Search strategy—Identifying relevant studies and national policies

We developed a search protocol which generated a list of possible search strings (related to climate-and water sensitive diseases in the south Asia context) used to locate relevant published literature. A pilot search was conducted in Web of Science Core Collection using a defined search-string, based on which iterations of search terms were developed (see S1 File). This was done using the authors’ expert knowledge and reference to literature. We adopted a broad search approach to allow for a comprehensive focus and retrieval of papers across different disciplines. No timeframe restriction was applied.

For the actual systematic search, the online bibliographic databases of PubMed, Web of Science Core Collections and Scopus were used to identify relevant peer-reviewed articles, given their wide scope of scientific publications and multidisciplinary contents [52]. In addition, the reference lists of randomly selected articles were manually searched using a “snowball” technique to identify any further literature that may have been missed in the first round of searches. This was repeated until saturation of the search had been reached. The key terms applied in the search are summarised in the accompanying excel scoring sheet (see S1 File). The search results were imported into Sciwheel reference management software (https://sciwheel.com) and duplicate articles were automatically removed.

Following the literature search, we conducted an online search of relevant national climate adaptation plans, health policies and hydroclimate and health surveillance systems of the focal countries. The keywords “climate policy”, “health policy or legislation”, “climate change adaptation policy”, “hydroclimate information systems”, “health surveillance systems” “climate-sensitive diseases” and “country” were used to search documents about policies and legislations of relevance to the subject-matter. The rationale for the policy mapping exercise was to establish the extent of national policy focus on climate-and-water sensitive diseases, and the opportunities for operationalising climate-driven early warning forecasting systems. This also afforded contextual insights into respective national government priorities and investment focus concerning tackling climate-and-water sensitive human disease risks as part of national climate-health adaptation planning. National policies in this context comprise the set of guidelines and legislative enactments adopted by the focal countries to support climate change adaptation, health and development planning. National adaptation policy thus refers to a formal national policy for identifying medium- and long-term adaptation needs and developing and implementing strategies and programmes to address them [53]. Source documents were collated from relevant government ministry and departmental websites for critical analysis and assessment of gaps. Both the peer-reviewed literature and policy searches were conducted between 2^nd January 2022 and 30th April 2022. A further peer-reviewed literature search was carried out (using the same keyword search terms) between 23^rd January 2024 and 1^st February 2024 based on which the corpus of relevant literature was updated.

Eligibility criteria.

After the systematic search, different inclusion and exclusion criteria were used to select peer-reviewed articles and policy documents for the review (Fig 2). Only electronically available, full text articles and policies written in English were considered. The identified articles were screened for eligibility in two phases. In the first phase, studies that did not report on forecasting systems linked to climate and/or water sensitive diseases and south Asia were excluded. If the article title and abstract information was insufficient relative to the inclusion/exclusion criteria, we retained the article in the Sciwheel database for further full text review. At the full text review stage, we further excluded articles that were inaccessible. In total 1,091 citations and 40 national climate and health policies and reports (latter are listed in Table 1) were imported to the Sciwheel database for the review.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Key national climate and health policies in the reviewed countries.

https://doi.org/10.1371/journal.pone.0309757.t001

The literature landscape

As illustrated in Fig 1, we identified 862 citations after removing duplicates, and 274 citations were selected for the abstract and full-text screening. After concluding the two-phase screening process, 187 articles published between 1992 and 2024 (120 studies between 2015 and 2024) were retained for the review, most of which were empirical studies (Figs 1 and 2A, n = 144). Of the 187 studies, 175 were country-specific and 12 covered the south Asia region (multi-country; Fig 2B). More than half of the selected studies (54.5% or n = 102 studies) came from study locations in India, followed by Bangladesh (20.9% or n = 39 studies), Sri Lanka (9.1% or n = 17 studies), Nepal (6.9% or n = 13 studies), and Pakistan (3.7% or n = 7 studies), as the top-most countries of research focus (Fig 2B). Bhutan (n = 2 studies) and Afghanistan (n = 1 study) represented the countries of the least research focus (for details of the papers, see Fig 2 and S1 Table (country distribution of included studies) in the Supplementary Information). The frequency of publications over time indicates that the number of climate-health studies from the south Asia region remained very low until 2012, followed by a steady increase, reaching maximum numbers in 2020 coincident with the publication of the Lancet 2020 seminal report dubbed the Lancet Countdown on health and climate change: responding to converging crises (Fig 2A).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Descriptive summary of included articles.

(A) For relevant abstracts, trends in publication over time indicate a continued increase in the volume of literature on climate and health in south Asia region. Literature published between 1^st January 1992, and 31^st January 2024 were included. (B) Study countries–each represented country is mentioned in at least one study. Pie charts show the number of publications by study design (C), model type for the modelling-based papers (D) and funding source reported (E).

https://doi.org/10.1371/journal.pone.0309757.g002

Concerning study design, we found a strong focus on studies adopting quantitative approaches (e.g. modelling) with 64.2% (n = 120), whereas 22.2% of studies (n = 43) employed qualitative-based (e.g. key-informant interviews, textual analysis), and 13.6% of studies (n = 24) mixed methods approaches (Fig 2C). This can be explained by the fact that most of the studies sought to quantify the degree of sensitivity of emerging infectious diseases to climate-related impacts. The flipside of this ‘quantitative’ focus, however, is the scarcity of interdisciplinary studies on the theme of climate-health, and therefore the understanding of complex socio-ecological and contextual drivers of vulnerability and adaptation to emerging climate-and water-sensitive diseases in the region. Of the 120 quantitative papers, slightly over half (52.5% or n = 63 studies) were modelling studies. Further disaggregation by modelling design indicates that 39.7% (n = 25 studies) of these 63 papers were space-time and temporal, with 28.6% (n = 18 studies) spatially focussed (Fig 2D). In terms of model type, nearly three-quarters of the modelling studies were correlative (71.4% or n = 45 studies) and 25.4% process-based (n = 16 studies), which is as expected since process-based models can only be parameterised if there is a sufficient information on the study system.

The reviewed papers were published in a diverse corpus of refereed journals, with 13 studies published in Malaria journal, 7 studies in Tropical Medicine & International Health, and 5 studies in each of Acta Tropica and Indian Journal of Medical Research. Further descriptive analysis of the citation bibliometrics indicate that 84 studies (representing 44.9% of the reviewed papers) involved international collaborations, meaning that one or more of the co-authors had their institutional affiliations outside of the reviewed countries. Of the 112 studies with funding disclosure statements, 58 studies (51.7%) were funded by external institutions based in Global North countries (US, Germany and UK) and supra-national institutions (World Health Organisation, United Nations Development Programme and World Bank) (Fig 2E). Most of the externally funded studies received support specifically from the Bill & Melinda Gates Foundation and joint WHO/TDR/UNDP/World Bank scheme. This perhaps suggests a future window of opportunity for increased funding support for more international collaborative research in the development of climate-driven early warning decision tools to support ongoing adaptation and health interventions in the region. Most of the funded studies were from Bangladesh and India (n = 30 studies apiece), Sri Lanka (n = 9 studies) and Nepal (n = 8 students). Pakistan, Afghanistan and Bhutan tended to lag in terms of international research funding in the climate-health theme, which perhaps is suggestive of priority contexts for future international funding investment. From a country perspective, we found 36.7% of the funded studies from India (n = 11 studies) were supported by the Indian Council for Medical Research (ICMR), the country’s flagship organisation for the formulation, coordination and promotion of biomedical research. A further cross-tabulation analysis of the authorship affiliation and publication year reveal a steady rise in the number of papers involving international collaborations from 2014 onwards (76 studies in total). The emphasis of climate change impacts on health in the IPCC 5^th Assessment Report and 2015 Lancet Commission Report on Climate and Health may have contributed to the growing research interest on the climate-health theme in the region.

Policy prioritisation of climate-and water-sensitive diseases

To understand the degree of representation or mention of climate-and-water sensitive diseases as a function of their prioritisation in policy, we reviewed 40 key national climate adaptation and health policies of the focal countries in the South Asia region. It emerged from our textual analysis that 30 out of the 40 reviewed documents had a (human) disease focus. The remaining 5 documents had no specific disease focus. Of the 30 documents with human disease focus, 45.7% (n = 14 documents) had an explicit mention of specific human diseases of interest in the respective countries. Most the documents were general in their identification, lacking specificity on the individual climate-sensitive diseases of priority in national adaptation planning and health interventions (Table 1). As an exception, Bangladesh’s Climate Change and Sustainable Development (2000–2050) specifically identifies cholera, diarrhoea, malaria, dengue and Lymphatic filariasis as the topmost diseases of policy interest. Similarly, India’s National Action Plan on Climate Change (undated) identifies malaria, visceral leishmaniasis, Japanese encephalitis, Lymphatic filariasis and dengue as the priority diseases of focus. With the notable exception of the National Health Policy 2009 (mentioning avian influenza, malaria and dengue as priority diseases of focus), relevant documents reviewed for Pakistan generally lacked clarity on specific priority climate-and-water sensitive diseases of interest. The federal system of governance in Pakistan has meant that prioritisation of specific diseases is often reflected in policy/programme documents at the sub-national level. The priority diseases climate-and water sensitive diseases of interest in the South Asia region were again reflected in the number of mentions in the policy documents (malaria (n = 14 documents), dengue (n = 9 documents), Lymphatic filariasis (n = 6 documents) and cholera (n = 5 documents), as was the case with the peer reviewed literature. The amount of evidence existing for the above-mentioned diseases, in terms of disease system understanding and societal impacts, may have influenced their translation into policy as a function of their importance in the respective countries. If this holds true, then it stands to reason that the more evidence available on the burden and impact of a specific disease or pathogen, the greater its visibility in the national policy arenas.

Priority climate-and water-sensitive diseases of interest

As evidenced in Fig 3A, most reviewed papers reported on malaria (n = 53 studies), Lymphatic filariasis (n = 46 studies), dengue (n = 41 studies) and cholera (n = 38 studies) as the topmost climate-and water sensitive diseases of research interest in the south Asia region. Of the 53 malaria focused studies, over two thirds of the papers (66.6% or n = 36 studies) had India as the focal country followed by 5 studies apiece conducted in Bangladesh and Nepal (Fig 3B). Papers related to Lymphatic filariasis were predominantly from India (n = 34 studies), Sri Lanka (n = 4 studies), Bangladesh (n = 3 studies) and Nepal (n = 3 studies) respectively. A significant proportion of the cholera-focussed studies were from Bangladesh (69.4% or n = 26 studies), reflecting the country’s high-burden status (Fig 3B). For instance, it is estimated that 66 million people are at risk of cholera in Bangladesh, with approximately 100,000 cases and 4500 deaths annually [58–61].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Priority climate and water sensitive risks to human health.

Bar chart shows the number of papers by disease focus (A), disease and country of focus (B), taxonomic group of reported diseases (C), and disease transmission routes (D).

https://doi.org/10.1371/journal.pone.0309757.g003

Contrasting disease focus and publication trends (Figs 2A and 3B), the results show that most of the papers related to priority diseases were published within the last eight years alone (i.e., 2015 and 2023). This highlights the view that scientific research focus on climate-health theme is still nascent in the region, at least from the standpoint of publications in international refereed journals. A case in point is leptospirosis-focussed papers, which for example, had 23 of the 24 related studies published within the last seven years alone (i. e. 2015–2022). The more intense research focus may be reflective of the leptospirosis policy and programme in India which started in 2015. It is plausible that there is a longer history of operational research work in respective government departments/ institutes in the region that is not represented in the peer-reviewed literature, in which case further exploration through policy interviews would be instructive. On the taxonomic classification, the reviewed papers predominantly focussed on protozoa (n = 66 studies), followed by viruses (n = 56 studies), parasites and bacteria (n = 44 studies respectively). Most of the papers reported on vector-borne diseases as the topmost climate-sensitive diseases in the region, which mirrors the common trend in the zoonoses scholarship, where most research continues to focus on climate change effects on the distribution and spread of vector-borne infectious diseases. Interestingly, the results show an under-representation of important, neglected climate-sensitive infectious vector-borne diseases (e.g. chikungunya and visceral leishmaniasis) in terms of scientific interest despite their wide-ranging societal impacts, particularly in the rural contexts of the South Asia region. Juxtaposing the disease focus of peer-reviewed papers and policy focus appear to give the indication that there is an overemphasis on known and epidemic climate-sensitive diseases (e.g. malaria, cholera) to the relative neglect of otherwise endemic high burden neglected diseases (e.g. leptospirosis, chikungunya) that disproportionately affect poor and vulnerable communities in the region.

Environmental drivers and disease focus associations

To the extent that the dynamics of disease distribution and spread is likely to be affected by environmental change factors, it was instructive to identify the range of climate and non-climatic drivers considered in the reviewed studies. Of the studies that considered climate drivers (n = 109 papers), 59 (representing 54.1%) studies analysed 1 or 2 climate drivers with the remaining 50 studies reported on 3 or more drivers (Fig 4C). The climate drivers of disease (hydrometeorological hazards) considered in these studies included extreme weather events, seasonal rainfall, temperature, drought and wind. The most frequently addressed climate drivers were rainfall 65.1% (n = 71 studies), temperature 60.6% (n = 66 studies), hydrology (e.g., soil moisture, drought, flood extent and waterbodies) 55.1% (n = 60 studies) and climate change 48.6% (n = 53 studies) (Fig 4B). Extreme weather events (e.g., heatwaves, cyclones, and flash floods) 17.4% (n = 19 studies) and wind (mostly storm events) 5.5% (n = 6 studies) were the climate drivers least often examined as linked to focal diseases reported in respective studies (Fig 4B). The relatively limited focus on extreme weather events as a known climate driver of focal diseases reported, could be indicative of the dearth of research evidence on related drivers linked to disease burdens and/or transmission pathways in the region. It may also reflect the fact that it has been relatively easier for modellers to access input data on mean climatology than extreme events. Further cross-tabulation of climate drivers and disease investigated showed that the topmost diseases associated with seasonal rainfall pattern as a climate driver were malaria (n = 33 studies of 66), dengue (n = 28 studies) and cholera (n = 8 studies). Similarly, these diseases (malaria, dengue, and cholera) were frequently associated with temperature and climate change (see Fig 6).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Climate drivers of diseases reported in the reviewed studies.

(A) frequency histogram of the climate drivers considered in reviewed studies; (B) stacked bar chart showing the proportion to which focal diseases are sensitive to different climate drivers, broken down by disease taxa (note: More than one climate driver can be associated with a disease; (C) frequency of climate drivers reported; D, E, F & G represent frequency histograms of the number of climate drivers (0–6 drivers) considered in studies that focussed on viruses, bacteria, protozoa, and parasites, respectively.

https://doi.org/10.1371/journal.pone.0309757.g004

For most studies of viruses, bacteria and protozoa, either 1 or no climate driver was considered within the study (Fig 4d–4g). The highest number of climate drivers (n = 6) considered for an individual disease was for malaria (protozoa, n = 2 studies), leishmaniasis (protozoa parasite, n = 2 studies), and dengue (viral, n = 2 studies), whereas Japanese encephalitis virus and avian influenza virus were each reported in single studies [62,63]. One cholera-related study had 5 associated climate drivers [22].

The non-climatic drivers examined in reviewed studies included altitude, vegetation, and land use change (n = 49 studies). Land use change and vegetation were the non-climatic drivers most often considered in studies that considered these types of drivers (Fig 5A). Cross-tabulation of non-climatic drivers and disease investigated revealed that malaria (n = 12 studies of 22), dengue (n = 8 studies) and cholera (n = 3 studies) were the topmost diseases associated with land use change as a non-climate driver (also see Fig 6). The highest proportion of non-climatic drivers (n = 3) addressed in respective studies was for protozoan-related diseases (malaria) (Fig 5F).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Non-climatic drivers of diseases reported in the reviewed studies.

(A) frequency histogram of non-climatic drivers reported, (B) stacked bar chart showing the proportion to which focal diseases are sensitive to different non-climatic drivers, broken down by disease taxa (note: More than one non-climatic driver can be associated with a disease), (C) frequency histogram of number of non-climatic drivers considered in reviewed studies, D, E, F,& G frequency histograms of the number of non-climatic drivers (0–3 drivers) considered in studies that focused on viruses, bacteria, protozoa and parasites, respectively.

https://doi.org/10.1371/journal.pone.0309757.g005

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Cross-tabulation of assignment of papers to a given environmental and/or social driver and disease lens in the South Asia region.

Displayed are selected environmental change and socio-demographic drivers (mentioned more than 5 times across the reviewed papers) in a matrix of disease lens in each reviewed paper. Disease lens and climate and non-climatic drivers are sorted to frequency of inclusion/consideration in respective studies (i.e., studies from the general climate change and rainfall driver lens respectively are most frequently focussed on malaria and dengue). Note: EWEs = Extreme Weather Events; * = soil moisture, drought, flood & waterbodies.

https://doi.org/10.1371/journal.pone.0309757.g006

Only 34 studies (18.2%) considered both climate and non-climatic environmental drivers, out of which 14 were modelling focussed. While the limited consideration of both climate and non-climatic drivers within studies is suggestive of data gaps on specific drivers, it also underscores the relevance of cross-sectoral information exchange, for example from hydroclimate and health, in better prioritising and planning of appropriate disease control interventions (see Section 3.4). Models describing the impact of climate change on infectious diseases considering a narrow range of climate drivers known to affect diseases may risk large uncertainties in comprehensively understanding and projecting how climate change will affect these same diseases in the future. This lends support to McIntyre et al. [55] observation about the urgent need for disease risk models, to examine a combination of climate drivers and multiple scenarios for changes in climate impacts [64].

A small proportion of all the reviewed studies (28.3% or n = 53 studies) identified socio-demographic factors as drivers of disease emergence or spread. For example, the studies that reported on social drivers of disease transmission focussed on malaria (44.4% or n = 20 studies), dengue (33.3% or n = 15 studies), cholera (17.8% or n = 8 studies) and Lymphatic filariasis (13.3% or n = 6 studies). Of the 45 studies that considered social drivers, 73.3% of them (n = 33 studies) also included climate and 46.7% (n = 21 studies) non-climatic drivers. The limited focus on the interplay of socio-demographic factors and disease dynamics (underscored by the difficulty of obtaining socio-demographic data at the same resolution as health data) perhaps highlights a strategic research priority for better understanding the biosocial factors that shape vulnerability to climate-sensitive disease risks in the region.

Forecasting capacities to address climate-driven infectious diseases transmission and spread: Taking stock

To better understand the forecasting capacities, priorities and needs for targeted early warning and anticipatory response, it was imperative to investigate the diversity of existing national hydroclimate and disease surveillance systems available to stakeholders in the region, including sectoral focus, scale of operationalisation and priority diseases being addressed. From the analysis of the reviewed papers, only 36 studies reported on national/state level health surveillance or information systems, with just 3 studies mentioning the existence of integrated surveillance systems (i.e. linking animal and human health information) for disease prevention and control [65–67]. None of the reviewed studies reported on linkage of health and environmental information systems to inform disease management in the reviewed countries. This indicates limited information systems in the reviewed countries, possibly reflecting the relative dearth of evidence on forecasting systems. It therefore follows that further investigation through targeted key informant interviews and workshops remains necessary to better understand and assess forecasting capabilities and opportunities to link information and models across sectors.

The evidence for existing hydroclimate and disease surveillance capacities in the reviewed countries is summarised in Fig 7. To better understand the needs for hydroclimate surveillance for different diseases, we used identified links between climatic drivers (different hydrometeorological hazards) and diseases outlined in Mora et al [10] with further detailed results available in S3–S6 Tables. Whereas some of the identified information systems have yet to be fully operationalised and thus are not mature to assess in terms of impact, they present a promising outlook and a fair representation of the existing capabilities and avenues for improvement, particularly in terms of strategic emphasis, sectoral focus and diseases being prioritised for interventions (S3–S6 Tables). As evidenced in Fig 7, our inventory of existing information systems indicate that the largest number of hydroclimate and health information systems are found in India (n = 18 platforms), followed by Bangladesh (n = 4 platforms) and Pakistan (n = 3 platforms), which somewhat reflects their socio-economic capacity (as upper middle-income nations) and investment priorities. Afghanistan and Sri Lanka had the least number of information systems (1 apiece). In Afghanistan, the prevailing political tensions, which has gravely impacted on national infrastructural investments including health and metrological systems [61,68–70], may provide part of the explanation for this lack of evidence. Besides, the restricted composition of the WHO South Asia region (excluding Afghanistan and Pakistan) may impact on broader regional cooperation in terms of information exchange and capacity building. Moreover, almost all the identified information systems (91.2% or n = 31 platforms) had national orientation and a seemingly limited disease focus, suggesting that few of the countries have their own surveillance systems to monitor and track the outbreak of any new diseases.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. Summary of existing hydroclimate and disease surveillance capacities.

Panels show all potential drivers (hydrometeorological hazards) and links with priority diseases obtained from Mora et al. [10] in grey. Orange nodes indicate that information is available for a driver or disease from existing information systems and orange links highlight where information may be available to explore the relationship between a driver and disease. Green links highlight instances where a link between a driver and disease was explored in a reviewed study and green nodes highlight instances where national information systems have fed into a reviewed study.

https://doi.org/10.1371/journal.pone.0309757.g007

In terms of cross-sectoral implementation, a large majority of the information systems are sectorally driven in terms of operationalisation, which poses a risk for late detection and cross-sectoral engagement with preparedness and response infrastructure. To the extent that most of the identified hydro-climate systems utilise hydro-metrological data highlights a potential for use in establishing relationships with health surveillance and information systems, where corresponding disease data exist. In this sense, it is imperative that hydro-climate forecasting outputs are available on timescales relevant to disease surveillance. Altogether, the foregoing highlights varying capabilities and sectoral representation, underscoring the need for crosscutting integrated efforts (through the development of cross-sectoral early warning systems and tools) to tackle risks and vulnerabilities as well as support long term adaptation planning, particularly in highly vulnerable countries like Sri Lanka, Bhutan and Bangladesh.

Developing climate-driven early warning systems to support disease adaptation and resilience building interventions—A research agenda

Hydro-climatic and disease information systems are increasingly recognised as vital for successful adaptation in the public health sectors in the face of climate change, through providing timely and tailored information to support targeted decision-making and resilience building interventions in highly vulnerable regions [49,71–73]. While conventional sectorally-driven hydrological, climate and disease early warning systems are useful, they remain inadequate for comprehensive capture of complex disease dynamics that straddle the human, animal, and environment interface. This necessitates cross-sectoral efforts to develop integrated climate-driven early warning systems that leverage existing meteorological and health surveillance capacities across scale. Yet, much of the research and efforts toward improving the delivery and use of hydroclimatic information and disease surveillance capabilities in the South Asia region has at best been fragmented and sectorally focussed [66,73–77]. The operationalisation of the One Health approach (which recognises the interconnectedness of human health, animal health and the environment, and advocates for integrated, holistic and transdisciplinary approaches to reduce disease impacts) is seen as a plausible vehicle for developing integrated solutions that support disease control interventions at scale [20,50,78].

In this paper, we considered the question of what is needed to develop integrated climate-driven early warning systems for climate-and water-sensitive diseases to support climate-health, health adaptation, and risk management needs at scale. Consistent with the notion that cross-sectoral collaboration is needed to better inform effective climate-health interventions, we argue that there is a need for a paradigm shift in research focus from a single pathogen/disease focus to multiple disease stressors, along with their burdens and impacts, to support the development and targeting of early warning systems. To our knowledge, this scoping review is the first of its kind to characterise the state of the art on climate-health scholarship in the South Asia region, highlighting the priority climate-and water-sensitive diseases and existing cross-sectoral information systems that could be linked in the future to better understand and predict how climate variability and climate change precipitate health impacts. We found that there exist marked differences in research foci in the region and existing information systems capabilities, with overrepresentation of evidence from India and Bangladesh compared to the lesser developed and highly vulnerable countries (Afghanistan, Sri Lanka and Bhutan), which are lagging in terms of research, and hydrological and surveillance infrastructural capacity. In Afghanistan, for example, research evidence on climate-health linkages is largely focussed on vector-borne diseases (malaria and dengue) [21,61,69,79,80], despite the potential climate change impacts on cholera, which is a priority epidemic disease of concern (128,278 cases reported between 2000 and 2017) in the country [81].

We thus put forward five directions for future research and operationalisation.

1. To guarantee more robust evidence on disease burdens and impacts, we need further research on priority climate-and water sensitive diseases particularly in least represented and data poor contexts such as Afghanistan, Sri Lanka, and Bhutan. As evidenced above, there is a disproportionate representation of research focus on India and Bangladesh which has meant a dearth of evidence on other countries to support long-term adaptation planning. Nonetheless, it is noteworthy that poor national security in some of the focal countries (e.g., Afghanistan) hampers the ability to undertake fieldwork to assess adaptation options and requirements [82].
2. Studies focussing on the social determinants of climate impacts on health are underrepresented among the topic areas in the literature on the South Asia region. Given that patterns of vulnerability and adaptation to climate induced disease risks and associated impacts manifests within dynamic socio-cultural and political contexts, it is critical to understand the context of at–risk populations, drivers of their vulnerability, and adaptive strategies to better inform the contextualisation and tailoring of early warning support systems for bolstering resilience and long-term adaptation. This could be delivery of fine scale vulnerability maps and risk information that integrate understanding of social vulnerability and improves targeting of interventions.
3. Better understanding of the different information needs, opportunities, and barriers for cross-sector information-sharing and analytics is needed across the region to support the development of locally appropriate climate-driven early warning systems for meaningful adaptation planning at the relevant scales. The paucity of studies assessing cross-sectoral forecasting and surveillance capacities and requirements calls for better inclusion of underrepresented countries (Afghanistan, Pakistan, Sri Lanka, and Bhutan) in future research. International funding support to address these questions, particularly in the underrepresented countries, and strengthening regional research networks could be a good starting point.
4. The non-accessibility of data and/or barriers to information-sharing between cross-sectoral actors remains a key challenge, as data is treated as proprietary. This calls for greater sensitisation of data owners and curators, and disciplinary experts in climate and epidemiology, about cross-sectoral information-sharing and improved mechanisms for fostering these and needs for developing integrated early-warning systems and decision support tools to better inform disease management and long-term adaptation planning, learning from demonstration cases in other contexts.
5. There is also a need to create improved information services in which producers and end-users of early warning information and decision-support tools interact to identify needs and priorities for reducing vulnerabilities and strengthening resilience to climate induce disease risks [50,83]. This suggests a need-based tailored information service for at-risk populations in the region. In doing so, co-production approaches remain an important aspect of tailoring hydroclimatic and disease information systems to stakeholder needs and priorities.

Notwithstanding the breadth and the wide regional coverage of the evidence reviewed, this scoping review is subject to some limitations. First, scoping reviews have an inherent limitation due to the focus on the breadth rather than depth of information on the subject-matter. As such, meta-analysis of the evidence base (including quality appraisal) is generally not conducted in a scoping review [84]. Second, owing to time and COVID-19-related constraints, it was not possible to engage directly with stakeholders about their needs and priorities as well as to ascertain capabilities and data access to enable a joined-up approach through key informant interviews or multi-stakeholder workshops. Third, the reliance on English-language peer-reviewed papers and national climate adaptation plans and health policies might have resulted in missing some relevant policy documents (e.g. regional/state policies) and papers published in restricted local journals and/or local languages. Moreover, although we focussed on policies current at the time of review, we cannot dismiss the possibility that we missed new documents or that a policy was updated before the review was completed. This scoping review nonetheless showcases some critical patterns in existing climate-health scholarship and related forecasting and surveillance systems in South Asia which have resulted in a number of key actionable steps for developing integrated climate-driven early warning infrastructure towards supporting long-term adaptation planning and interventions.