Finding

After searching PubMed for papers published over the last 10 years, using a customized API and keywords for scoping climate and health software, we found only 37 fully developed and named tools.

This suggests that few studies progress from providing the initial evidence of climate-health linkages to the operationalization of a decision support tool that could inform actions to reduce the burden of disease. Our qualitative research indicated that this might be due to lack of incentives for academics to publish full models, and lack of multi-disciplinary communities which bring together scientists and software developers.

Although only 37 tools met our full list of characteristics, we found several models of climate and infectious disease that had well described, freely accessible model outputs but no access to code repositories.

Examples of two such models were an Aedes mosquito model from Ryan et al. (2019)1, and a dengue climate forecasting model from Lowe et al. (2021)2. The former had no associated code repository and the latter had no named software platform. Both models can potentially be developed into software packages or platforms.

Recommendation

Transition validated models into tools.

Figures from example models

Finding

The majority of tools identified in our review (81.1%) focused on vector-borne disease systems. Only 10% of modelling tools were applied to infectious diseases with other modes of transmission, including respiratory (5.4%), foodborne (2.7%), and waterborne (2.7%). There were no specific tools for soil-borne diseases.

We targeted specific pathogens and disease classes and of 42 disease terms we searched for, there were robust tools for only six. Most notably, we could not identify a robust tool for Ebola.

The closest match was Ebola-Spread3, a predictive niche modelling tool from Oxford University. The associated GitHub repository has incomplete code, requiring the user to have the data inputs that are not specified or linked, and the repository itself was not up to date, meaning it is effectively unusable.

Recommendation

Tools for neglected disease groups.

There is an opportunity to develop tools for climate-sensitive diseases transmission modes that have been neglected (e.g., respiratory, foodborne, soilborne, waterborne) to increase the preparedness of the public health sector for the next pandemic.

3. Kraemer, M. U., Golding, N., Bisanzio, D., Bhatt, S., Pigott, D. M., Ray, S. E., ... & Reiner, R. C. (2019). Utilizing general human movement models to predict the spread of emerging infectious diseases in resource poor settings. Scientific reports, 9(1), 1-11.

Finding

More than half of the tools were described as operationalized; however, only one quarter were freely available online and one quarter had interfaces legible to finance, policy and regional decision makers. Those that did have legible interfaces were funded by an institutional or country-level partner.

By a legible interface we mean a graphical user interface which enables decision makers to understand the disease risks, without diving into code or mathematical formulas.

An example of a legible model was HYDREMATS4, where the current and future scenarios for malaria risk in West Africa are visualised over a map. The interface enables a user to select a year and explore the mean monthly rainfall (mm), temperature, immunity index, malaria prevalence, entomological inoculation rate and R0 as predicted by the model. It also provides a brief explanation of the model itself.

Recommendation

Co-creation of tools

Tools with Legible Interface

4. Yamana TK, Eltahir EAB.2013. Projected Impacts of Climate Change on Environmental Suitability for Malaria Transmission in West Africa. 121:10

Finding

Most tools were developed for, and implemented in geographic areas where the infectious disease of interest is currently endemic. The tools found in this review had been implemented in several WHO regions, spanning Africa (43.8%), the Americas (14.6%), Europe (10.4%), the Western Pacific (10.4)%, South-East Asia (6.3%), and the Eastern Mediterranean (2.1%). Four tools (8.3%) did not focus on a single geographic region, but rather a global study.

A recent study showed tropical highland areas in the African region, the Eastern Mediterranean region, and the region of the Americas could experience an annual increase in the number of months that are climatically suitable months for malaria transmission by the end of the century -- an increase of about 1.6 months per year5. Tropical lowland areas in the Western Pacific region and the Eastern Mediterranean region will experience the greatest increase in climatic suitability for dengue transmission by approximately 4 additional months per year. Without bold action to curb emissions, the population at risk of both diseases is predicted to increase by an additional 4·7 billion people, particularly in lowlands and urban areas. These findings are consistent with other estimates that dengue transmission could newly impact an additional 1 billion people by 2080, with the largest increases in the western European region6. It is estimated that an additional 1.3 billion people will be at risk of Zika fever by 2050, with different regions experiencing an increase in risk depending on the climate scenario (e.g., Eastern sub-Saharan Africa experiencing the largest increase in risk under RCP 8.5, and North America experiencing the highest under RCP 4.5)7.

Recommendation

Tools for zones of emergence

There is a need for modelling tools that address disease dynamics in zones of current or future emergence, as indicated by the projected shift in the burden of climate-sensitive disease transmission under climate change scenarios.

There is also an opportunity to cross-pollinate knowledge and experiences between regions that are currently endemic for climate sensitive diseases and regions of projected disease emergence due to climate change.

Map of geographic regions where tools and studies were implemented

5. Felipe J Colón-González, Maquins Odhiambo Sewe, Adrian M Tompkins, Henrik Sjödin, Alejandro Casallas, Joacim Rocklöv, Cyril Caminade, Rachel Lowe (2021). Projecting the risk of mosquito-borne diseases in a warmer and more populated world: a multi-model, multi-scenario intercomparison modelling study. Lancet Planet Health 2021;5: e404–14.

6. Ryan SJ, Carlson CJ, Mordecai EA, Johnson LR. Global expansion and redistribution of Aedes-borne virus transmission risk with climate change. PLoS Negl Trop Dis. 2019;13: e0007213.

7. Ryan SJ, Carlson CJ, Tesla B, Bonds MH, Ngonghala CN, Mordecai EA, et al. Warming temperatures could expose more than 1.3 billion new people to Zika virus risk by 2050. Glob Chang Biol. 2021;27: 84–93.

Finding

North American and European institutions are disproportionately represented as tool-creators (38% in the USA and UK alone). There is a need for greater representation of the Global South, where many of the tools were designed to be used.

There were 102 institutions represented in the author list of the 48 publications screened in our review. Over one third (38.2%) of these institutions were based in the USA or UK. Only 24 institutions were associated with more than one paper, which included universities and agencies located in Europe (n=11), the Americas (n=8), the Western Pacific (n=3), and South-east Asia (n=1).

Designated corresponding authors for the 37 tools were based in the USA (32.4%), UK (27.0%), other European countries (29.7%), Australia (5.4%), South-east Asian countries (5.4%), and Tanzania (2.7%).

An analysis of 9000 publications related to modelling at the intersection of climate change and infectious disease revealed:

• LSHTM acting as bridge between North America and Asia & Pacific
• Norwegian Institute of Public Health acting as a conduit into Scandinavian Institutions
• Africa and South/Central America under-represented, but certain institutions are building outreach such as the University of Florida.
• Separate network with Liverpool University at the centre.

Recommendation

Equity and diversity of investigators

There is a need to support teams that are led or co-led by researchers and other partners from the Global South, where the impacts of climate sensitive infectious diseases are the greatest.

Network Analysis of Institutes publishing together

Finding

We identified large variations in the spatial and temporal scale of the 37 tools.

The spatial scale of tools, which includes geographic extent and unit of model output, varied considerably for tools in our final list. Model scales ranged from highly localized foci (8.1%), for example simulations for individual villages, to tools with a global or continental extent (16.2%).

Some examples of local tools from our final list include ANOSPEX, SLIM, and FleaTickRisk, while MARA, Vibrio Map Viewer, and the albopictus package operate on broader, continental scales.

The ideal scale of tool is dependent on goals; local tools may be limited in terms of where they can be operationalized, and their transferability, while being well validated locally. Tools with very coarse resolutions may aid broad planning efforts, but be of limited use to local stakeholders. The takeaway here is that there is no “best” scale for models, and it is important to encourage models at different scales to help facilitate decision-making at different granularities.

Few models forecast disease risk beyond the seasonal climate outlook.

Recommendation

Multi-scalar tools

There is a need to develop tools across a range of spatio-temporal scales to capture climate and disease processes at different scales and to support decision making needs across scales. An analysis of the gaps in the spatial and temporal data infrastructure would provide important guidance on gold standards.

Vibrio Map Viewer

Finding

We found that climate datasets are readily available, but sharing of health datasets is politically sensitive. Most public health and climate sectors don’t have a mandate to focus on the climate impacts on health (yet). There is a need to influence and inform policy and dataset sharing.

Recommendation

Policy Action

There is interest from the climate and health sectors to work together to co-create modelling tools. However, often the sectors lack clear mechanisms for data sharing or lack a political mandate to engage in these efforts. There is a need to influence and inform policies that encourage intersectoral collaboration and data sharing to address climate impacts on health.

Finding

Common barriers to tool implementation include lack of effective communication between modelers and decision makers, lack of personnel to maintain and operate systems, and difficulty in training new users to operate and update the tools. The lack of communication often leads to confusion about tool goals. We also identified a lack of information on climate-disease modelling tools for non-English speaking countries, increasing the inequities in the field. There is an opportunity to improve science communications, to produce information in multiple languages, and to build the capacity of local experts who are tool users.

Recommendation

Science communication needs

There is an opportunity to improve the communication between modelers and decision makers, such as the creation of useful public visuals that decision makers can use to create and enact evidenced based policy. Science communication experts and graphic designers should be incorporated into research teams. There is also an opportunity to translate available information on tools for non-English speaking countries.

Capacity building for tool users

There is a critical need to train and sustain the next generation of tool users, such as local experts in ministries of health, through training materials that are freely available in different languages.