear in the text that we make an exception for actions that aim to decarbonise the healthcare sector, and this is the only exception we make.