Final Report Summary - HEALTHY FUTURES (Health, environmental change and adaptive capacity: mapping, examining and anticipating future risks of water-related vector-borne diseases in eastern Africa)
Executive Summary:
HEALTHY FUTURES (Health, environmental change and adaptive capacity: mapping, examining and anticipating future risks of water-related vector-borne diseases in eastern Africa [HF]), a FP7 Cooperative project, aimed to build a disease risk mapping system for three water-related high-impact vector-borne diseases (VBDs) (malaria, Rift Valley fever (RVF) and schistosomiasis) in eastern Africa. Projections of future risk were based on a combination of environment, notably climate, change data, output from dynamic disease models and spatially explicit assessments of vulnerability.
The major successes of HF include:
1) Development of a publically accessible Geoportal which facilitates searching and use of data produced in the project.
2) Production of new, state-of-the-art dynamic disease models for malaria, RVF and schistosomiasis. A second dynamic disease model for RVF has also been developed through the project that allows the efficacy of different interventions (e.g. vaccination campaigns) to be examined. These models have either been published or are in the process of being published in scientific journals.
3) Release of the HF Atlas, an interactive, web-based mapping and decision support tool (DST), built within an open-source framework and aiming to provide meaningful and guided access to information on environmental change, disease hazard and risk, and vulnerability for the three target VBDs in eastern Africa. An important aspect of the Atlas is use of climate models and future scenarios referred to in the latest (the 5th) IPCC Assessment Report.
4) Development and release of Decision Support frameworks (DSF) for the three target diseases in consultation with relevant decision makers in the East African Community (EAC) region.
5) Four PhDs, including the first ever PhD to be awarded in Science by the University of Rwanda (UR).
6) Extensive dissemination and exploitation of results through various channels including: 1) HF website receiving ~40,000 hits to date (http://www.healthyfutures.eu/) 2) seven contractual HF news releases, 3) project factsheet, 4) seven HF newsletters, 5) four project stakeholder updates, 6) a non-technical tri-fold project brochure, 7) inclusion in EC-produced promotional material and websites, 8) a short promotional video (www.vimeo.com/70318624) 9) two promotional items (a key-ring with bottle opener and torch, and post-it notes to increase the brand awareness), 10) contribution to a short documentary “Health and climate change in Africa” focusing on the collaborations between HF and its sister project, QWeCI - produced by the charity, Africa Turns Green (https://www.youtube.com/watch?v=oYL4Nc-qnKE) 11) HF Symposium (jointly coordinated by members of HF and QWeCI and held at the 4th Annual East African Health and Scientific Conference in Kigali, Rwanda in March 2013), 12) HF-focused and organised/hosted session at the ‘Impact of Environmental Changes on Infectious Disease’ (IECID) Conference, Sitges, Spain in March 2015, 13) 19 papers published in scientific journals with one more awaiting decision. Two submissions in a WHO-WMO special publication on climate services and health. One special issue of the journal Geospatial Health is currently in preparation, 14) oral presentations of HF results and networking through attendance and presentation at many key conferences and meetings and 15) inclusion as a case study in the World Health Organization/the World Meteorological Organization (WHO/WMO) special publication on Climate Services for Health.
With its integrated holistic approach, HF has considerably enhanced current understanding and dynamic modelling of links between the three target, water related VBDs and environmental change drivers and effects. As stated in the external evaluator’s report commissioned by project members, “HF has been an ambitious project that successfully joined the divergent scientific interests of climate specialists and health practitioners in a single project to address the health impacts of climate change on three VBDs in the EAC. It has provided key scientific information and models on present and future climate impacts on public health and the possibility of further development of the Atlas-Metadata portal as a practical research and learning tool for stakeholders in the EAC.”
Project Context and Objectives:
HEALTHY FUTURES Context
The health effects of future environmental, including climate, changes have been projected to be substantial, often negative and to vary geographically. The effects will be felt most acutely among the poorest members of society, who already carry a disproportionately high share of the burden of environmentally sensitive diseases, with sub-Saharan Africa a focus of adverse health impacts. Environmental change also has implications for health in the developed world, as changes in environment drive the emergence of new diseases, and changes in the distribution and epidemic potential of existing infectious diseases. Environmental change will impact health in a multitude of ways. Concern has, however, tended to focus on the future distribution and spread of infectious diseases, and in particular the negative health impacts of changes in transmission and outbreaks of VBDs as a result of anthropogenic climate change. The impacts may be direct, in terms of outbreaks of disease among human populations, or indirect, in the form of outbreaks of diseases that affect domesticated animals or plants, and therefore jeopardise food security, agriculture-based economic activities and trade.
These concerns provided the motivation for HF, which aimed to build a disease risk mapping system for three water-related high-impact VBDs (malaria, RVF and schistosomiasis) in eastern Africa, accounting for environmental/climatic trends to project future risk. The project involved a comprehensive, inter-disciplinary consortium of health, environment, socio-economic and climate experts in addition to governmental health departments, and throughout the four years of EU FP7 funding, sought and obtained input from stakeholders, including health decision makers. Concentrating on the three target VBDs in eastern Africa, HF produced improved understanding of the sensitivities to environmental conditions, new state-of-the-art dynamic disease models, down-scaled environmental change model outputs and scenarios of future conditions from the IPCC’s 5th Assessment Report (2013 & 2014), and a novel, interactive, online disease risk-mapping platform (the HF Atlas).
HEALTHY FUTURES Objectives
The main aims of HF were to develop (1) a basis for anticipating future environmental changes and their impacts on water-related VBDs in eastern Africa, and (2) the capacity of health and veterinary services in the study area to respond to early warnings of future outbreaks. These aims were met through seven, interlinked Work Packages (WPs) that:
1) provided effective management of the research (WP1);
2) ensured that full benefits are derived from synergies between this research project and other related research projects (WP7);
3) developed policy-relevant simulations of future levels of environmental correlates of the three VBDs and mapped spatial variations in vulnerabilities across the region (WPs 2,3,4);
4) combined results of the work outlined in 3) with improved understanding of the epidemiology of the targeted VBDs to model, dynamically, VBD emergence and spread and to improve (constrain) assessments of risks of future outbreaks (WPs 3,4);
6) enhanced the capacity of health and veterinary services at various geographic scales to respond to changes in risk of transmission and outbreaks of the targeted VBDs as a result of a convergence of changing climatic, land use and socio-economic conditions (WP5);
7) ensured that the findings of the research are widely disseminated among stakeholders, policy-makers and the global scientific community (WP6).
One further activity of crucial importance was the training of early stage researchers in the trans-disciplinary fields of environmental change and animal and human health. As part of this, four students (two from Africa and two from Europe) obtained their PhDs through direct support from HF.
Activities carried out per workpackage are detailed below:
WP1 facilitated coordination and management of the project. During M37-48 there were two partner meetings. The fifth partners’ meeting was held from 26-27 February 2014 in Nairobi, Kenya and was hosted by the International Livestock Research Institute (ILRI). The sixth partners’ meeting was held from 9-10 September 2014 in Salzburg, Austria and was hosted by the Paris-Lodron University of Salzburg (PLUS).
WP2 largely involved construction of a project database (‘geoportal’) comprising information drawn from a range of primary and secondary sources, including: historical, socio-economic, migration, settlement and conflict data; earth sciences data; direct climate observations and the results of regional downscaling of global reanalyses; and information on disease-environment relationships. Geoportal is at: http://41.204.190.50/geoportal/catalog/main/home.page
WP3 focused on field-based evaluations of environment and hydrological data constructed in WPs 2 and 4, respectively, with the collection and collation of disease vector/host information, and with the development of new dynamic models and statistical multivariate regression models for infection rates of the three VBDs. Outputs from these models were input to WP4 as part of disease risk mapping. Two PhD theses were also linked to WP3: 1) ‘Modelling the effects of temperature changes on Schistosoma mansoni transmission’ by Nicky McCreesh at UDUR; and 2) ‘A simulation model of Rift Valley fever transmission in Kenya’ by John Gachohi at ILRI/University of Nairobi.
WP4 collated information relating to historical disease drivers and occurrence (collected in WP2), along with knowledge and models of dynamics of the three target VBDs (WP3). Spatial assessments of the current and projected future hazards and risks of the three target VBDs in eastern Africa were combined with a spatially explicit database relating to historical outbreaks with the aim of identifying hotspots of high risk. Projections were closely linked to procedures adopted in the IPCC’s 5th Assessment reports. WP4 focused on two scales: the region comprising the five East African Community (EAC) countries and on smaller sub-national scales. The assessments of hazard and risk informed WP5, through input to discussions on DSF carried out with other consortium members and stakeholders in the findings of the research. The HF Atlas (http://zgis186.geo.sbg.ac.at/hf_atlas/) an interactive, web-based mapping and visualisation tool, was built largely as part of WP4. Two PhD theses were also linked to WP4: 1) ‘Climate change and malaria in Rwanda: Spatial assessment of social vulnerability at different scale levels’ completed by Jean Pierre Bizimana at UR (formerly NUR); and 2) ‘Integrated spatial indicators for modeling, exploring and visualizing vulnerability to vector-borne diseases’ completed by Michael Hagenlocher at PLUS.
WP5 was concerned with identifying and deploying the most appropriate environmental change adaptation strategies and decision support approaches (including relevant tools and information platforms) for adoption in the project. As part of this, input from decision-makers associated with the VBDs in the EAC area was obtained through a series of stakeholder engagement workshops hosted in the study region by HF consortium members.
WP6 focused on the engagement of stakeholders in the research, the effective dissemination of research findings, and the training of Early Stage Researchers involved in the project. As part of this WP, HF led very successful sessions at the ‘4th Annual East African Health and Scientific Conference’ Kigali, Rwanda (March 2013), and at the ‘Impact of Environmental Changes on Infectious Disease’ Conference, Sitges, Spain (March 2015).
WP7 promoted synergy, cross-fertilisation, networking, and coordination with other climate and health-related research projects funded by the EU FP and other non-EU stakeholders and researchers internationally and regionally. The WP also provided independent, external oversight, results review and evaluation, assurance of quality-control, guidance on next steps and use of best practices, including highest compliance with ethical guidelines. The WP operated mainly through an Expert Review Panel (ERP), chaired by an expert in the field and comprising four other members who were external to HF.
Project Results:
WORK PACKAGE 1 - Project Management
Lead beneficiary: AquaTT
All deliverables in WP1, which ensured the efficiency and effectiveness of the work performed within the project, were successfully developed and submitted. Six formal partnership meetings took place throughout the project duration.
WORK PACKAGE 2 - Disease information and database construction
Lead beneficiary: Trinity College Dublin
WORK PACKAGE 2 - Disease information and database construction
Lead beneficiary: Trinity College Dublin
Task 2.1 Collection of historical data
This task involved an investigation of past outbreaks of the three target diseases (malaria, schistosomiasis and RVF) in the eastern African study region (Burundi, Kenya, Rwanda, Tanzania and Uganda). This was principally achieved through archival research involving a wide range of primary sources, including colonial reports, private papers and ministry files. Historical archival data were collected from documents held in the National Archives (London), Rhodes House (Oxford), the Wellcome Unit for the History of Medicine (Oxford), the School of African and Oriental Studies (London), the Kenya National Archives (Nairobi), Ministry of Public Health and Sanitation (Nairobi), the Uganda National Archives (Entebbe) and the National Archives of Tanzania (Dar es Salaam) (most of the data collected from the Tanzania National Archives in 2012, however, was unfortunately lost due to the theft of a laptop). The collected data were coded, compiled in an Access database and visually displayed in Google Earth. The resulting historical database is incorporated in the HF Atlas, publically accessible through the project’s webpage. Thus this information can be, and is being, used to examine the historical framework of outbreaks of the target diseases in the study region in environmental, social and political contexts.
The HF Atlas is at: http://zgis186.geo.sbg.ac.at/hf_atlas/
Task 2.2 Collection of socio-economic data
This task gathered information on current policies, regulations, strategies and challenges in addressing outbreaks of the three target diseases and on other socio-economic data thought to affect, and be affected by, the transmission and maintenance of the diseases in the study region (e.g. present-day migration, poverty, settlement patterns and settlement densities, including informal settlements). In this context the task has generated knowledge (baseline information) on the extent to which outbreaks of malaria and RVF are a reflection of the socio-economic conditions in Ijara sub-county. Among the key findings in relation to malaria are that health education appears to be improving awareness of the benefits of owning mosquito nets in general, and insecticide treated nets (ITN) in particular. However, bednet control programmes are still grappling with distribution, and affordability and equity issues, as well as the trade-off between social marketing and commercial approaches. In relation to RVF, the main barriers restricting the abilities of health and veterinary services to respond to warnings of heightened disease outbreaks were identified. With respect to the vaccination of animals, which is one of the primary methods of preventing RVF outbreaks and the spread of the disease to humans, coverage is not universal. An important finding of the survey was the diversity in the frequency of vaccinations, and the leading reason for non-vaccination was non-affordability. A report based on all findings has been written and these results can better contextualise the socio-economic conditions of the diseases to enable the implementation of more effective response mechanisms in the framework of changing environments.
Task 2.3 Development of land cover/land use/terrain/surface water databases
Under this activity, spatially-referenced data on land cover and elevation, soil types, normalised difference vegetation indices for the period 1999 – 2010, livestock, and precipitation for the entire eastern Africa region were processed by ILRI and used to map the risk of RVF. Additional data on surface hydrology and livelihood zones were compiled only for Kenya.
Under this task, NUR also compiled spatially referenced data on land cover and surface hydrology, as well as other secondary data on land use, elevation and population in eastern Africa, with a focus on Rwanda. These data have been used in the examination of malaria.
Furthermore, all the datasets produced and collected as part of this task were uploaded to an online database which was created within this task. The online database comprises data collected, quality assured and collated within WP 2 as well as other WPs of HF. It includes metadata (Task 2.7) of the produced output datasets as well direct access to data via a File Transfer Protocol (FTP). Where restrictions apply, relevant contacts are identified. Initially only accessible to members of the HF consortium and to non-consortium members of the project’s Executive Committee, the information platform was made open access at the project end and it thus facilitates the dissemination of, and public access to, results and data produced in the project.
Task 2.4 Developing climate databases and regional climate simulations
Four key disease models have been developed entirely or partly within HF and they have various data input requirements. These were provided by both externally generated datasets (e.g. satellite and ground based rainfall observations, land use change projections, global climate model integrations) and by new datasets created within the project research (such as regional climate model simulations, newly downscaled global model runs, and high resolution simulations of surface hydrological conditions). Task 2.4 was concerned with the generation of these datasets, the key ones being the high resolution regional climate simulations.
The two participating institutions (SMHI and ICTP) produced two sets of climate simulations at high resolution. Given the computational requirements and the availability of validation datasets at 10km, a lower resolution target than originally envisaged was set. In this regard, ICTP used its regional climate model (RegCM4), driven by reanalysis, to generate data at 24km resolution, covering an area between 200S to 200N in latitude and 60E to 650E in longitude. SMHI produced a simulation at a resolution of 0.15 degrees (~ 17km) covering an area between 130S to 130N in latitude and 180E to 600E in longitude. The model domain covers all the HF target countries and includes the important surrounding regional features that are known to affect the region's climate, such as the western Indian Ocean and the Congo basin. The model outputs were compared against Climate Research Unit (CRU) observations and ERA-Interim reanalysis to examine how well the model reproduces the spatial, inter-seasonal and inter-annual variability of the region’s climate. The results are reported in deliverable D2.2.
In addition, the project applied a statistical downscaling technique to CMIP5 climate models in order to provide a wider range of lower-resolution benchmark driving fields. Self-Organizing Map based Downscaling (SOMD) is a leading empirical downscaling technique for Africa and provides meteorological station level or gridded data in response to global climate change forcing. The downscaling of a global climate model (GCM) is accomplished by deriving the normative local response from the atmospheric state on a given day (predictors), as defined from historical observed data (predictants). The method recognises that the regional response is both stochastic as well as a function of the large scale synoptics. As such it generates a statistical distribution of observed responses to past large scale observed synoptic states. This statistical downscaling methodology was used to downscale climate change projections from the latest two versions of the Coupled Model Intercomparison Projects (CMIP3 and CMIP5) over the HF study area.
Finally, research was also conducted to produce gridded socio-economic data for use in the dynamical disease modelling, in particular the assessment of population growth on fine-spatial scales. Total population estimates for the baseline period 1950-2010 were produced, and for five different future Shared Socio-Economic Pathways (SSP) for the period 2011-2100. Population estimates for the baseline period and each of the five SSP scenarios were obtained at the national level at five year intervals (from 1950-2010, and from 2015-2100 respectively). Data for the baseline period were retrieved from the United Nations World Population Prospects (UNWPP) 2012. Revision Population data for the five SSPs were obtained from the Inter-Sectoral Impact Model Intercomparison Project. Yearly population estimates were computed for each country using a linear interpolation method. Population change rates were used as scaling factors to compute gridded population data for both the baseline period and the five SSPs. Not all countries present in the gridded dataset were present in the UNWPP and SSP data. The 37 countries not present in these datasets were assigned an average scaling factor per year by taking the mean of the population change rate of the 193 countries in the UNWPP and SSP datasets.
Task 2.5 Bringing Climate Databases to the kilometre-scale for use in disease analysis
According to the original Description of Work (DoW) temperature and surface hydrology data, at ultra-high (1-10km) resolutions were to drive the Liverpool malaria model (LMM). In the course of the project development, however, a decision was reached to develop these modules within the disease models themselves to allow direct and computationally efficient interaction with other components of the disease models. In this way the task accomplished the direct incorporation of the surface hydrology within the new dynamical disease model VECTRI, and also directly downscaled the temperature to the high resolution topography.
The development of the new VECTRI model (a model not envisaged in the DoW) within the project framework allowed the integration of the surface hydrology and temperature downscaling directly into the model design from the outset. As a result, the consortium and scientists more widely now have access to a disease model that explicitly incorporates the direct interaction with the human population, allowing future population growth and urbanisation to be accounted for in disease risk projections, and that also directly incorporates the surface hydrology in a pond parametrisation.
The surface hydrology component has two parts:
1. A fraction of breeding sites associated with permanent water bodies. Initially this was set to a constant, but during the project several methods were investigated such as the use of existing databases including the global lakes and wetlands database of the WWF.
2. A pond parametrisation, which is driven by rainfall (available on scales down to 10km). The pond parametrisation is described in Tompkins and Ermert (2013).
The pond parametrisation is a novel aspect of the VECTRI model that allows for a more realistic association between rainfall and vector densities. The pond parametrisation accounts for topography in the maximum ponding fraction. This implies that while the input rainfall data do not have information below the 10km scale, the ponding fraction can be provided on scales down to 1km accounting for topography. The pond parameterisation has since been validated in comparisons with AMSR-E satellite derived statistics and also ultra-high resolution (10m) simulation with an explicit surface hydrology scheme. The magnitude of the pond fraction is much smaller for VECTRI indicating the need to tune the wmax parameter appropriately with observations, or the need to include soil texture in the parametrisation (see Tompkins and Ermert, 2013). Nevertheless, it is extremely encouraging that the VECTRI parametrisation scheme at 1km is able to almost exactly reproduce the pond temporal evolution. VECTRI tends to over-predict the pond fraction during the rainy season onset, and under-predict it during the main season. This is due to the fact that VECTRI does not account for soil moisture in the calculation of infiltration, which is apparently a second order effect.
The temperature information is also downscaled to the model resolution, up to a finest resolution of 1km as per the DoW, again using the topography to drive a simple lapse rate adjustment scheme. In the earlier version of VECTRI (up to v.1.3.1) the pond temperature is simply related to the atmospheric mean temperature with a fixed offset. This will be updated post project, with the implementation of an optional energy balance that has been developed and tested with in situ pond temperature measurements. The energy balance model is optional since it relies on the availability of radiation fluxes from the driving climate model, which can be subject to considerable biases associated with errors in the cloud fields, and which are not always archived on a daily timestep in the output of archives such as CMIP5. Thus, the option of using a simple offset from atmospheric 2 metre temperatures will be retained.
Task 2.6 Compilation of disease data
Data on RVF outbreaks in Kenya for the period 1912-2010 were obtained from the Department of Veterinary services and Centres for Disease Control (Kenya). These data were used by ILRI to develop Milestone 16 and have also been analysed and presented in various workshops as part of the QWeCI (HF’s sister project) research reports. These records identify the years when outbreaks occurred and include administrative units (province, district, division and area/village) affected. An RVF outbreak was defined as above normal occurrence of abortions, perinatal mortality and hemorrhagic syndrome in livestock with or without human involvement. In most cases, primary cases were screened using real-time reverse transcription polymerase chain reaction (RT-PCR) to satisfy requirements for official declarations. Secondary cases for each outbreak were often diagnosed based on clinical signs. The total number of records in the database is 599,832 given repeated records per division. Similar data have been obtained from Tanzania identifying districts that have reported RVF outbreaks since 1930.
The georeferencing of RVF hotspot sites was also completed in Kenya, to enable a more intensive analysis using ecological niche models. The survey utilised a check list that collected additional data from each site on:
- Vegetation cover
- Types of livestock and wildlife found in each site
- Whether there were both human and livestock cases in the recent 2006/2007 outbreak
- Main livelihood activity
UDUR’s work on this task was linked to WP3 Task 3.3 and formed part of the validation of the schistosomiasis dynamic modelling exercise. Georeferenced data on S. mansoni infection prevalence in human populations were extracted from the open access Global Neglected Tropical Disease (GNTD) database. Prevalence data were available from 2965 records in total, and 594 records when surveys that did not meet stricter inclusion criteria were excluded (‘selected data’). When prevalence data was plotted on a map of the mean (across scenarios) model output ‘infection risk’ at baseline (2006-2015), the prevalence data yielded estimates of prevalence for 19% (279/1470) of grid squares when all data were used, and 7% (100/1470) when selected data were used. Prevalence estimates for each square were calculated from 1-119 (median=4) and 1-37 (median=4) individual estimates when all and selected data were used respectively.
Task 2.7 Creation of an online project database (information platform)
This task was dedicated to the creation of an online metadata entry platform. It was decided to produce this with the ESRI Geoportal Server (version 1.2.2) which is free of charge and open source. The platform (/portal) is hosted by an Apache Tomcat Server (version 6.0.37) which is a virtual server running on a physical server (Ubuntu Server) located at ILRI. This platform has a MySQL database in the background where the (meta-)data are stored. The platform (a HTML/JavaServer Pages based portal) is also fully customisable and has been redesigned by PLUS to meet the needs of the project. For example, it fits the HF colour scheme, while the homepage now has buttons for searching and adding (meta-)data that help the user to more efficiently navigate the portal (http://41.204.190.50/geoportal/catalog/main/home.page). Different login levels are available: administrator, data editors identified for each partner institution, and guest account. However, the entries, which are set to be public, can also be viewed without logging in. Finally a data upload function was developed, whereby data can be uploaded during the fill-out process of metadata. The data are stored in folders and the name of a folder is the ID of the metadata file. This helps the organisation of the data, particularly in case of deleting a metadata record. ICTP is providing the FTP storage. Once the data are uploaded they are reviewed and quality assured by an administrator at NUR. This process of uploading data will continue beyond project end as the metadata portal will be maintained post project.
WORK PACKAGE 3 - Environment-disease transmission relationships & modelling
Lead beneficiary: University of Durham (UDUR)
Task 3.1 Field Studies for the three target diseases
Schistosomiasis (UDUR)
Collecting background data on schistosomiasis in this task also contributed to both task 2.6 (compilation of disease data) and task 3.3 (Development of dynamic models). Briefly. geo-referenced data on S. mansoni infection prevalence in human populations were extracted from the open access Global Neglected Tropical Disease (GNTD) database. Prevalence data were available from 2965 records in total, and 594 records when surveys that did not meet stricter inclusion criteria were excluded from modelling.
Previous field-based work at Lake Albert was reviewed with respect to understanding where major gaps existed in terms of data that could be used to give values to model parameters in task 3.3c. The conclusion of this review was a lack of information on the natural history of snail species responsible for transmission of S. mansoni in the region. To address these gaps UDUR undertook a series of experiments to test how fecundity and longeveity of two species are affected by changes in water temperature. Laboratory experiments were conducted at Vector Control Division in Kampala, Uganda, to estimate Biomphalaria sudanica mortality, fecundity and growth rates at ten different constant water temperatures, ranging from 13-32°C. Field experiments took place on the shore of Lake Albert. Snail cages placed in open water at the Lake Albert field site were used to determine the effects of snail densities on B. sudanica and B. stanleyi mortality and fecundity rates in semi-natural conditions. B. sudanica survival and fecundity were found to be highest at 20°C and 22°C respectively. Growth in shell diameter was estimated to be highest at 23°C in small and medium sized snails, but the relationship between temperature and growth was not clear. The fecundity of both B. sudanica and B. stanleyi decreased by 72-75% with a four-fold increase in population density. Increasing densities four-fold also doubled B. stanleyi mortality rates, but had no effect on the survival of B. sudanica. The optimum temperature for fecundity was lower for B. sudanica than for previously studied species of Biomphalaria. In contrast to other Biomphalaria species, B. sudanica have a distinct peak temperature for survival, as opposed to a plateau of highly suitable temperatures. For both B. stanleyi and B. sudanica, fecundity decreased with increasing population densities. This means that snail populations may experience large fluctuations in numbers, even in the absence of any external factors such as seasonal temperature changes. Survival also decreased with increasing density for B. stanleyi, in contrast to B. sudanica and other studied Biomphalaria species where only fecundity has been shown to decrease.
RVF and Malaria (ILRI)
Field surveys on RVF were conducted in the Ijara study site in Kenya to generate empirical data that have been used in modelling RVF transmission dynamics. These activities included participatory, entomological and serological surveys; they were implemented in the second and third years of the project and their results were reported in deliverable D3.2 entitled: RVF/malaria study site analysis and major findings for RVF & malaria transmission. The main results from this task included:
- Livestock movement patterns and their population structures required for RVF modelling were described
- Mosquito vectors that are prevalent in the area were sampled and characterised
- RVF sero-prevalences in sheep, goats and cattle were estimated
KEMRI organised the collection of in-patient malaria data from the project site in Ijara District hospital for the period 2006-2011. Data on the distribution of insecticide impregnated bednets was also collected. The malaria data were tested for seasonality and seemed to be consisted with trends in rainfall. Monthly Malaria case anomalies were computed. These data were made available to the modeling team.
Task 3.2 Evaluation and development of statistical disease models
ICTP intended to develop a range of statistical models to estimate the relations between schistosomiasis, climate, environmental, and socio-demographic parameters in Uganda in collaboration with Durham University. This was not possible, however, owing to the lack of epidemiological surveillance data for schistosomiasis beyond the prevalence data used in task 3.3c. In addition, the collection of human schistosomiasis incidence data in Uganda was not possible due to issues for obtaining the ethical clearance to conduct a site analysis. Consequently, the development of statistical models was exclusively conducted on malaria data, which were more freely available in the project.
ICTP developed a range of statistical malaria models to examine how malaria incidence changes as a function of climatic, environmental, demographic and socioeconomic predictors in Uganda and Rwanda. The key component of this task was a dataset of epidemiological surveillance data (i.e. number of malaria cases) retrieved from our partners in the ministries of Health of Uganda and Rwanda. ICTP also collected environmental, socioeconomic and demographic data from various sources, including but not restricted to bureaus of statistics, remote sensing datasets, climate reanalyses, international malaria programmes, and censuses. It was possible to retrieve climate data for the same period of time, and at the same spatiotemporal resolution as the epidemiological data. Socioeconomic data were only available for selected periods of time, and for selected geographical regions. These features of the socioeconomic data prevented their incorporation into the statistical models. They also compromised the skill of the statistical models. In the end, a dataset of malaria, climate and non-climatic predictors was developed covering the whole of Rwanda and Uganda. To our knowledge, these are the longest and more spatially diverse malaria-related datasets yet assembled for both countries.
Some problems were inherent to the epidemiological data. First, some entries were missing or duplicated, and so statistical interpolation methods were implemented by ICTP to estimate a value for those data. Second, some of the entries in the malaria dataset had levels that were significantly lower or higher than the surrounding entries. For most of these entries, there was no available information in the ministries of health as to whether those values corresponded to real observations or were simply data entry errors. Since it was impossible to determine the veracity of such values, an algorithm was developed by ICTP to filter the data and reduce the impact of those suspicious entries in the performance of statistical models. Third, the length of the malaria time series was relatively short (just over a decade), posing challenges for estimating significant effects of climate drivers on malaria incidence using statistical models.
Statistical models were then developed using a time series cross-validation algorithm to detect the model specification resulting in the lowest prediction errors. These models accommodated the delayed and nonlinear effects of climate variables on malaria incidence. Given that data were unavailable for key malaria predictors, the statistical models incorporated random effects terms to account for the potential effects of unobserved variables. This modelling approach was used on both the Uganda and Rwanda datasets.
The risk of infection estimated by the model was evaluated through comparisons with malaria observations in both countries. Model outputs were found to reproduce the spatiotemporal dynamics of malaria in both countries. The estimated relationships between malaria and climate were statistically significant in both countries, and the functional form of such relations was in agreement with previous research. The relations between malaria and socioeconomic development, however, were significant only for a small number of predictors. This finding may indicate that some of these variables are not important for describing changes in malaria incidence. However, it is likely that these results are due to the very short time series of socioeconomic data preventing the model from estimating significant effects. The significant relationships estimated between malaria incidence and socioeconomic drivers have to be cautiously considered too, because they arise from data aggregated at a different temporal resolution from the epidemiological data.
Task 3.3 Evaluation and development of dynamical disease models
Task 3.3a - Malaria
This task involved the development and application of dynamical climate-dependent malaria transmission models for eastern Africa. The two malaria models produced or developed through HF, LMM and VECTRI, were driven using climate projections provided by a large ensemble of climate models derived from multiple bias correction techniques. A large inter-comparison exercise was carried out (Task 3.4) to provide a quantitative method for analysing the impact of the long-term effects of climate change on malaria transmission in eastern Africa.
The LMM, initially formulated in 2004, was further developed within HF with the addition of components representing transmission from a chronically infected immune population at the start of the season and influx of mosquitoes from permanent water bodies. These new components were tested, along with different LMM parameter settings, against records of observed malaria transmission parameters in Senegal and in South Africa. Additionally, LMM-simulated malaria transmission and variability patterns driven by various climate datasets have been compared with malaria maps from the Malaria Atlas Project 2010 (MAP¬2010), a statistical analysis of malaria prevalence based on observations. LMM is largely able to reproduce the spatial distribution and seasonal variability of malaria incidence in the eastern Africa when driven by observational climate timeseries (from the ECMWF interim reanalysis, Global Precipitation Climatology Project and the Tropical Rainfall Measurement Mission Multi-satellite Precipitation Analysis) and compared with Mapping Malaria Risk in Africa (MARA) and MAP¬2010 products. Analysis comparing LMM hindcasts driven by ECMWF’s state of the art seasonal forecast system (System 4) and the observationally driven simulations indicates some probabilistic skill for high (above upper tercile), above average (above median) and low (below lower tercile) malaria incidence with values (>0.7) of the area under the Relative Operator Characteristic (ROC) curve with a forecast lead time of approximately three months.
Different dynamical and statistical malaria models (MIASMA, MARA and UMEA) were also compared with LMM and VECTRI, both at the global scale and for Africa. This was carried out for both the recent context and under different climate change and population growth scenarios (consistent with Task 3.4). Results showed that the climate might become more suitable for malaria transmission in the highlands of eastern Africa and, to a lesser extent, less suitable over west Africa. This first multi-malaria model comparison exercise was carried out within the Inter-Sectoral Impact Model Intercomparison Project (ISI-MIP), QWeCI and HF frameworks and this involved research teams at UNILIV, ICTP, UMEA, Maastricht University and the London School of Hygiene and Tropical Medicine.
A large ensemble of long-term climate projections under various emissions scenarios were integrated with the epidemiological models LMM and VECTRI to provide a range of projections of climate-related malaria risk in eastern Africa over the next century. This undertaking incorporated the two malaria transmission models, a range of emission scenarios (as defined by the IPCC), and four separate bias correction techniques comprising a total of 27 climate model datasets, many of which were developed by partners in HF. This implementation allowed us to compare between the two malaria models, two of the common RCPs and three of the bias-corrected climate model streams. Analysis was performed on these multi-model malaria risk projections by calculating the mean, spread (standard deviation) and relative differences in time (anomalies). For the historical period (1980 to 2005) both the malaria models were compared with MAP2010. Both LMM and VECTRI overestimate the prevalence for this period as these models represent the effects of climate on malaria transmission only, and do not take into account interventions that have been carried out during this period or any immunity effects. Despite the wide array of driving input data for the disease model projections there was some consistency to be found in the results, which was tested by quantifying multi-model agreement. Multi-model agreement is said to occur for a given period if all the component mean values for that period are within two standard deviations of the overall mean. For example, if all the Global Climate Models (GCMs) used within a modelling stream (particular bias-correction technique) have mean values within two standard deviations of the modelling stream mean then the multiple climate models are deemed to share a certain level of agreement. The multi-model spread (uncertainty) is generally higher near the epidemic fringes of the distribution (for low prevalence values).
Numerical malaria models solely driven by climate parameters cannot explain the decline in malaria endemicity observed over the 20th century at the continental scale (mostly due to human intervention, land use management and other socio-economic parameters). Generally, more significant climate impacts and greater consistencies across the multi-model ensemble are shown under higher emission scenarios (RCP8.5) for the end of the 21st century. There is no clear agreement between models for near-term projections (2020s). The largest uncertainties are associated with the methodology, e.g. the numerical malaria models, as opposed to the driving climate model projections. The uncertainties related to the spread in rainfall and temperature changes as simulated by the climate models are large over the northern fringe of the Sahel. This is consistent with the diverging rainfall projections shown by the various climate models over these regions. The uncertainties related to the emission scenarios are relatively small, but they linearly grow as a function of time over highland areas. For future climate projections, all numerical models generally agree on the increase in climate suitability for malaria transmission over the eastern African highlands of the Rift Valley and Ethiopia. Furthermore there appears to be general agreement between models on a southward shift of the epidemic fringe that lies to the north of the central African endemic belt.
Task 3.3b RVF
As anticipated in the HF description of work, the sister EU FP7 project QWeCI developed a RVF version of the LMM, named the Liverpool Rift Valley fever model, or LRVF. This new epidemiological, dynamic transmission model underwent further development through detailed parameterisation in this project using literature-based data and data collected as part of HF. The LRVF model was further developed as part of Task 3.3 to include a complex dynamical host component to take into account the greater impact of transmission dynamics that livestock has for this disease. The relationship between environmental conditions and RVF outbreaks in Mauritania has also been investigated. Results show that the four reported RVF outbreaks over Mauritania (1998, 2003, 2010 and 2012) were preceded by similar rainfall conditions e.g. a rainless period lasting at least seven days followed by heavy precipitation. This result is consistent with former published studies focusing on RVF outbreaks over Senegal and was used to inform LRVF development.
Disease model development carried out as part of the FP7 QWeCI project resulted in the creation of a generalised disease-modelling library, EpiCS. EpiCS (EPIdemiological modelling toolkit for Climate Sensitive disease) is a C/C++ library of generic functions that allows any host or vector process (such as mortality, population growth, biting rate) to be associated with any transmission model structure. The toolkit has been tested by recreating the existing LMM. A prototype two-vector, single host dynamic RVF model was also implemented using the EPICS toolkit. The amplifying vectors for RVF, Culex spp mosquitoes, were modelled using the same structure as the dynamic mosquito model in LMM. The population dynamics of the reservoir vectors for RVF, Aedes spp mosquitoes, were modelled using a drying/wetting trigger for the larval stage, driven by rainfall, a physically based version of the rainfall criteria used in other RVF models. In LRVF, the eggs of Aedes require a (configurable) period of drying and then re-wetting before they can proceed to larval and pupal stages. Prototype LRVF model parameters were taken from the literature, where available, and LMM settings were used as a starting point for parameters for which there are no published values. Plugins derived from the new version of the LMM and the newly developed LRVF model were embedded and tested for the Disease Model Cradle (DMC) GUI. The plugin deployment procedure used on each operating system platform (Linux, Mac OS X and Windows) has also been streamlined.
Task 3.3b required the development of a new climate driven dynamical model of RVF transmission based on LMM. Following the formulation of an initial prototype model, redevelopment of the host component of the LRVF model and model parameterisation was finalised. A more complex host component for LRVF was developed to take into account the greater impact of transmission dynamics that livestock has for RVF compared with malaria. In the new host module, the population is divided into two sub-populations based on age division due to the much higher susceptibility of young livestock to RVF infection. The sub-populations are dynamically coupled via new births and the young mature into adult livestock. The LRVF model was qualitatively calibrated based on past RVF epizootics in Kenya and Tanzania, in collaboration with ILRI, following anomalously high rainfall events using reanalysis climate data. Various parameterisations of the model were used in an iterative process that combined qualitative, inverse modelling and knowledge of the system (both local and from literature) to produce Culex EIR spikes in the expected years based on observations. This process guided previously unmeasured parameters away from the default LMM-derived set-up to a unique LRVF based parameterisation via an improved understanding of model parameter sensitivity.
The new LRVF model was integrated with climate model data in order to provide long-term projections for RVF dynamics in eastern Africa. LRVF was driven by five different GCMs comprising the models calibrated by ISI-MIP for moderate and extreme emission scenarios (respectively Representative Concentration Pathway (RCP)4.5 and RCP8.5). These climate models provide daily temperature and precipitation data for the period 1980 to 2099. The outputs of LRVF provide an estimate for areas that are vulnerable to RVF epizootics as a result of the state of the climate and predicted current livestock immunity. By driving LRVF with climate model projections, the long-term changes in epizootic susceptibility for eastern Africa were predicted. The precise nature of these changes was dependent on the climate model used but certain dynamics exhibited general trends from which a consensus could be drawn. In the long-term, LRVF considers potential shifts in regions that are especially suited to enzootic and epizootic activity based on climate-driven projections of vector dynamics. Results of LRVF suggest that there is a threat of the highland areas of eastern Africa becoming more suitable for RVF transmission as temperature increases and these regions support greater populations of the RVF vectors. Whether regions susceptible to increased RVF transmission in the future are capable of supporting a major outbreak depends on shorter timescale rainfall dynamics, as well as on the local vector population, state of host immunity and level of vulnerability.
Task 3.3c Schistosomiasis
An agent-based model of the temperature-sensitive stages of the Schistosoma and intermediate host snail life-cycles, parameterised using data from S. mansoni and Biomphalaria pfeifferi laboratory and field-based observations, was developed through HF (Task 3.1). Infection risk is calculated as the number of cercariae in the model, adjusted for their probability of causing infection. The number of snails in the model is approximately estimated to remain constant between 15–31°C. Outside this range, snail numbers drop sharply, and the snail population cannot survive outside the range 14–32°C. Mean snail generation time decreases with increasing temperature from 176 days at 14°C to 46 days at 26°C. Human infection risk is highest between 16–18°C and 1 pm and 6–10 pm in calm water, and 20–25°C and 12–4 pm in flowing water. Infection risk increases sharply when temperatures increase above the minimum necessary for sustained transmission. The model suggests that, in areas where S. mansoni is already endemic, warming of the water at transmission sites will have differential effects on both snails and parasites depending on abiotic properties of the water-body. Snail generation times will decrease in most areas, meaning that snail populations will recover faster from natural population reductions and from snail-control efforts. Links between the ecological properties of transmission sites and infection risk that could significantly affect the outcomes of interventions designed to alter water contact behaviour, with such interventions more likely to reduce infection levels at river locations than lakes, where infection risk remains high for longer. In cooler areas where snails are currently found, increasing temperatures may significantly increase infection risk, potentially leading to new, high-intensity foci of infection.
The model was run using low, moderate and high warming climate projections over eastern Africa. For each climate projection, eight model scenarios were used to determine the sensitivity of predictions to different relationships between air and water temperature, and different snail mortality rates. Maps were produced showing predicted changes in risk as a result of increasing temperatures over the next 20 and 50 years. Comparing baseline model output with prevalence data indicates suitable temperatures are necessary but not sufficient for both S. mansoni transmission and high infection prevalences. All else being equal, infection risk may increase by up to 20% over most of eastern Africa over the next 20 and 50 years. Increases may be higher in Rwanda, Burundi, south-west Kenya and eastern Zambia, and S. mansoni may become newly endemic in some areas. Results for 20-year projections are robust to changes in simulated intermediate host snail habitat conditions. The baseline estimates of S. mansoni transmission potential were validated using data collected as part of Task 2.6.
An ensemble of regional climate simulations over Africa generated through work in task 4. 5 (provision of high resolution projections) was used to provide projected daily maximum and minimum temperature data for the eastern Africa study region. The ensemble consists of three members of the Rossby Centre Regional Climate Model (RCM) – RCA4 [22], driven by a coupled atmosphere ocean general circulation model (AOGCM) – EC-EARTH [23]. Three RCP scenarios were used – RCP2.6 RCP4.5 and RCP8.5 – which represent low, moderate and high levels of warming respectively. All three regional simulations were made within the African branch of the Coordinated Regional Downscaling Experiment (CORDEX), and cover the whole African continent at about a 50 km (0.44°) resolution. A smaller sub-domain in eastern Africa, with an area of 1470 (35 x 42) grid boxes, was selected for the study.
Task 3.4 Scenario integration
This task was completed towards the end of the project in the form of the HF Atlas, an online system for analysing and visualising integrations of data and models in order to assess the projected risks of the three VBDs targeted by HF. A summary of activities related to this task, undertaken by UDUR, UNILV, ICTP and ILRI, is provided below for each of the three VBDs.
RVF
How peak-year frequency in RVF incidence changes in comparison to the historical period for both immature and mature livestock was tested. The two regions previously highlighted as hotspots for vector abundance (central/western Kenya and Rwanda) showed significant increases in incidence while most other areas show a significant decrease. These results appear to correlate with suitable mosquito habitat based largely on temperature. As the entire region heats up throughout the century, mosquito abundance is effectively forced into cooler, more highland regions in order to survive. It should be noted that this is a feature of the chosen mosquito-survival scheme implemented in the model, which would benefit from further rigorous analysis due to its evident sensitivity. The decrease in relatively warmer, lower areas was seen in both vectors in the late century. Aedes start to decrease significantly even earlier as large parts of the region also become wetter, but this does not affect Culex proliferation. These trends in vector population dynamics propagate through to the transmission characteristics of the model.
The change in incidence peaks for immature and mature livestock were similar (with significant increases in highland regions) but the increase around Rwanda was less pronounced for mature livestock. Since livestock characteristics were the same in all but case fatality ratio this suggests that livestock immunity has a role to play. Since infected immature livestock either die from the disease (with a probability of approximately 70%) or mature into adults at approximately 2 months, the effects of immunity are negligible. Hence we can think of these immature livestock dynamics as representing an entirely susceptible population.
Information from this mapping exercise formed the basis of the hazard layer for the HF Atlas.
Schistosomiasis
UDUR undertook analysis of different scenarios using the dynamic agent-based model developed in Task 3.3b. Model output was used as the basis for disease risk mapping in the HF Atlas.
Malaria
Different malaria simulations were carried out within HF for the historical period (1980 to 2005).
The impact of future climate change on simulated malaria distribution was tested for LMM-VECTRI based on the super climate ensemble for two emission scenarios (RCP4.5 or RCP4p5 and RCP8.5 or RCP8p5) and for different time slices (2020s, 2050s, 2080s). The results of the super ensemble generally agree with former results obtained during the ISI-MIP and QWeCI EU projects. Those former results were based on a greater malaria model ensemble (including MARA, MIASMA and UMEA) using fewer climate model inputs as drivers (five GCMs were used whereas here we combined 23 different GCMs and RCMs). The climate seems to become increasingly suitable for malaria transmission over the highlands of eastern Africa. The LMM alone tends to simulate similar changes which are, however, smaller in magnitude with respect to the super ensemble and VECTRI.
As with RVF and Schistosomiasis, the HF Atlas provided a means of integrating projected disease hazard with indicators of vulnerability.
WORK PACKAGE 4 - Disease risk and vulnerability mapping
Lead beneficiary: University of Rwanda (UR)
Task 4.1 Identification of current vulnerability hotspots for the 3 target diseases
PLUS, ILRI and UR together adopted a risk and vulnerability framework that viewed humans and their environment as part of coupled socio-ecological systems (SES). Based on a comprehensive literature review and expert consultations, a list of indicators and related data for social vulnerability was agreed for malaria, RVF and schistosomiasis. Building on a methodology developed by PLUS, spatial vulnerability units/regions (geons) were modelled, representing the social vulnerability for the three target diseases (deliverable D4.1).
Task 4.2 Assessment and provision of downscaled climate change projections for the study region from past and on-going projects
An ensemble of statistically downscaled future projections of daily rainfall and temperature for 11 CMIP5 General Circulation Models (GCMs) using the historical and the RCP4.5and RCP8.5 scenarios was provided through this task. The output covered the 140-year period 1960-2099 at 50 km resolution for the eastern Africa region. This ensemble of runs allowed for an exploration of the range of uncertainty introduced by the different GCMs and due to the different emission scenarios. The results were used as input into the different disease models used within HF.
Task 4.3 Provision of land-use change scenarios
To assess the effects of land use change (LUC) on the malaria risk, data from four of the five available Earth System Models (ESMs) that contributed to the LUC experiment of the CMPI5 were used to drive a spatially explicit, dynamical malaria model. Previous efforts to incorporate LUC in dynamical disease models have been extremely limited, and have mostly examined the indirect effect by which LUC impacts temperature and precipitation, which in turn impacts malaria. This study used the suite of ESMs that participated in the CMIP5 process and conducted twin investigations with and without LUC. Results from the task provide the first multi-model assessment of the potential impact of LUC on malaria transmission in Africa via its impact on climate.
To isolate the impact of LUC on climate, the five CMIP5 groups performed two sets of simulations for the period 2006-2100 using the same forcing as for RCP2.6 and RCP8.5 but with the land-use invariant over time. Daily precipitation and temperature output from four of the five ESMs was used to drive the VECTRI model for integrations to 2100 for the model’s historical run (1960-2005) and the two RCPs (2005-2100). A malaria model was run for each available ESM member (three CanESM, two MPI and one member for IPSL and MIROC).
The impact of LUC on climate was assessed in terms of two variables only, the 2 metre temperature and precipitation, as these are key inputs to the malaria model. The first notable aspect of the impact of LUC on precipitation is that, for each ESM, the response is considerably different between the two RCPs. This is in distinct contrast to the changes that occur between the present and future, where the higher greenhouse gas concentrations essentially amplify the magnitude of the spatial changes observed in the lower emissions scenarios (see for example section 2 of HF D3.4). This highlights the fact that the LUC maps are considerably different spatially between the two scenarios. The magnitudes of the changes are relatively small and considerably smaller than the changes observed due to greenhouse gas emissions.
As the impact of LUC on climate is minor in three of the four ESMs, LUC was expected to have a limited indirect impact on malaria in those models. Using the metric of the malaria prevalence (parasite ratio, PR) and length of the transmission season (LTS) this was indeed seen to be the case. In the IPSL, MPI and CanESM models, the change in prevalence was less than 5% everywhere within Africa, and the LTS were mostly less than 10 days. Only the MIROC model produced substantial changes, with prevalence increasing strongly in the southeast of Africa over the northern part of Mozambique and southern Tanzania. Over the Sahel region, the increases in temperature led to decreases in prevalence. This is because above about 30°C mean temperature, malaria transmission starts to decrease with temperature.
In conclusion, this task was the first attempt to isolate and assess the potential indirect impact of LUC on malaria transmission using a dynamical modelling framework, and multiple ESMs based on the latest projection of potential LUC that contributed to the latest round of the IPCC climate assessment. The study showed that in terms of indirect impact, whereby LUC impacts local climate, which can then alter transmission, the effect was very limited in three of the four available models. In one model, the impact was more significant.
The study also highlighted the uncertainty in projecting such extremely model dependent impacts.
Task 4.4 Provision of socio-economic change scenarios
The original intention was to integrate socio-economic indicators and their respective quantities, and future projected values for these, with spatially explicit output from dynamic disease models in order to generate risk maps for the three VBDs. Projected future changes in socio-economic indicators were to be based on IPCC scenarios. Two main activities were carried out in relation to this original intention. A joint scenario workshop was organised by ILRI from 5-7 November 2012. This was organised together with colleagues from CCAFS, where regionally based scenarios were adapted to the specific disease context. These scenarios are qualitative and help to better understand future development pathways in the VBD context. Additionally, how levels for key demographic and economic social vulnerability indicators might be projected for the future and integrated within the HF Atlas was discussed (deliverable D4.4).
Task 4.5 High resolution regional climate projections for eastern Africa study area
A set of high resolution Regional Climate Model (RCM) simulations for eastern Africa was created for use by HF consortium members, in particular those directly involved in projecting future environmental change-driven disease outcomes. Simulations covered the period 1970 to 2100 and generally followed the WCRP CORDEX experiment protocol (http://wcrp-cordex.ipsl.jussieu.fr/). While the majority of Africa CORDEX simulations employed an RCM grid covering the entire African continent at a spatial resolution of 50km, special resolution of the RCM was increased and targeted eastern Africa in order to provide finely resolved spatial detail in simulated present and future climate variables. These simulated data were then input to subsequent assessments of disease risk.
Task 4.6 Risk and vulnerability mapping of disease morbidity and related impacts
The primary aim of Task 4.6 was to produce spatial assessments of current and projected future risks of, and vulnerability to, the three target VBDs in the EAC region. Future projections of risk were based on environmental change data-driven output from dynamic disease models and their subsequent integration with vulnerability indices. Achieving this aim also involved development of the HF Atlas though a highly consultative and reflective process.
Mr Jean Pierre Bizimana was recruited as PhD student for HF in October 2011. The title of his PhD, submitted before project end, was ‘Climate Variability and Malaria in Rwanda: Spatial Assessment of Social Vulnerability at Different Scale Levels’. The specific objectives of the research were: to conceptualise social vulnerability to malaria in the context of environmental and socioeconomic changes; to identify suitable indicators of social vulnerability to malaria; to apply an indicator framework for spatial assessment and explicitly model the social vulnerability to malaria; and to examine the extent to which malaria incidence is the interplay of climate variability and socioeconomic conditions of highland communities in Rwanda.
WORK PACKAGE 5 - Adaptation and support tools: development of decision support tools
Lead beneficiary: Stockholm Environment Institute (SEI)
Task 5.1 Identification and assessment of environmental management and climate change adaptation strategies
An inventory of all strategies relevant to the project – health, environmental, water and sanitation – was created and a report summarising the findings was produced. A total of 50 country documents were posted to the internal project website.
Task 5.2 Identifying key stakeholders for strengthening human and animal health systems
A database of relevant stakeholders and domain experts in the EAC included representatives of numerous sectors (ministries of environment, health, water and irrigation; environmental management agencies; climate change directorates; civil society; and donor agencies). The development of the database involved multiple visits to all five countries comprising the EAC, thus raising awareness of HF. The database was also instrumental in disseminating project outputs and identifying the appropriate mix of participants for the RVF and malaria/schistosomiasis workshops that provided a forum for discussing the architecture and content of the decision support facilities (frameworks and tools) that were important products of WP5 and the project in general.
Task 5.3 Use and assessment of current environmental management and climate change adaptation tools
This task assessed environmental management and climate change adaptation tools that address the environment-related vulnerability of human health. Relevant literature on environmental health impact assessment tools and their application, including those developed specifically for addressing environmental health concerns, was evaluated. Feedback on recent and current use of assessment tools was also sought from domain experts in the EAC. The task informed later stages in the research project when decision support tools (DSTs) and frameworks (DSFs) were in the process of being discussed and developed.
Task 5.4 Refining the RVF Decision Support Tool
Activities carried out as part of this task refined the RVF DSF that had been developed earlier by domain experts involved in attempts to mitigate the impact of RVF in eastern Africa. Three RVF workshops were organised by project members during 2014: 24-25 February (Nairobi, Kenya); 30 September – 1 October (Naivasha Kenya); and 12-13 November (Dar e Salaam, Tanzania). Discussions at the workshop enabled the DSF to be modified in line with up-to-date information, including data generated by HF, and recent past experience, with the amended DSF forming part of deliverable D5.4.
Task 5.5 Developing decision support framework
This task resulted in the production of risk-based DSFs for malaria and schistosomiasis. These two DSFs address changing risk for the two diseases in light of environmental, including climate, change. The DSFs were developed in a workshop held in Nairobi in November 2014 that brought together HF researchers and decision-makers involved in managing the response to these two diseases.
WORK PACKAGE 6 - Stakeholder engagement and empowerment through sharing of knowledge and training
Lead beneficiary: Trinity College Dublin (TCD)
Task 6.1 Identification – and if necessary grouping - of stakeholders and target end users of the HF project
A number of partners contributed to identifying and grouping potential stakeholders and target end users of HF, including TCD, PLUS, NUS and CH. For example, PLUS provided contacts for relevant stakeholders in eastern Africa, within the EC and internationally. The specific focus provided by PLUS was on the identification of stakeholders in the domain of Earth Observation and GIS, including initiatives such as GEO, Copernicus and the WHO. CH helped identify key stakeholders in government and private institutions in Uganda. Information on stakeholders was also harvested from WP5, Task 5.2 ‘Identifying stakeholders for strengthening human and animal health systems’ and other relevant WPs. Project deliverable D5.2 ‘Identification and engagement of key stakeholders’ illustrates the grouping of stakeholders and maps the linkages and flows of information between various groups.
All identified stakeholders were input to a database with fields comprising: type of stakeholder; contact details; and preferred communication language (English or Kiswahili). A protocol was devised by AquaTT for updating and managing the database, and throughout the project the database continued to grow with the addition of the contact details of relevant stakeholders.
The stakeholder database was actively employed when carrying out dissemination activities to ensure those with an interest in HF outputs were kept up-to-date, to increase impact and to facilitate translation of research findings into meaningful actions on the ground. The database was also utilised to invite the participation of relevant stakeholders for the project’s various stakeholder meetings and engagement workshops.
Task 6.2 General dissemination
Over the course of the project, HF has employed a range of methods to disseminate project results and outputs to stakeholders and end users. Dissemination of information to project members was facilitated through the use of Basecamp, the project intranet site. A publically accessible website, available in both English and Kiswahili, was developed and maintained throughout the project providing regular updates on project news and outputs. The website has proven to be highly popular having received ~40,000 hits since its launch midway through the first year of the project. AquaTT will continue to maintain this website for up to five years after the completion of the project to ensure the continued and uninterrupted dissemination of HF outputs.
In addition to the website, a number of other dissemination activities has taken place including the production and distribution of:
- Seven contractual HF news releases (available in both English and Kiswahili) and additional news releases disseminated through such channels as AlphaGalileo (http://www.alphagalileo.org) the EC’s CORDIS News services (http://cordis.europa.eu/news/) CORDIS WIRE, Twitter, LinkedIn and AquaTT’s Training News.
- Seven HF newsletters (non-contractual)
A non-technical, tri-fold project brochure was developed for general release outlining the project and English and Kiswahili versions are available to download from the project website.
A short promotional video (c.90 seconds) that summarises HF was also developed. The video, hosted by VIMEO and aimed at raising awareness of the project, can be viewed at www.vimeo.com/70318624 or on the project website.
Two promotional items that include the HF logo and website address were designed to increase the brand awareness of the project: a key-ring with bottle opener and torch, and post-it notes. These have been distributed at a number of stakeholder engagement workshops and events over the course of the project. These items will continue to be disseminated after the completion of the project, for example, at the HF session at the ‘Impact of Environmental Changes on Infectious Disease’ Conference which will be held from the 23 – 25 March 2014 in Sitges, Spain.
HF also contributed to the completion of a short documentary produced by the charity Africa Turns Green. The documentary “Health and climate change in Africa” was released in August 2014 and is available to view on the project website:
http://www.healthyfutures.eu/index.php?option=com_k2&view=item&layout=item&id=176&Itemid=276&lang=en
and on YouTube:
https://www.youtube.com/watch?v=oYL4Nc-qnKE
The video focuses on the collaborations between HF and its sister project, QWeCI, and outlines how both projects have worked to increase capacity in Africa and reduce the impacts of the target VBDs in the study region. A number of representatives from HF partners (UDUR, UNILIV and ILRI) were featured in the documentary, while Dr Laragh Larsen and Professor David Taylor were consulted over the content of the script. This documentary has been used during stakeholder engagement events and will be employed in future to inform the general public about HF and about project outcomes.
HF has been included as a case study in the World Health Organization/the World Meteorological Organization (WHO/WMO) special publication on Climate Services for Health. The submission focuses on the HF Atlas, and is part of a small group of accepted pieces from more than 75 submissions globally. The final expected outputs of the Climate Services for Health project comprise an online e-book, interactive web repository and a hardcopy publication in the six UN languages.
Task 6.3 Targeted dissemination
Four project updates have been distributed to all stakeholders listed in the stakeholder database. These project updates summarise the work that was carried out each year, highlighting noteworthy achievements. These short briefs provided readers with the means to obtain further information if desired.
Task 6.4 Communication and feedback
A ‘FAQ & Feedback’ section on the project website:
http://www.healthyfutures.eu/index.php?option=com_k2&view=item&layout=item&id=133&Itemid=265&lang=en
provided a channel for communication and feedback. Furthermore, all stakeholder engagement workshops carried out during the project facilitated an interactive forum for two-way communication and feedback.
Task 6.5 Compilation of all knowledge generated from the project and translation (appropriate level, terminology, language) for end users
All content on the HF website is available in both English and Kiswahili in order to be accessible to a wide range of target stakeholders. The same applies to all contractual news releases, and the project factsheet and brochure.
Within several work packages and tasks, user manuals and guidance documents on technical issues were produced, particularly those provided by PLUS on the metadata portal and the HF Atlas.
Task 6.6 Scientific publications of research and presentations at scientific conferences
A full list of peer reviewed publications can be found in deliverable D6.15. HF research, results and outputs have been presented in a variety of formats and capacities since the project’s start.
Task 6.7 Engagement and training workshops (for refinement of Decision Support tools developed in WP5)
A number of engagement and training workshops for the refinement of DSTs developed in WP5 have been held.
1. One-day stakeholders meeting on malaria and climate change, Kampala, Uganda, 25 October 2013
CH (part of the Ministry of Health (MoH) Uganda) in collaboration with ICTP, NUS and SEI organised a one-day meeting themed ‘Malaria seasonal forecasting and Climate Change adaptation’. The meeting, held in Kampala, Uganda, was attended by a variety of stakeholders, including those from the health sector, academia, researchers and malaria-related institutions. Key attendees included the Director General of Uganda National Health Research Organisation (UNHRO) Dr Sam Okware and Dr Myers Lugemwa of the Uganda National Malaria Control Programme. The meeting reviewed available malaria and climate data, current malaria forecasting by ICTP and the European Centre for Medium Range Weather Forecasts (ECMWF) and possible DSTs and approaches. One of the main aims of the meeting was the sharing of current knowledge on malaria and climate modelling using available data from Uganda. In his opening remarks to the meeting Dr Sam Okware observed that climate change affects health and the whole health system, noting that most diseases in the country are climate-sensitive, especially vector-borne and water-related diseases such as malaria and diarrhoea. Subsequent presentations at the meeting covered broad areas of data collection and quality, especially related to malaria data, malaria early warning systems and forecasting, the climate change national adaptation programme of action in the health sector in Uganda and DST. The workshop highlighted the need for the different stakeholders to use the outputs of the research to improve preparedness, quantify disease impacts, and above all to enable the decision-making/disease outbreak forecasting and response processes. Recommendations results from the meeting included: the development of a hands-on module by ICTP and supported by MoH, Uganda, for training in the use of the malaria forecasting system output in the operational planning environment; the evaluation of the potential for directly using the VECTRI modelling system for intervention planning and also investigating potential adaptation strategies with regard to sensitivity to climate change; the identification and exploration of key areas of collaboration in vulnerability of health to climate change with a view to improving upon current adaptation strategies; undertaking further research on how to extend climate and disease modelling and forecasting to other diseases. A full report on the meeting including the presentations given is available to download from the project website.
2. Stakeholder Engagement Workshop, Nairobi, Kenya, 24-25 February 2014
A second HF stakeholder workshop was held from 24-25 February 2014 at the International Livestock Research Institute (ILRI), Nairobi, Kenya. The aims of the workshop were to raise the profile of HF research among selected stakeholders in eastern Africa and to further the process of enabling the development and effective uptake of outputs from the project. Specifically, the workshop had the following objectives: to communicate the project research outputs (modelling, risk mapping etc.) that were currently in progress for each of the three target VBDs; to collect feedback on the value of the research outputs to date and on how to support their uptake by decision makers, for example, integration of modelling results into DSTs; to develop case studies for the application of DSTs and DSFs; to test different decision-making methods through the use of case study examples to compare the value of various intervention strategies; and, to define the outputs and needs for further stakeholder engagement in HF and beyond.
Participants in the workshop included a number of HF partners, external members of the HF Expert Review Panel (ERP) and representatives from government ministries and research institutes in eastern Africa. During the first day of the workshop, plenary presentations were made by project partners to introduce the project and provide updates on case studies of the three diseases. The presentation included: an overview of HF; malaria projections for 2050 and how climate information can be integrated into health planning; sources of climate data for national decision-making; guidance that models can provide for schistosomiasis control; and a DST for RVF. On the second day of the workshop, participants carried on working in three groups, each focused on one of the diseases using the case studies presented during day one. The groups tested two of the methods or ‘engines’ of the adaptation decision explorer (ADx) tool: the voting method and Analytic Hierarchy Process. ADx was being trialled at the workshop as a possible component of a DST.
This stakeholder engagement workshop was a positive step forward in communicating the latest research from the project to stakeholders, in receiving feedback about what specific concerns decision makers have, and in sharing insights into how those concerns could be best addressed. However, weaknesses in the ADx DST became apparent, as did a divide between time frame of main interest to HF (up to century scale) and the much shorter policy-oriented time scales that most concerned many decision-makers present at the workshop. One specific request to emerge from the workshop was for concise, user-friendly summaries of the likely risks of the three target diseases under changing conditions predicted for coming decades. These summaries were subsequently produced by members of the HF consortium and are available on the HF website. Overall, this stakeholder engagement workshop proved to be a valuable exercise in identifying how the outputs of HF could be designed to meet the needs of decision makers in the EAC. More information on this workshop can be found on the project website.
3. Decision-Makers Workshop for Malaria and Schistosomiasis, Nairobi, Kenya, 18-20 November 2014
The third stakeholder engagement workshop for the refinement of DSTs was held in Nairobi from 18-20 November 2014. The objective of the meeting was to develop two separate risk-based DSFs for schistosomiasis and malaria that could assist decision-makers in the EAC by serving as guides to responses to these two diseases in the face of climate change. HF outputs include spatially explicit assessments of risks of the three target VBDs in eastern Africa as a result of projected changes in environmental conditions. These assessments are based on combined hazard and vulnerability data. This HF workshop reviewed the schistosomiasis and malaria research outputs of the project in the context of the general state of knowledge for these diseases and assessed the implications of the findings for future decision making and action. Risk-based DSFs are decision-making guides developed by decision-makers. This approach was pioneered in applications to RVF epidemic response; DSFs are developed through participatory consultations that bring together health decision-makers, control programme implementers and researchers. They systematically break down the decision-making process into sets of steps by defining decision points and action categories where decisions need to be taken. They are not prescriptive. Instead DSFs act as aide memoires to remind decision-makers of areas where action should be considered at appropriate points in time. DSFs are specifically intended to reduce the impact of uncertainty and risk in decision-making. DSFs are living documents that are part of an adaptive management process.
This workshop began in plenary session with a brief technical overview leading to a discussion to distil key action points coming from the research that should be captured in the DSFs. Next the meeting discussed the objectives of the two frameworks and the time frame that they should cover. Thereafter the meeting broke into two parallel sessions: one on schistosomiasis and the other on malaria. These parallel sessions built the timelines describing the process of epidemiological change, identified the decision points and action categories and articulated the DSFs in detail. Many of the key action points arising from the technical discussion were captured in the statements of objectives and assumptions that guided the detailed construction of the DSFs.
In the case of malaria, major international control programmes are in place and the DSF is intended to be complementary to existing policy and strategy documents. The DSF accommodates the future effects on malaria epidemiology of three inter-related phenomena: climate, environment and land use change, and population migration and displacement.
In the case of schistosomiasis, international control efforts are lacking and the institutional context of schistosomiasis control is not well developed. The schistosomiasis DSF is intended to complement to regional and national policies and strategies, and to catalyse appropriate institutional development.
Task 6.8 Africa-EU Workshop
The original intention was for the University of Rwanda to host a workshop, focusing on environmental change and VBDs, in the final year of the project to coincide with the release of the IPCC’s 5th Assessment reports, particularly the report from Working Group II dealing with climate change impacts, including projected health effects. Unfortunately organisational delays, including delays releasing information about the meeting, key note speakers etc, led to a low level of international interest in the meeting and forced a decision, taken at the last (the 6th) partners’ meeting in September 2014 to cancel the workshop.
Permission was sought and fortunately obtained from the project officer in the EC to organise and host a HF-focused session at the ‘Impact of Environmental Changes on Infectious Disease’ (IECID) Conference, Sitges, Spain, 23-25 March 2015. The session attracted a large, engaged audience and was deemed a major success. Papers delivered in the session are forming the core of a special issue of the journal Geospatial Health, which is now in the process of being produced. In addition to the HF-focused session at the IECID Conference in Sitges, Spain, HF project members organised a session (CL2.5 Climate and infectious disease interactions) at the European Geophysical Union (EGU) annual meeting in Vienna, Austria, in April 2012 (i.e. during the second year of the project). During the following year, HF project members also jointly coorganised and hosted a symposium, along with its sister project QWeCI, as part of the 4th Annual East African Health and Scientific Conference that took place in Kigali, Rwanda from 27 to 29 March 2013.
HF Symposium, 4th EAC Health And Scientific Annual Conference, Kigali, Rwanda, 27 to 29 March 2014
The conference programmed comprised four sub-themes (Maternal and Child Health, Non Communicable Diseases and Trauma, Health Systems Strengthening, and Quality of Health Care) and four symposia: HIV and AIDS; Integrated disease surveillance and disaster preparedness; Tobacco Control; and the HF/QWeCI symposium on ‘Environment and Health in Africa: climate and vector-borne diseases’. Facilitated by NUR, the symposium was well-attended and brought together researchers at the cutting edge of efforts to understand the relationships between health and environment, and in particular the links between climate and VBDs in Africa. There were a total of 17 presentations in this session, including the plenary by Dr. Margaran Bagayoko (Protection of Human Environment Programme, World Health Organization Regional Office for Africa) whose talk centred on the potential of climate-based early warning systems for improved management of VBDs in Africa. The presentations that followed were wide ranging and delivered by speakers who had travelled from other parts of Africa and from Europe and Asia (the full programme is published in the third issue of the HF newsletter available on the HF website).
WORK PACKAGE 7 - Enhancing research synergy & application
Lead beneficiary: AquaTT
Task 7.1 Identify relevant projects and conferences
Over the course of the project, a list of relevant projects, events and conferences has been uploaded to the project intranet site and maintained. This list has been managed by AquaTT, which has worked to identify and add relevant information on an ongoing basis. Other consortium members have also contributed to the list by informing AquaTT of any relevant events and projects they have identified.
In addition to the Basecamp file, AquaTT has added information on all events and conferences to the HF website on both the homepage, under ‘Upcoming Events’, and to the dedicated ‘Events’ page, which displays all events on a Google calendar and provides the ability to search for events and conferences by date. There is also a dedicated ‘Search’ page that allows users to search using particular keywords.
The project website also contains a ‘Useful Links’ page. This page contains a list of relevant projects identified and links to those project websites as well as links to further Information. Currently there are eighteen links to relevant research projects, four links to relevant tools and four links to further Information.
Task 7.2 Establish an Expert Review Panel
The ERP was established at the beginning of the project in order to provide guidance and direction to HF project members. The ERP advised and evaluated HF activities using a robust system of internal and external controls. Focusing on providing advice on project direction, the ERP did not have any authority to vote on project matters, nor had it a legal responsibility in the operation of the project or the partner organisations.
The ERP initially consisted of eight panel members comprising of five experts in the research field who were not members of HF and three members of HF, one of whom was based in Africa and two of whom were originally based in Europe (one of the two subsequently moved to Asia, but continued to attend ERP meetings). The composition of the ERP was designed to ensure some independent scrutiny of the operation and aims of the project, while also facilitating communication between external members of the ERP and the wider HF consortium.
The members of the ERP were:
- John B. Malone, Professor, School of Veterinary Medicine, Baton Rouge, LA, USA (Chair of Expert Review Panel)
- Madeleine Thomson, Senior Research Scientist (Climate Information for Public Health), Africa Program, International Research Institute for Climate and Society, Columbia Earth Institute, USA
- Simon Brooker, Reader in Tropical Epidemiology in the Department of Infectious and Tropical Diseases at the London School of Hygiene and Tropical Medicine & Welcome Trust Research Fellow, KEMRI, Kenya
- Timothy Wesonga, East African Community Secretariat (EAC) Senior Livestock and Fisheries Officer, Tanzania
- Dr. Maurice Owuor Ope, East African Community Secretariat (EAC) Disease Surveillance and Epidemiology Officer, Tanzania
- Jan Semenza, European Centre for Disease Prevention and Control (ECDC), Sweden
- Paul Lowen/Ciara Egan (AquaTT), Ireland
- David Taylor (National University of Singapore), Singapore (originally Trinity College Dublin, Ireland)
- Theophile Niyonzima (NUR), Rwanda
Task 7.3 Formal peer-review and synergy events
Over the course of the project four ERP meetings were held. Each meeting was organised to coincide with Partner meetings and other project workshops in order to provide ERP members the opportunity to meet with members of the HF consortium to discuss the project.
The first meeting took place on 13 October 2011 in Kampala, Uganda. This meeting established a relationship between the ERP members and HEALTHY FUTURERS partners. The ERP members reviewed and discussed the work done to date within each work package and provided a number of recommendations to be incorporated by partners.
The second meeting took place on 10 May 2012 in Arusha, Tanzania. The ERP was bolstered by the presence of Dr. Maurice Ope, an expert in disease surveillance and epidemiology in eastern Africa, representing the Secretariat of the EAC. Maurice made a highly valued contribution to the ERP and was able to brief other members of the ERP on the fourth Annual East African Health and Scientific Conference that was scheduled to take place in Kigali in March 2012. As a result of this synergy, HF organised and ran a session at the conference on the theme of ‘Environment and health’.
The third ERP meeting was held on the 1 May 2013 in Trieste, Italy. Again the ERP met and reviewed the progress made in each work package including how the previous recommendations were incorporated by partners.
The fourth and final ERP meeting took place on the 28 February 2014 in Nairobi, Kenya. The ERP provided invaluable advice during this meeting on the strengths and weaknesses of the research approaches taken within the project, particularly in relation to the development of a DST.
Task 7.4 Report on meeting outcomes
Reports from the each of the ERP meetings are available to download from the project website: http://www.healthyfutures.eu/index.php?option=com_k2&view=item&layout=item&id=140&Itemid=267&lang=en
and have been disseminated to key stakeholders.
A report on the collaborations and synergies developed over the course of the project was also produced, and provides a comprehensive overview of the main collaborations established over the course of the project. The report has been disseminated to key stakeholders and is available to download from the HF website: http://www.healthyfutures.eu/images/HEALTHY_FUTURES_Synergies_and_Collaborations.pdf
Potential Impact:
HF with its integrated holistic approach has considerably enhanced current understanding and dynamic modelling of links between the three target, water related VBDs (malaria, schistosomiasis and RVF) and environmental change drivers and effects. The project has also raised awareness of potential future environmental change impacts on health among decision makers in the countries comprising the East African Community (EAC), and among the secretariat for the EAC. The potential impact of HF is described in relation to the expected impacts identified in the project proposal:
(1) More accurate and reliable predictions for the distribution of three water related vector-borne diseases (VBDs) in Africa
HF collected a wide range of existing and new data and collated these data within a project database (‘Geoportal’) (WP 2). These data comprise historical, socio-economic, migration, settlement and conflict data; earth sciences data; direct climate observations and the results of regional downscaling of global reanalyses; and information on disease-environment relationships. The Geoportal acts as a main repository for data and information on the current and future impact of environmental and socio-economic change on VBDs in eastern Africa, specifically the countries comprising the East Africa Community (EAC), and is an important resource for disease planning, further research and teaching and training in the study area and farther afield. The Geoportal provides a standardised format that enables communication between databases. The link, accessible to all interested parties, is: http://41.204.190.50/geoportal/catalog/main/home.page
The HF Atlas, intended as a resource to support teaching, research and health planning, utilises the data collected and collated through HF to highlight hotspots of projected disease risk in the region. This interactive, web-based mapping and DST, aims to provide meaningful and guided access to information on environmental change, and disease hazard, vulnerability and risk relating to the three target VBDs in eastern Africa. Changes in projected disease hazard and risk can be visualised for different time periods up to the end of the current century and for different geographical scales. The Atlas also compiles searchable information on historical disease outbreaks in the study area. An important aspect of the Atlas is use of climate models and future scenarios referred to in the latest (the 5th) IPCC Assessment Report.
HF did not carry out any disease modelling for the European region owing to budget constraints. However, environmental change impacts have a transboundary nature, and the distributions of vectors and related diseases are variable. In a highly globalised and dynamic world, VBD surveillance and early warning systems targeted at Africa have therefore relevance for Europe. Moreover, research carried out through HF is, with further financing, extendable to neighbouring parts of continental Africa, and to continental Europe to the north.
(2) Strengthening of the early warning, surveillance and monitoring systems for vector-borne diseases
A key outcome of HF has been release of the DSF designed specifically to strengthen existing early warning, surveillance and monitoring systems in eastern Africa for the three target VBDs and developed in consultation with decision makers in the East African Community (EAC).
(3) Support to policies on climate change and health
HF identified and deployed the most appropriate environmental change adaptation strategies and decision support approaches (including relevant frameworks and information platforms) for adoption in the project. One of these tools is the RVF DSF, which has been established and field-tested in the study area. The RVF DSF has been designed to guide timely, evidenced based decision-making in the control of RVF, and decomposes the RVF epidemic cycle into explicit steps that are then matched against specified actions (including interventions). Many of the steps are explicitly time dependent. The integration of inputs from multiple partners, decision-makers and experts (through a series of stakeholder engagement workshops hosted in the study region) was designed to ensure ownership of the product, and relevance to the decision-making challenges that have been experienced during previous RVF outbreaks.
Refinement of the RVF DSF was based on new outputs from research, particularly those evaluating the impacts of climate change on RVF epidemiology and the incorporation of human health interventions into the framework in line with One Health principles. The RVF DSF makes clear links to policy. For example, during an inter-epidemic period the RVF intervention for animal health disease prevention is to develop a clear policy on vaccination against RVF, including during inter-epidemic periods, when risk of RVF outbreak is high and in the face of an outbreak; during a pre-outbreak phase the RVF intervention for human health and animal health is to conduct a rapid risk assessment incorporating the level of vaccination in the area and informing decision and policy-makers of results.
DSFs have also been developed for malaria and schistosomiasis, also in consultation with domain experts in the EAC. Guidance notes covering projected changes in disease rises for the three target VBDs over coming decades, driven by environment, including climate, change, have also been made publically available.
In recognition of the project’s policy relevance in the field of environment and health, HF has been included as a case study in the World Health Organization/the World Meteorological Organization (WHO/WMO) special publication on Climate Services for Health. The case study focuses on the HF Atlas, and is part of a small group of accepted pieces from more than 75 submissions globally. The final expected outputs of the Climate Services for Health project comprise an online e-book, interactive web repository and a hardcopy publication in the six UN languages, produced by the WHO/WMO.
(4) Promotion of sustainable management of the natural and human environment and its resources by advancing our knowledge on the interactions between the biosphere, ecosystems and human activities
Water-related VBDs are transboundary in their distribution and effects. An integrated (i.e. regional and global), multi-national response is therefore required. The transmission and outbreaks of VBDs are also not only related to health and to environment. Their occurrence and severity are also products of socio- economic conditions, and the abilities of health and veterinary services to respond to early warnings of possible outbreaks. Joint action, across international boundaries, therefore represents the only effective way to ensure secure livelihoods and the health security of citizens. HF was guided by the concept that the stability of livelihoods, i.e. their strong resilience, crucially depends on maintaining environmental integrity. In this context, the consortium partners saw a strong link between animal and human health and environmental sustainability.
(5) Developing new technologies, tools and services, in order to address in an integrated way global environmental issues [...tools and on technologies for monitoring, prevention and mitigation of environmental pressures and risks...]
The HF project has led to the development and application of the following new technologies, tools and services:
- A publically accessible Geoportal comprising information drawn from a range of primary and secondary sources, including: historical, socio-economic, migration, settlement and conflict data; earth sciences data; direct climate observations and the results of regional downscaling of global reanalyses; and information on disease-environment relationships. The Geoportal is at: http://41.204.190.50/geoportal/catalog/main/home.page
- Improved understanding of environment-disease dependencies in eastern Africa, and of the factors that act to confound any relationships
- A new generation of dynamic disease models for the three target, water related VBDs
- The HF Atlas (http://zgis186.geo.sbg.ac.at/hf_atlas/) an interactive, web-based mapping and visualisation tool
- DSFs for the three target VBDs developed in consultation with relevant decision makers in the EAC region
- Environmental change and health in eastern Africa is now regarded as an important case study in discussions concerning the nature and effectiveness of climate services provision for health by the World Health Organization/the World Meteorological Organization (WHO/WMO). The case study focuses on the HF Atlas, and is part of a small group of accepted pieces from more than 75 submissions globally that will form the basis of an online e-book, interactive web repository and a hardcopy publication in the six UN languages produced by the WHO/WMO.
(6) Addressing 'Environmental safety and welfare' as an overarching theme
By contributing to enhanced understanding of links between environmental, including climate, change and the emergence and spread of the three target VBDs, by improving related outbreak early warning systems and by having a relevance beyond the geographic boundaries of the study area – including raising awareness of the potential health and economic impacts of climate change in the developing world - HF addressed environmental safety and welfare as an overarching theme.
Moreover, by communicating the potential links between environment, including climate, change and animal and human health, HF added further weight to arguments in favour of the need to mitigate environmental/climate change impacts, through the reduction of greenhouse gas emissions and increased environmental protection and wise use of resources, and through supporting adaptation measures among the most vulnerable human populations.
(7) Assessing, reducing and preventing tensions and conflicts related to the depletion of natural resources and environmental services [...of all means including technologies, earth observation, modelling and socio- economic research approaches...]
The HF project utilised Earth observation and environmental change technologies, in combination with existing data, to assess, monitor and model dynamic environmental conditions – and their potential implications for VBD burdens – in eastern Africa. The multi-disciplinary and consultative approach, adopted in the results generated, are fully compatible with ensuring sustainable development and the wise exploitation of natural resources in the study region.
Moreover, HF obtained input from decision-makers associated with the VBDs in the EAC region through a series of stakeholder engagement workshops. This series was an important foundation for the development of DSFs (WP 5). By working with the EAC, which is mandated to improve the wellbeing of citizens in eastern Africa, HF contributed to easing tensions within the region, and to the maintenance of good relationships within and between countries comprising the EAC.
(8) Consideration given to the various geographical, sectoral and cultural differences which exist within Africa, and broader socio-economic factors
The HF study area incorporated the eastern African countries of Burundi, Kenya, Rwanda, Tanzania and Uganda. The boundaries of the study area therefore equate to those of the EAC. The EAC is the regional intergovernmental organisation of the republics of Burundi, Kenya, Rwanda and Uganda and the United Republic of Tanzania. Regional cooperation towards a completely healthy society within member states is a key, stated aim of the EAC.
The choice of eastern Africa as a study area acknowledges the necessity of willingness at all levels of society – including supra-national organisations such as the EAC - to implement the findings and increased understanding that accrue from scientific research in order to mitigate effectively the negative health and economic effects of environmental, including climate, change. The cooperation of supra-national organisations is required because of the transboundary dimensions of environmental change causes and effects, the latter including health impacts. The study area also serves as a useful model in which to examine interlinkages between environment and water-related VBDs in Africa. A range of altitude- and latitude-related environmental conditions, including extensive highland areas and a lowland coastal plain, and a range of humidity, from more or less permanently humid highlands to arid and semi-arid plains, are accommodated within the study area. Moreover, large freshwater bodies and extensive wetlands, some of which are only now being exploited for food production, are also present.
Eastern Africa can also be viewed as a model of the African continent as a whole, in terms of human life. Environmental diversity in the study area is matched by a rich variety of human populations, languages, cultures and religions, and unevenness in the distribution of settlements, economic activities and investment in health. Highlands in the study area support disproportionately high densities of human populations. Many of the people in the study area are living in rural areas (rural population densities in some parts of the study area are among the highest in the world) as subsistence farmers. As such they are often distant from health and veterinary services. Parts of the study area have also experienced political instability in the relatively recent past, one consequence of which is the presence of temporary camps for displaced people, while urbanisation has also become a major factor. These are all factors that will influence the health effects of environmental change.
Within the project itself, there has been a strong representation of Africa-based partners, including in positions of responsibility, and stakeholders in the research. Several Work Packages were jointly-led by representatives from Africa-based institutional members of the HF consortium, while the Expert Review Panel was chaired by the representative of NUR (now University of Rwanda) in the project. Moreover, three of the six partners’ meetings were held in eastern Africa – including in Arusha, Tanzania, in order to facilitate attendance by a representative of the EAC, as was one of the two research symposia and one of the two Early Stage Researcher Generic Skills and Networking workshops. All three meetings with stakeholders were held in eastern Africa. Two of the four PhD scholarships funded through the project when to Africa-based students. Finally news releases and summaries of HF research and results were made available in both English and Kiswahili, while project updates were frequently sent to stakeholders in the research. More than 30% of the c. 40,000 hits on the project website (www.healthyfutures.eu) have originated in Africa – with Kenya (over 6000 hits) and Rwanda (over 4500 hits) showing the greatest online interest in the project, according to the number of hits.
(9) Training activities and exchange of staff
HF had a strong commitment to training activities. There were two Early Stage Researcher (ESR) Generic Skills and Networking workshops held:
- The first took place in Arusha, Tanzania in May 2012. Eight ESRs funded through HF participated in the workshop, which was also attended by seven research trainers.
- The second took place at the ICTP, Trieste, Italy in April 2013, and included attendance at the Spring School on Modelling Tools and Capacity Building in Climate and Public Health in the ICTP. The content of this school was of particular relevance to the work being carried out in the HF project. Six ESRs from five of the HF partner institutions participated in the two-week school, which was attended by more than fifty people in total (the vast majority from developing world countries). In addition, several other members of the HF team contributed to teaching at the school.
Four PhDs, funded through HF, has led to the publication of several papers in high impact, internationally refereed journals, in addition to theses. The titles of the PhD thesis research and the students concerned are:
a. ‘Modelling the effects of temperature changes on Schistosoma mansoni transmission’ completed by Nicky McCreesh at Durham University.
b. ‘Integrated spatial indicators for modeling, exploring and visualising vulnerability to vector-borne diseases’ completed by Michael Hagenlocher at PLUS.
c. ‘A simulation model of Rift Valley fever transmission in Kenya’ completed by John Gachohi at the ILRI.
d. ‘Climate change and malaria in Rwanda: Spatial assessment of social vulnerability at different scale levels’ completed by Jean Pierre Bizimana at the University of Rwanda (UR) - The first ever PhD to be awarded in Science by the UR.
(10) Innovative management and governance tools and adaptive technologies suitable for the relevant authorities and stakeholders in Africa
As mentioned above, HF developed and implemented innovative management and governance tools through its online Geoportal and DSTs. These were designed and implemented together with key local partners – essential players in sustaining the technologies after the project end - but also through the integration of regional and local stakeholders through stakeholder engagement workshops. Furthermore, environmental change and health in eastern Africa was chosen as an important case study for Climate services for health by the World Health Organization/the World Meteorological Organization (WHO/WMO). The case study focuses on the HF Atlas, and is part of a small group of accepted pieces from more than 75 submissions globally. The final expected outputs of the Climate Services for Health project comprise an online e-book, interactive web repository and a hardcopy publication in the six UN languages.
(11) Integration of local stakeholders, and/or regional actors, and the necessary networking; clustering and coordination activities between the relevant selected projects
HF partners consulted actively with local and regional stakeholders throughout the project. Besides strategic involvement of key local stakeholders in a number of engagement and training workshops for the refinement of DSTs, partners also networked with regional actors. For example, partners from TCD travelled to Burundi to investigate the country’s archival collection and to establish better links with the country, meeting representatives from Ministère de la Santé Publique et de la Lutte contre le Sida (Minisanté) (Ministry of Public Health and Fight against AIDS) and the Programme National Intégré de Lutte contre les Maladies Tropicales Négligées et la Cécité (PNIMTNC) (National Integrated Program for Neglected Tropical Diseases and Blindness Control) amongst others. Burundi is the only country member of the EAC that had no institutional representatives in the HF project.
Furthermore, HF with its sister project QWECI (which focused on western Africa) jointly coordinated a symposium at the 4th annual East African Community Health & Scientific Conference, Kigali, Rwanda in 2013. Both projects also appeared together in the video documentary entitled Health and Climate Change in Africa, produced by Africa Turns Green, a charity that showcases the work of African green entrepreneurs who are protecting their environment.
As previously mentioned, within the project itself, there has been a strong representation of Africa-based partners, including in positions of responsibility, and stakeholders in the research. Several Work Packages were jointly-led by representatives from Africa-based institutional members of the HF consortium, while the Expert Review Panel was chaired by the representative of NUR (now University of Rwanda) in the project. Moreover, three of the six partners’ meetings were held in eastern Africa – including in Arusha, Tanzania, in order to facilitate attendance by a representative of the EAC, as was one of the two research symposia and one of the two Early Stage Researcher training and networking workshops. All three meetings with stakeholders were held in eastern Africa. Two of the four PhD scholarships funded through the project when to Africa-based students. Finally news releases and summaries of HF research and results were made available in both English and Kiswahili, while project updates were frequently sent to stakeholders in the research. More than 30% of the c. 40,000 hits on the project website (www.healthyfutures.eu) have originated in Africa – with Kenya (over 6000 hits) and Rwanda (over 4500 hits) showing the greatest online interest in the project, according to the number of hits.
Main Dissemination Activities and the Exploitation of Results
HF results were exploited through the prompt publication of research findings in a diversity of fora, including high impact scientific journals, webpages, frequent news releases, workshops and international conferences. HF deliverables were disseminated in English and, where relevant, material was translated into Kiswahili, which is also widely spoken in the study area. The variety of platforms and forms of media used – and the hosting of deliverables on an open access, project website - maximised the impact of the results of the project. Main dissemination activities included:
- HF website - has to date received over 40,000 hits and is available in both English and Kiswahili (http://www.healthyfutures.eu/)
- Seven contractual HF news releases - available in both English and Kiswahili and additional news releases disseminated through such channels as AlphaGalileo (http://www.alphagalileo.org) the EC’s CORDIS News services (http://cordis.europa.eu/news/) CORDIS WIRE, Twitter, LinkedIn and AquaTT’s Training News (also available on project website)
- Project factsheet - available in both English and Kiswahili (also available on project website)
- Seven HF newsletters - non-contractual (also available on project website)
- Four project updates - distributed to all stakeholders listed in the Stakeholder Database (also available on project website)
- A non-technical, tri-fold project brochure - available in both English and Kiswahili developed for general release outlining the project (also available on project website)
- A short promotional video - (c.90 seconds) summarising the HF project and hosted by VIMEO aimed at raising awareness of the project, viewable at www.vimeo.com/70318624 or on the project website
- Two promotional items: a key-ring with bottle opener and torch, and post-it notes - which include the HF logo and website address designed to increase the brand awareness of the project. These have been distributed at a number of stakeholder engagement workshops and events over the course of the project.
- A short documentary “Health and climate change in Africa” - produced by the charity, Africa Turns Green and HF contributed to its completion. The documentary was released in August 2014 and is available to view on the project website (http://www.healthyfutures.eu/index.php?option=com_k2&view=item&layout=item&id=176&Itemid=276&lang=en) and on YouTube (https://www.youtube.com/watch?v=oYL4Nc-qnKE). The video focuses on the collaborations between HF and its sister project, QWeCI, and outlines how both projects have worked to increase capacity in Africa and reduce the impacts of the target VBDs in the study region.
- HF Symposium - Jointly coordinated by members of HF and its sister project QWeCI and held at the 4th Annual East African Health and Scientific Conference in Kigali, Rwanda in March 2013. The conference programmed comprised four sub-themes (Maternal and Child Health, Non Communicable Diseases and Trauma, Health Systems Strengthening, and Quality of Health Care) and four symposia: HIV and AIDS; Integrated disease surveillance and disaster preparedness; Tobacco Control; and the HF/QWeCI symposium on ‘Environment and Health in Africa: climate and vector-borne diseases’. The symposium was well-attended and brought together researchers at the cutting edge of efforts to understand the relationships between health and environment, and in particular the links between climate and VBDs in Africa. There were a total of 17 presentations in this session, including the plenary by Dr. Margaran Bagayoko (Protection of Human Environment Programme, World Health Organization Regional Office for Africa) whose talk centred on the potential of climate-based early warning systems for improved management of VBDs in Africa. The presentations that followed ranged in topics related to the symposium title, from speakers who had travelled from other parts of Africa and from Europe and Asia.
- HF-focused and organised/hosted session - at the ‘Impact of Environmental Changes on Infectious Disease’ (IECID) Conference, Sitges, Spain in March 2015
- Scientific Papers - 19 papers published in scientific journals with one more awaiting decision. Two submissions in a WHO-WMO special publication on climate services and health. One special issue of the journal Geospatial Health is currently in preparation.
- EC Promotion - HF has featured prominently in two widely circulated magazines produced by the EC: Horizon (“Predicting disease outbreaks in East Africa”) and Research EU (“Anticipating climate change, tackling disease”) and most recently through the EC Horizon 2020 website (http://ec.europa.eu/programmes/horizon2020/en/news/mapping-effects-climate-change-deadly-diseases).
- Oral presentations of HF results and networking - partners have carried out dissemination activities through attendance (and in a lot of cases presentation of HF results) at many key conferences and meetings, details of which have been provided in the participant portal.
- Case study in a special publication on Climate and Health (WHO/WMO) – HF has been included as a case study in the World Health Organization/the World Meteorological Organization (WHO/WMO) special publication on Climate Services for Health. The case study focuses on the HF Atlas, and is part of a small group of accepted pieces from more than 75 submissions globally. The final expected outputs of the Climate Services for Health project comprise an online e-book, interactive web repository and a hardcopy publication in the six UN languages, produced by the WHO/WMO.
List of Websites:
The address of the public website is: www.healthyfutures.eu
The following is a list of the email contact details and affiliations of HF project partners:
[Partner number (Beneficiary Short name, Country) - Contact person]
P1 (TCD, Ireland) - Laragh Larsen llarsen@tcd.ie/Gayle McGlynn mcglyng@tcd.ie
P2 (ICTP, Italy) - Adrian Tompkins tompkins@ictp.it/Diro Gulilat Tefera gtefera@ictp.it/Felipe de Jesús Colón-González fcolon_g@ictp.it/Riccardo Biondi rbiondi@ictp.it/Susanne Henningsen hennings@ictp.it
P3 (PLUS, Austria) - Peter Zeil peter.zeil@sbg.ac.at/Stefan Kienberger stefan.kienberger@sbg.ac.at/Michael Hagenlocher michael.hagenlocher@sbg.ac.at
P4 (SMHI, Sweden) - Colin Jones Colin.Jones@smhi.se/Grigory Nikulin grigory.nikulin@smhi.se/Monica Wallgren Monica.Wallgren@smhi.se
P5 (UoN, Kenya) - Winnie Mitullah wvmitullah@swiftkenya.com/Eric Othieno Nyanjom othieno_n@yahoo.co.uk/Fredrick Obonyo Mukanga fredmukanga@yahoo.com
P6 (AQUATT, Ireland) - David Murphy david@aquatt.ie/Olivia Daly olivia@aquatt.ie
P7 (ILRI, Kenya) - An Notenbaert A.notenbaert@cgiar.org/Jeffery Mariner j.mariner@cgiar.org/Bernard Bett b.bett@cgiar.org/Wachira Theuri w.theuri@cigar.org/Jusper Kiplimo jusronohk@gmail.com/Nancy Ajima n.ajima@cgiar.og/John Gachohi j.gachohi@cgiar.org
P8 (UR, Rwanda) - Theophile Niyonzima tniyonzima@nur.ac.rw/Jean Pierre Bizimana jpbizimana@nur.ac.rw/Jean Damascene Mazimpaka jdmazimpaka@nur.ac.rw/Rachel Murekatete rmurekatete@nur.ac.rw/Caritas Thereza Gasengayire tgasengayire@nur.ac.rw
P9 (SEI Tanz/York,Tanzania/United Kingdom) - Stacey Noel stacey.noel@sei.se/Neela Matin neela.mation@sei.se/Victor Kongo victor.kongo@sei.se/Richard Taylor richardtaylor.sei@googlemail.com/Sukaina Bharwani sukaina.bharwani@sei.se
P10 (Community Health, Uganda) - Agaba E. Friday agabafriday@hotmail.com/Didacus B. Namanya didamanya@yahoo.com
P11 (KEMRI, Kenya) - Andrew Githeko githeko@yahoo.com/Diana Karanja Diana@cohesu.com
P13 (UCT, South Africa) - Bruce Hewitson: hewitson@csag.uct.ac.za/Lisa Coop: lcoop@csag.uct.ac.za
P14 (UDUR, United Kingdom) - Mark Booth mark.booth@durham.ac.uk/Nicky McCreesh Nicky.mccreesh@durham.ac.uk/Dajana Dzanovic dajana.dzanovic@durham.ac.uk/Gary Mitchell Gary.Mitchell@durham.ac.uk
P15 (UNILIV, United Kingdom) - Andy Morse A.P.Morse@liverpool.ac.uk/Cyril Caminade Cyril.Caminade@liverpool.ac.uk/Anne Jones Anne.Jones@liverpool.ac.uk/Scott McGee S.Mcgee@liverpool.ac.uk
P16 (NUS, Singapore) - David Taylor geodmt@nus.edu.sg
HEALTHY FUTURES (Health, environmental change and adaptive capacity: mapping, examining and anticipating future risks of water-related vector-borne diseases in eastern Africa [HF]), a FP7 Cooperative project, aimed to build a disease risk mapping system for three water-related high-impact vector-borne diseases (VBDs) (malaria, Rift Valley fever (RVF) and schistosomiasis) in eastern Africa. Projections of future risk were based on a combination of environment, notably climate, change data, output from dynamic disease models and spatially explicit assessments of vulnerability.
The major successes of HF include:
1) Development of a publically accessible Geoportal which facilitates searching and use of data produced in the project.
2) Production of new, state-of-the-art dynamic disease models for malaria, RVF and schistosomiasis. A second dynamic disease model for RVF has also been developed through the project that allows the efficacy of different interventions (e.g. vaccination campaigns) to be examined. These models have either been published or are in the process of being published in scientific journals.
3) Release of the HF Atlas, an interactive, web-based mapping and decision support tool (DST), built within an open-source framework and aiming to provide meaningful and guided access to information on environmental change, disease hazard and risk, and vulnerability for the three target VBDs in eastern Africa. An important aspect of the Atlas is use of climate models and future scenarios referred to in the latest (the 5th) IPCC Assessment Report.
4) Development and release of Decision Support frameworks (DSF) for the three target diseases in consultation with relevant decision makers in the East African Community (EAC) region.
5) Four PhDs, including the first ever PhD to be awarded in Science by the University of Rwanda (UR).
6) Extensive dissemination and exploitation of results through various channels including: 1) HF website receiving ~40,000 hits to date (http://www.healthyfutures.eu/) 2) seven contractual HF news releases, 3) project factsheet, 4) seven HF newsletters, 5) four project stakeholder updates, 6) a non-technical tri-fold project brochure, 7) inclusion in EC-produced promotional material and websites, 8) a short promotional video (www.vimeo.com/70318624) 9) two promotional items (a key-ring with bottle opener and torch, and post-it notes to increase the brand awareness), 10) contribution to a short documentary “Health and climate change in Africa” focusing on the collaborations between HF and its sister project, QWeCI - produced by the charity, Africa Turns Green (https://www.youtube.com/watch?v=oYL4Nc-qnKE) 11) HF Symposium (jointly coordinated by members of HF and QWeCI and held at the 4th Annual East African Health and Scientific Conference in Kigali, Rwanda in March 2013), 12) HF-focused and organised/hosted session at the ‘Impact of Environmental Changes on Infectious Disease’ (IECID) Conference, Sitges, Spain in March 2015, 13) 19 papers published in scientific journals with one more awaiting decision. Two submissions in a WHO-WMO special publication on climate services and health. One special issue of the journal Geospatial Health is currently in preparation, 14) oral presentations of HF results and networking through attendance and presentation at many key conferences and meetings and 15) inclusion as a case study in the World Health Organization/the World Meteorological Organization (WHO/WMO) special publication on Climate Services for Health.
With its integrated holistic approach, HF has considerably enhanced current understanding and dynamic modelling of links between the three target, water related VBDs and environmental change drivers and effects. As stated in the external evaluator’s report commissioned by project members, “HF has been an ambitious project that successfully joined the divergent scientific interests of climate specialists and health practitioners in a single project to address the health impacts of climate change on three VBDs in the EAC. It has provided key scientific information and models on present and future climate impacts on public health and the possibility of further development of the Atlas-Metadata portal as a practical research and learning tool for stakeholders in the EAC.”
Project Context and Objectives:
HEALTHY FUTURES Context
The health effects of future environmental, including climate, changes have been projected to be substantial, often negative and to vary geographically. The effects will be felt most acutely among the poorest members of society, who already carry a disproportionately high share of the burden of environmentally sensitive diseases, with sub-Saharan Africa a focus of adverse health impacts. Environmental change also has implications for health in the developed world, as changes in environment drive the emergence of new diseases, and changes in the distribution and epidemic potential of existing infectious diseases. Environmental change will impact health in a multitude of ways. Concern has, however, tended to focus on the future distribution and spread of infectious diseases, and in particular the negative health impacts of changes in transmission and outbreaks of VBDs as a result of anthropogenic climate change. The impacts may be direct, in terms of outbreaks of disease among human populations, or indirect, in the form of outbreaks of diseases that affect domesticated animals or plants, and therefore jeopardise food security, agriculture-based economic activities and trade.
These concerns provided the motivation for HF, which aimed to build a disease risk mapping system for three water-related high-impact VBDs (malaria, RVF and schistosomiasis) in eastern Africa, accounting for environmental/climatic trends to project future risk. The project involved a comprehensive, inter-disciplinary consortium of health, environment, socio-economic and climate experts in addition to governmental health departments, and throughout the four years of EU FP7 funding, sought and obtained input from stakeholders, including health decision makers. Concentrating on the three target VBDs in eastern Africa, HF produced improved understanding of the sensitivities to environmental conditions, new state-of-the-art dynamic disease models, down-scaled environmental change model outputs and scenarios of future conditions from the IPCC’s 5th Assessment Report (2013 & 2014), and a novel, interactive, online disease risk-mapping platform (the HF Atlas).
HEALTHY FUTURES Objectives
The main aims of HF were to develop (1) a basis for anticipating future environmental changes and their impacts on water-related VBDs in eastern Africa, and (2) the capacity of health and veterinary services in the study area to respond to early warnings of future outbreaks. These aims were met through seven, interlinked Work Packages (WPs) that:
1) provided effective management of the research (WP1);
2) ensured that full benefits are derived from synergies between this research project and other related research projects (WP7);
3) developed policy-relevant simulations of future levels of environmental correlates of the three VBDs and mapped spatial variations in vulnerabilities across the region (WPs 2,3,4);
4) combined results of the work outlined in 3) with improved understanding of the epidemiology of the targeted VBDs to model, dynamically, VBD emergence and spread and to improve (constrain) assessments of risks of future outbreaks (WPs 3,4);
6) enhanced the capacity of health and veterinary services at various geographic scales to respond to changes in risk of transmission and outbreaks of the targeted VBDs as a result of a convergence of changing climatic, land use and socio-economic conditions (WP5);
7) ensured that the findings of the research are widely disseminated among stakeholders, policy-makers and the global scientific community (WP6).
One further activity of crucial importance was the training of early stage researchers in the trans-disciplinary fields of environmental change and animal and human health. As part of this, four students (two from Africa and two from Europe) obtained their PhDs through direct support from HF.
Activities carried out per workpackage are detailed below:
WP1 facilitated coordination and management of the project. During M37-48 there were two partner meetings. The fifth partners’ meeting was held from 26-27 February 2014 in Nairobi, Kenya and was hosted by the International Livestock Research Institute (ILRI). The sixth partners’ meeting was held from 9-10 September 2014 in Salzburg, Austria and was hosted by the Paris-Lodron University of Salzburg (PLUS).
WP2 largely involved construction of a project database (‘geoportal’) comprising information drawn from a range of primary and secondary sources, including: historical, socio-economic, migration, settlement and conflict data; earth sciences data; direct climate observations and the results of regional downscaling of global reanalyses; and information on disease-environment relationships. Geoportal is at: http://41.204.190.50/geoportal/catalog/main/home.page
WP3 focused on field-based evaluations of environment and hydrological data constructed in WPs 2 and 4, respectively, with the collection and collation of disease vector/host information, and with the development of new dynamic models and statistical multivariate regression models for infection rates of the three VBDs. Outputs from these models were input to WP4 as part of disease risk mapping. Two PhD theses were also linked to WP3: 1) ‘Modelling the effects of temperature changes on Schistosoma mansoni transmission’ by Nicky McCreesh at UDUR; and 2) ‘A simulation model of Rift Valley fever transmission in Kenya’ by John Gachohi at ILRI/University of Nairobi.
WP4 collated information relating to historical disease drivers and occurrence (collected in WP2), along with knowledge and models of dynamics of the three target VBDs (WP3). Spatial assessments of the current and projected future hazards and risks of the three target VBDs in eastern Africa were combined with a spatially explicit database relating to historical outbreaks with the aim of identifying hotspots of high risk. Projections were closely linked to procedures adopted in the IPCC’s 5th Assessment reports. WP4 focused on two scales: the region comprising the five East African Community (EAC) countries and on smaller sub-national scales. The assessments of hazard and risk informed WP5, through input to discussions on DSF carried out with other consortium members and stakeholders in the findings of the research. The HF Atlas (http://zgis186.geo.sbg.ac.at/hf_atlas/) an interactive, web-based mapping and visualisation tool, was built largely as part of WP4. Two PhD theses were also linked to WP4: 1) ‘Climate change and malaria in Rwanda: Spatial assessment of social vulnerability at different scale levels’ completed by Jean Pierre Bizimana at UR (formerly NUR); and 2) ‘Integrated spatial indicators for modeling, exploring and visualizing vulnerability to vector-borne diseases’ completed by Michael Hagenlocher at PLUS.
WP5 was concerned with identifying and deploying the most appropriate environmental change adaptation strategies and decision support approaches (including relevant tools and information platforms) for adoption in the project. As part of this, input from decision-makers associated with the VBDs in the EAC area was obtained through a series of stakeholder engagement workshops hosted in the study region by HF consortium members.
WP6 focused on the engagement of stakeholders in the research, the effective dissemination of research findings, and the training of Early Stage Researchers involved in the project. As part of this WP, HF led very successful sessions at the ‘4th Annual East African Health and Scientific Conference’ Kigali, Rwanda (March 2013), and at the ‘Impact of Environmental Changes on Infectious Disease’ Conference, Sitges, Spain (March 2015).
WP7 promoted synergy, cross-fertilisation, networking, and coordination with other climate and health-related research projects funded by the EU FP and other non-EU stakeholders and researchers internationally and regionally. The WP also provided independent, external oversight, results review and evaluation, assurance of quality-control, guidance on next steps and use of best practices, including highest compliance with ethical guidelines. The WP operated mainly through an Expert Review Panel (ERP), chaired by an expert in the field and comprising four other members who were external to HF.
Project Results:
WORK PACKAGE 1 - Project Management
Lead beneficiary: AquaTT
All deliverables in WP1, which ensured the efficiency and effectiveness of the work performed within the project, were successfully developed and submitted. Six formal partnership meetings took place throughout the project duration.
WORK PACKAGE 2 - Disease information and database construction
Lead beneficiary: Trinity College Dublin
WORK PACKAGE 2 - Disease information and database construction
Lead beneficiary: Trinity College Dublin
Task 2.1 Collection of historical data
This task involved an investigation of past outbreaks of the three target diseases (malaria, schistosomiasis and RVF) in the eastern African study region (Burundi, Kenya, Rwanda, Tanzania and Uganda). This was principally achieved through archival research involving a wide range of primary sources, including colonial reports, private papers and ministry files. Historical archival data were collected from documents held in the National Archives (London), Rhodes House (Oxford), the Wellcome Unit for the History of Medicine (Oxford), the School of African and Oriental Studies (London), the Kenya National Archives (Nairobi), Ministry of Public Health and Sanitation (Nairobi), the Uganda National Archives (Entebbe) and the National Archives of Tanzania (Dar es Salaam) (most of the data collected from the Tanzania National Archives in 2012, however, was unfortunately lost due to the theft of a laptop). The collected data were coded, compiled in an Access database and visually displayed in Google Earth. The resulting historical database is incorporated in the HF Atlas, publically accessible through the project’s webpage. Thus this information can be, and is being, used to examine the historical framework of outbreaks of the target diseases in the study region in environmental, social and political contexts.
The HF Atlas is at: http://zgis186.geo.sbg.ac.at/hf_atlas/
Task 2.2 Collection of socio-economic data
This task gathered information on current policies, regulations, strategies and challenges in addressing outbreaks of the three target diseases and on other socio-economic data thought to affect, and be affected by, the transmission and maintenance of the diseases in the study region (e.g. present-day migration, poverty, settlement patterns and settlement densities, including informal settlements). In this context the task has generated knowledge (baseline information) on the extent to which outbreaks of malaria and RVF are a reflection of the socio-economic conditions in Ijara sub-county. Among the key findings in relation to malaria are that health education appears to be improving awareness of the benefits of owning mosquito nets in general, and insecticide treated nets (ITN) in particular. However, bednet control programmes are still grappling with distribution, and affordability and equity issues, as well as the trade-off between social marketing and commercial approaches. In relation to RVF, the main barriers restricting the abilities of health and veterinary services to respond to warnings of heightened disease outbreaks were identified. With respect to the vaccination of animals, which is one of the primary methods of preventing RVF outbreaks and the spread of the disease to humans, coverage is not universal. An important finding of the survey was the diversity in the frequency of vaccinations, and the leading reason for non-vaccination was non-affordability. A report based on all findings has been written and these results can better contextualise the socio-economic conditions of the diseases to enable the implementation of more effective response mechanisms in the framework of changing environments.
Task 2.3 Development of land cover/land use/terrain/surface water databases
Under this activity, spatially-referenced data on land cover and elevation, soil types, normalised difference vegetation indices for the period 1999 – 2010, livestock, and precipitation for the entire eastern Africa region were processed by ILRI and used to map the risk of RVF. Additional data on surface hydrology and livelihood zones were compiled only for Kenya.
Under this task, NUR also compiled spatially referenced data on land cover and surface hydrology, as well as other secondary data on land use, elevation and population in eastern Africa, with a focus on Rwanda. These data have been used in the examination of malaria.
Furthermore, all the datasets produced and collected as part of this task were uploaded to an online database which was created within this task. The online database comprises data collected, quality assured and collated within WP 2 as well as other WPs of HF. It includes metadata (Task 2.7) of the produced output datasets as well direct access to data via a File Transfer Protocol (FTP). Where restrictions apply, relevant contacts are identified. Initially only accessible to members of the HF consortium and to non-consortium members of the project’s Executive Committee, the information platform was made open access at the project end and it thus facilitates the dissemination of, and public access to, results and data produced in the project.
Task 2.4 Developing climate databases and regional climate simulations
Four key disease models have been developed entirely or partly within HF and they have various data input requirements. These were provided by both externally generated datasets (e.g. satellite and ground based rainfall observations, land use change projections, global climate model integrations) and by new datasets created within the project research (such as regional climate model simulations, newly downscaled global model runs, and high resolution simulations of surface hydrological conditions). Task 2.4 was concerned with the generation of these datasets, the key ones being the high resolution regional climate simulations.
The two participating institutions (SMHI and ICTP) produced two sets of climate simulations at high resolution. Given the computational requirements and the availability of validation datasets at 10km, a lower resolution target than originally envisaged was set. In this regard, ICTP used its regional climate model (RegCM4), driven by reanalysis, to generate data at 24km resolution, covering an area between 200S to 200N in latitude and 60E to 650E in longitude. SMHI produced a simulation at a resolution of 0.15 degrees (~ 17km) covering an area between 130S to 130N in latitude and 180E to 600E in longitude. The model domain covers all the HF target countries and includes the important surrounding regional features that are known to affect the region's climate, such as the western Indian Ocean and the Congo basin. The model outputs were compared against Climate Research Unit (CRU) observations and ERA-Interim reanalysis to examine how well the model reproduces the spatial, inter-seasonal and inter-annual variability of the region’s climate. The results are reported in deliverable D2.2.
In addition, the project applied a statistical downscaling technique to CMIP5 climate models in order to provide a wider range of lower-resolution benchmark driving fields. Self-Organizing Map based Downscaling (SOMD) is a leading empirical downscaling technique for Africa and provides meteorological station level or gridded data in response to global climate change forcing. The downscaling of a global climate model (GCM) is accomplished by deriving the normative local response from the atmospheric state on a given day (predictors), as defined from historical observed data (predictants). The method recognises that the regional response is both stochastic as well as a function of the large scale synoptics. As such it generates a statistical distribution of observed responses to past large scale observed synoptic states. This statistical downscaling methodology was used to downscale climate change projections from the latest two versions of the Coupled Model Intercomparison Projects (CMIP3 and CMIP5) over the HF study area.
Finally, research was also conducted to produce gridded socio-economic data for use in the dynamical disease modelling, in particular the assessment of population growth on fine-spatial scales. Total population estimates for the baseline period 1950-2010 were produced, and for five different future Shared Socio-Economic Pathways (SSP) for the period 2011-2100. Population estimates for the baseline period and each of the five SSP scenarios were obtained at the national level at five year intervals (from 1950-2010, and from 2015-2100 respectively). Data for the baseline period were retrieved from the United Nations World Population Prospects (UNWPP) 2012. Revision Population data for the five SSPs were obtained from the Inter-Sectoral Impact Model Intercomparison Project. Yearly population estimates were computed for each country using a linear interpolation method. Population change rates were used as scaling factors to compute gridded population data for both the baseline period and the five SSPs. Not all countries present in the gridded dataset were present in the UNWPP and SSP data. The 37 countries not present in these datasets were assigned an average scaling factor per year by taking the mean of the population change rate of the 193 countries in the UNWPP and SSP datasets.
Task 2.5 Bringing Climate Databases to the kilometre-scale for use in disease analysis
According to the original Description of Work (DoW) temperature and surface hydrology data, at ultra-high (1-10km) resolutions were to drive the Liverpool malaria model (LMM). In the course of the project development, however, a decision was reached to develop these modules within the disease models themselves to allow direct and computationally efficient interaction with other components of the disease models. In this way the task accomplished the direct incorporation of the surface hydrology within the new dynamical disease model VECTRI, and also directly downscaled the temperature to the high resolution topography.
The development of the new VECTRI model (a model not envisaged in the DoW) within the project framework allowed the integration of the surface hydrology and temperature downscaling directly into the model design from the outset. As a result, the consortium and scientists more widely now have access to a disease model that explicitly incorporates the direct interaction with the human population, allowing future population growth and urbanisation to be accounted for in disease risk projections, and that also directly incorporates the surface hydrology in a pond parametrisation.
The surface hydrology component has two parts:
1. A fraction of breeding sites associated with permanent water bodies. Initially this was set to a constant, but during the project several methods were investigated such as the use of existing databases including the global lakes and wetlands database of the WWF.
2. A pond parametrisation, which is driven by rainfall (available on scales down to 10km). The pond parametrisation is described in Tompkins and Ermert (2013).
The pond parametrisation is a novel aspect of the VECTRI model that allows for a more realistic association between rainfall and vector densities. The pond parametrisation accounts for topography in the maximum ponding fraction. This implies that while the input rainfall data do not have information below the 10km scale, the ponding fraction can be provided on scales down to 1km accounting for topography. The pond parameterisation has since been validated in comparisons with AMSR-E satellite derived statistics and also ultra-high resolution (10m) simulation with an explicit surface hydrology scheme. The magnitude of the pond fraction is much smaller for VECTRI indicating the need to tune the wmax parameter appropriately with observations, or the need to include soil texture in the parametrisation (see Tompkins and Ermert, 2013). Nevertheless, it is extremely encouraging that the VECTRI parametrisation scheme at 1km is able to almost exactly reproduce the pond temporal evolution. VECTRI tends to over-predict the pond fraction during the rainy season onset, and under-predict it during the main season. This is due to the fact that VECTRI does not account for soil moisture in the calculation of infiltration, which is apparently a second order effect.
The temperature information is also downscaled to the model resolution, up to a finest resolution of 1km as per the DoW, again using the topography to drive a simple lapse rate adjustment scheme. In the earlier version of VECTRI (up to v.1.3.1) the pond temperature is simply related to the atmospheric mean temperature with a fixed offset. This will be updated post project, with the implementation of an optional energy balance that has been developed and tested with in situ pond temperature measurements. The energy balance model is optional since it relies on the availability of radiation fluxes from the driving climate model, which can be subject to considerable biases associated with errors in the cloud fields, and which are not always archived on a daily timestep in the output of archives such as CMIP5. Thus, the option of using a simple offset from atmospheric 2 metre temperatures will be retained.
Task 2.6 Compilation of disease data
Data on RVF outbreaks in Kenya for the period 1912-2010 were obtained from the Department of Veterinary services and Centres for Disease Control (Kenya). These data were used by ILRI to develop Milestone 16 and have also been analysed and presented in various workshops as part of the QWeCI (HF’s sister project) research reports. These records identify the years when outbreaks occurred and include administrative units (province, district, division and area/village) affected. An RVF outbreak was defined as above normal occurrence of abortions, perinatal mortality and hemorrhagic syndrome in livestock with or without human involvement. In most cases, primary cases were screened using real-time reverse transcription polymerase chain reaction (RT-PCR) to satisfy requirements for official declarations. Secondary cases for each outbreak were often diagnosed based on clinical signs. The total number of records in the database is 599,832 given repeated records per division. Similar data have been obtained from Tanzania identifying districts that have reported RVF outbreaks since 1930.
The georeferencing of RVF hotspot sites was also completed in Kenya, to enable a more intensive analysis using ecological niche models. The survey utilised a check list that collected additional data from each site on:
- Vegetation cover
- Types of livestock and wildlife found in each site
- Whether there were both human and livestock cases in the recent 2006/2007 outbreak
- Main livelihood activity
UDUR’s work on this task was linked to WP3 Task 3.3 and formed part of the validation of the schistosomiasis dynamic modelling exercise. Georeferenced data on S. mansoni infection prevalence in human populations were extracted from the open access Global Neglected Tropical Disease (GNTD) database. Prevalence data were available from 2965 records in total, and 594 records when surveys that did not meet stricter inclusion criteria were excluded (‘selected data’). When prevalence data was plotted on a map of the mean (across scenarios) model output ‘infection risk’ at baseline (2006-2015), the prevalence data yielded estimates of prevalence for 19% (279/1470) of grid squares when all data were used, and 7% (100/1470) when selected data were used. Prevalence estimates for each square were calculated from 1-119 (median=4) and 1-37 (median=4) individual estimates when all and selected data were used respectively.
Task 2.7 Creation of an online project database (information platform)
This task was dedicated to the creation of an online metadata entry platform. It was decided to produce this with the ESRI Geoportal Server (version 1.2.2) which is free of charge and open source. The platform (/portal) is hosted by an Apache Tomcat Server (version 6.0.37) which is a virtual server running on a physical server (Ubuntu Server) located at ILRI. This platform has a MySQL database in the background where the (meta-)data are stored. The platform (a HTML/JavaServer Pages based portal) is also fully customisable and has been redesigned by PLUS to meet the needs of the project. For example, it fits the HF colour scheme, while the homepage now has buttons for searching and adding (meta-)data that help the user to more efficiently navigate the portal (http://41.204.190.50/geoportal/catalog/main/home.page). Different login levels are available: administrator, data editors identified for each partner institution, and guest account. However, the entries, which are set to be public, can also be viewed without logging in. Finally a data upload function was developed, whereby data can be uploaded during the fill-out process of metadata. The data are stored in folders and the name of a folder is the ID of the metadata file. This helps the organisation of the data, particularly in case of deleting a metadata record. ICTP is providing the FTP storage. Once the data are uploaded they are reviewed and quality assured by an administrator at NUR. This process of uploading data will continue beyond project end as the metadata portal will be maintained post project.
WORK PACKAGE 3 - Environment-disease transmission relationships & modelling
Lead beneficiary: University of Durham (UDUR)
Task 3.1 Field Studies for the three target diseases
Schistosomiasis (UDUR)
Collecting background data on schistosomiasis in this task also contributed to both task 2.6 (compilation of disease data) and task 3.3 (Development of dynamic models). Briefly. geo-referenced data on S. mansoni infection prevalence in human populations were extracted from the open access Global Neglected Tropical Disease (GNTD) database. Prevalence data were available from 2965 records in total, and 594 records when surveys that did not meet stricter inclusion criteria were excluded from modelling.
Previous field-based work at Lake Albert was reviewed with respect to understanding where major gaps existed in terms of data that could be used to give values to model parameters in task 3.3c. The conclusion of this review was a lack of information on the natural history of snail species responsible for transmission of S. mansoni in the region. To address these gaps UDUR undertook a series of experiments to test how fecundity and longeveity of two species are affected by changes in water temperature. Laboratory experiments were conducted at Vector Control Division in Kampala, Uganda, to estimate Biomphalaria sudanica mortality, fecundity and growth rates at ten different constant water temperatures, ranging from 13-32°C. Field experiments took place on the shore of Lake Albert. Snail cages placed in open water at the Lake Albert field site were used to determine the effects of snail densities on B. sudanica and B. stanleyi mortality and fecundity rates in semi-natural conditions. B. sudanica survival and fecundity were found to be highest at 20°C and 22°C respectively. Growth in shell diameter was estimated to be highest at 23°C in small and medium sized snails, but the relationship between temperature and growth was not clear. The fecundity of both B. sudanica and B. stanleyi decreased by 72-75% with a four-fold increase in population density. Increasing densities four-fold also doubled B. stanleyi mortality rates, but had no effect on the survival of B. sudanica. The optimum temperature for fecundity was lower for B. sudanica than for previously studied species of Biomphalaria. In contrast to other Biomphalaria species, B. sudanica have a distinct peak temperature for survival, as opposed to a plateau of highly suitable temperatures. For both B. stanleyi and B. sudanica, fecundity decreased with increasing population densities. This means that snail populations may experience large fluctuations in numbers, even in the absence of any external factors such as seasonal temperature changes. Survival also decreased with increasing density for B. stanleyi, in contrast to B. sudanica and other studied Biomphalaria species where only fecundity has been shown to decrease.
RVF and Malaria (ILRI)
Field surveys on RVF were conducted in the Ijara study site in Kenya to generate empirical data that have been used in modelling RVF transmission dynamics. These activities included participatory, entomological and serological surveys; they were implemented in the second and third years of the project and their results were reported in deliverable D3.2 entitled: RVF/malaria study site analysis and major findings for RVF & malaria transmission. The main results from this task included:
- Livestock movement patterns and their population structures required for RVF modelling were described
- Mosquito vectors that are prevalent in the area were sampled and characterised
- RVF sero-prevalences in sheep, goats and cattle were estimated
KEMRI organised the collection of in-patient malaria data from the project site in Ijara District hospital for the period 2006-2011. Data on the distribution of insecticide impregnated bednets was also collected. The malaria data were tested for seasonality and seemed to be consisted with trends in rainfall. Monthly Malaria case anomalies were computed. These data were made available to the modeling team.
Task 3.2 Evaluation and development of statistical disease models
ICTP intended to develop a range of statistical models to estimate the relations between schistosomiasis, climate, environmental, and socio-demographic parameters in Uganda in collaboration with Durham University. This was not possible, however, owing to the lack of epidemiological surveillance data for schistosomiasis beyond the prevalence data used in task 3.3c. In addition, the collection of human schistosomiasis incidence data in Uganda was not possible due to issues for obtaining the ethical clearance to conduct a site analysis. Consequently, the development of statistical models was exclusively conducted on malaria data, which were more freely available in the project.
ICTP developed a range of statistical malaria models to examine how malaria incidence changes as a function of climatic, environmental, demographic and socioeconomic predictors in Uganda and Rwanda. The key component of this task was a dataset of epidemiological surveillance data (i.e. number of malaria cases) retrieved from our partners in the ministries of Health of Uganda and Rwanda. ICTP also collected environmental, socioeconomic and demographic data from various sources, including but not restricted to bureaus of statistics, remote sensing datasets, climate reanalyses, international malaria programmes, and censuses. It was possible to retrieve climate data for the same period of time, and at the same spatiotemporal resolution as the epidemiological data. Socioeconomic data were only available for selected periods of time, and for selected geographical regions. These features of the socioeconomic data prevented their incorporation into the statistical models. They also compromised the skill of the statistical models. In the end, a dataset of malaria, climate and non-climatic predictors was developed covering the whole of Rwanda and Uganda. To our knowledge, these are the longest and more spatially diverse malaria-related datasets yet assembled for both countries.
Some problems were inherent to the epidemiological data. First, some entries were missing or duplicated, and so statistical interpolation methods were implemented by ICTP to estimate a value for those data. Second, some of the entries in the malaria dataset had levels that were significantly lower or higher than the surrounding entries. For most of these entries, there was no available information in the ministries of health as to whether those values corresponded to real observations or were simply data entry errors. Since it was impossible to determine the veracity of such values, an algorithm was developed by ICTP to filter the data and reduce the impact of those suspicious entries in the performance of statistical models. Third, the length of the malaria time series was relatively short (just over a decade), posing challenges for estimating significant effects of climate drivers on malaria incidence using statistical models.
Statistical models were then developed using a time series cross-validation algorithm to detect the model specification resulting in the lowest prediction errors. These models accommodated the delayed and nonlinear effects of climate variables on malaria incidence. Given that data were unavailable for key malaria predictors, the statistical models incorporated random effects terms to account for the potential effects of unobserved variables. This modelling approach was used on both the Uganda and Rwanda datasets.
The risk of infection estimated by the model was evaluated through comparisons with malaria observations in both countries. Model outputs were found to reproduce the spatiotemporal dynamics of malaria in both countries. The estimated relationships between malaria and climate were statistically significant in both countries, and the functional form of such relations was in agreement with previous research. The relations between malaria and socioeconomic development, however, were significant only for a small number of predictors. This finding may indicate that some of these variables are not important for describing changes in malaria incidence. However, it is likely that these results are due to the very short time series of socioeconomic data preventing the model from estimating significant effects. The significant relationships estimated between malaria incidence and socioeconomic drivers have to be cautiously considered too, because they arise from data aggregated at a different temporal resolution from the epidemiological data.
Task 3.3 Evaluation and development of dynamical disease models
Task 3.3a - Malaria
This task involved the development and application of dynamical climate-dependent malaria transmission models for eastern Africa. The two malaria models produced or developed through HF, LMM and VECTRI, were driven using climate projections provided by a large ensemble of climate models derived from multiple bias correction techniques. A large inter-comparison exercise was carried out (Task 3.4) to provide a quantitative method for analysing the impact of the long-term effects of climate change on malaria transmission in eastern Africa.
The LMM, initially formulated in 2004, was further developed within HF with the addition of components representing transmission from a chronically infected immune population at the start of the season and influx of mosquitoes from permanent water bodies. These new components were tested, along with different LMM parameter settings, against records of observed malaria transmission parameters in Senegal and in South Africa. Additionally, LMM-simulated malaria transmission and variability patterns driven by various climate datasets have been compared with malaria maps from the Malaria Atlas Project 2010 (MAP¬2010), a statistical analysis of malaria prevalence based on observations. LMM is largely able to reproduce the spatial distribution and seasonal variability of malaria incidence in the eastern Africa when driven by observational climate timeseries (from the ECMWF interim reanalysis, Global Precipitation Climatology Project and the Tropical Rainfall Measurement Mission Multi-satellite Precipitation Analysis) and compared with Mapping Malaria Risk in Africa (MARA) and MAP¬2010 products. Analysis comparing LMM hindcasts driven by ECMWF’s state of the art seasonal forecast system (System 4) and the observationally driven simulations indicates some probabilistic skill for high (above upper tercile), above average (above median) and low (below lower tercile) malaria incidence with values (>0.7) of the area under the Relative Operator Characteristic (ROC) curve with a forecast lead time of approximately three months.
Different dynamical and statistical malaria models (MIASMA, MARA and UMEA) were also compared with LMM and VECTRI, both at the global scale and for Africa. This was carried out for both the recent context and under different climate change and population growth scenarios (consistent with Task 3.4). Results showed that the climate might become more suitable for malaria transmission in the highlands of eastern Africa and, to a lesser extent, less suitable over west Africa. This first multi-malaria model comparison exercise was carried out within the Inter-Sectoral Impact Model Intercomparison Project (ISI-MIP), QWeCI and HF frameworks and this involved research teams at UNILIV, ICTP, UMEA, Maastricht University and the London School of Hygiene and Tropical Medicine.
A large ensemble of long-term climate projections under various emissions scenarios were integrated with the epidemiological models LMM and VECTRI to provide a range of projections of climate-related malaria risk in eastern Africa over the next century. This undertaking incorporated the two malaria transmission models, a range of emission scenarios (as defined by the IPCC), and four separate bias correction techniques comprising a total of 27 climate model datasets, many of which were developed by partners in HF. This implementation allowed us to compare between the two malaria models, two of the common RCPs and three of the bias-corrected climate model streams. Analysis was performed on these multi-model malaria risk projections by calculating the mean, spread (standard deviation) and relative differences in time (anomalies). For the historical period (1980 to 2005) both the malaria models were compared with MAP2010. Both LMM and VECTRI overestimate the prevalence for this period as these models represent the effects of climate on malaria transmission only, and do not take into account interventions that have been carried out during this period or any immunity effects. Despite the wide array of driving input data for the disease model projections there was some consistency to be found in the results, which was tested by quantifying multi-model agreement. Multi-model agreement is said to occur for a given period if all the component mean values for that period are within two standard deviations of the overall mean. For example, if all the Global Climate Models (GCMs) used within a modelling stream (particular bias-correction technique) have mean values within two standard deviations of the modelling stream mean then the multiple climate models are deemed to share a certain level of agreement. The multi-model spread (uncertainty) is generally higher near the epidemic fringes of the distribution (for low prevalence values).
Numerical malaria models solely driven by climate parameters cannot explain the decline in malaria endemicity observed over the 20th century at the continental scale (mostly due to human intervention, land use management and other socio-economic parameters). Generally, more significant climate impacts and greater consistencies across the multi-model ensemble are shown under higher emission scenarios (RCP8.5) for the end of the 21st century. There is no clear agreement between models for near-term projections (2020s). The largest uncertainties are associated with the methodology, e.g. the numerical malaria models, as opposed to the driving climate model projections. The uncertainties related to the spread in rainfall and temperature changes as simulated by the climate models are large over the northern fringe of the Sahel. This is consistent with the diverging rainfall projections shown by the various climate models over these regions. The uncertainties related to the emission scenarios are relatively small, but they linearly grow as a function of time over highland areas. For future climate projections, all numerical models generally agree on the increase in climate suitability for malaria transmission over the eastern African highlands of the Rift Valley and Ethiopia. Furthermore there appears to be general agreement between models on a southward shift of the epidemic fringe that lies to the north of the central African endemic belt.
Task 3.3b RVF
As anticipated in the HF description of work, the sister EU FP7 project QWeCI developed a RVF version of the LMM, named the Liverpool Rift Valley fever model, or LRVF. This new epidemiological, dynamic transmission model underwent further development through detailed parameterisation in this project using literature-based data and data collected as part of HF. The LRVF model was further developed as part of Task 3.3 to include a complex dynamical host component to take into account the greater impact of transmission dynamics that livestock has for this disease. The relationship between environmental conditions and RVF outbreaks in Mauritania has also been investigated. Results show that the four reported RVF outbreaks over Mauritania (1998, 2003, 2010 and 2012) were preceded by similar rainfall conditions e.g. a rainless period lasting at least seven days followed by heavy precipitation. This result is consistent with former published studies focusing on RVF outbreaks over Senegal and was used to inform LRVF development.
Disease model development carried out as part of the FP7 QWeCI project resulted in the creation of a generalised disease-modelling library, EpiCS. EpiCS (EPIdemiological modelling toolkit for Climate Sensitive disease) is a C/C++ library of generic functions that allows any host or vector process (such as mortality, population growth, biting rate) to be associated with any transmission model structure. The toolkit has been tested by recreating the existing LMM. A prototype two-vector, single host dynamic RVF model was also implemented using the EPICS toolkit. The amplifying vectors for RVF, Culex spp mosquitoes, were modelled using the same structure as the dynamic mosquito model in LMM. The population dynamics of the reservoir vectors for RVF, Aedes spp mosquitoes, were modelled using a drying/wetting trigger for the larval stage, driven by rainfall, a physically based version of the rainfall criteria used in other RVF models. In LRVF, the eggs of Aedes require a (configurable) period of drying and then re-wetting before they can proceed to larval and pupal stages. Prototype LRVF model parameters were taken from the literature, where available, and LMM settings were used as a starting point for parameters for which there are no published values. Plugins derived from the new version of the LMM and the newly developed LRVF model were embedded and tested for the Disease Model Cradle (DMC) GUI. The plugin deployment procedure used on each operating system platform (Linux, Mac OS X and Windows) has also been streamlined.
Task 3.3b required the development of a new climate driven dynamical model of RVF transmission based on LMM. Following the formulation of an initial prototype model, redevelopment of the host component of the LRVF model and model parameterisation was finalised. A more complex host component for LRVF was developed to take into account the greater impact of transmission dynamics that livestock has for RVF compared with malaria. In the new host module, the population is divided into two sub-populations based on age division due to the much higher susceptibility of young livestock to RVF infection. The sub-populations are dynamically coupled via new births and the young mature into adult livestock. The LRVF model was qualitatively calibrated based on past RVF epizootics in Kenya and Tanzania, in collaboration with ILRI, following anomalously high rainfall events using reanalysis climate data. Various parameterisations of the model were used in an iterative process that combined qualitative, inverse modelling and knowledge of the system (both local and from literature) to produce Culex EIR spikes in the expected years based on observations. This process guided previously unmeasured parameters away from the default LMM-derived set-up to a unique LRVF based parameterisation via an improved understanding of model parameter sensitivity.
The new LRVF model was integrated with climate model data in order to provide long-term projections for RVF dynamics in eastern Africa. LRVF was driven by five different GCMs comprising the models calibrated by ISI-MIP for moderate and extreme emission scenarios (respectively Representative Concentration Pathway (RCP)4.5 and RCP8.5). These climate models provide daily temperature and precipitation data for the period 1980 to 2099. The outputs of LRVF provide an estimate for areas that are vulnerable to RVF epizootics as a result of the state of the climate and predicted current livestock immunity. By driving LRVF with climate model projections, the long-term changes in epizootic susceptibility for eastern Africa were predicted. The precise nature of these changes was dependent on the climate model used but certain dynamics exhibited general trends from which a consensus could be drawn. In the long-term, LRVF considers potential shifts in regions that are especially suited to enzootic and epizootic activity based on climate-driven projections of vector dynamics. Results of LRVF suggest that there is a threat of the highland areas of eastern Africa becoming more suitable for RVF transmission as temperature increases and these regions support greater populations of the RVF vectors. Whether regions susceptible to increased RVF transmission in the future are capable of supporting a major outbreak depends on shorter timescale rainfall dynamics, as well as on the local vector population, state of host immunity and level of vulnerability.
Task 3.3c Schistosomiasis
An agent-based model of the temperature-sensitive stages of the Schistosoma and intermediate host snail life-cycles, parameterised using data from S. mansoni and Biomphalaria pfeifferi laboratory and field-based observations, was developed through HF (Task 3.1). Infection risk is calculated as the number of cercariae in the model, adjusted for their probability of causing infection. The number of snails in the model is approximately estimated to remain constant between 15–31°C. Outside this range, snail numbers drop sharply, and the snail population cannot survive outside the range 14–32°C. Mean snail generation time decreases with increasing temperature from 176 days at 14°C to 46 days at 26°C. Human infection risk is highest between 16–18°C and 1 pm and 6–10 pm in calm water, and 20–25°C and 12–4 pm in flowing water. Infection risk increases sharply when temperatures increase above the minimum necessary for sustained transmission. The model suggests that, in areas where S. mansoni is already endemic, warming of the water at transmission sites will have differential effects on both snails and parasites depending on abiotic properties of the water-body. Snail generation times will decrease in most areas, meaning that snail populations will recover faster from natural population reductions and from snail-control efforts. Links between the ecological properties of transmission sites and infection risk that could significantly affect the outcomes of interventions designed to alter water contact behaviour, with such interventions more likely to reduce infection levels at river locations than lakes, where infection risk remains high for longer. In cooler areas where snails are currently found, increasing temperatures may significantly increase infection risk, potentially leading to new, high-intensity foci of infection.
The model was run using low, moderate and high warming climate projections over eastern Africa. For each climate projection, eight model scenarios were used to determine the sensitivity of predictions to different relationships between air and water temperature, and different snail mortality rates. Maps were produced showing predicted changes in risk as a result of increasing temperatures over the next 20 and 50 years. Comparing baseline model output with prevalence data indicates suitable temperatures are necessary but not sufficient for both S. mansoni transmission and high infection prevalences. All else being equal, infection risk may increase by up to 20% over most of eastern Africa over the next 20 and 50 years. Increases may be higher in Rwanda, Burundi, south-west Kenya and eastern Zambia, and S. mansoni may become newly endemic in some areas. Results for 20-year projections are robust to changes in simulated intermediate host snail habitat conditions. The baseline estimates of S. mansoni transmission potential were validated using data collected as part of Task 2.6.
An ensemble of regional climate simulations over Africa generated through work in task 4. 5 (provision of high resolution projections) was used to provide projected daily maximum and minimum temperature data for the eastern Africa study region. The ensemble consists of three members of the Rossby Centre Regional Climate Model (RCM) – RCA4 [22], driven by a coupled atmosphere ocean general circulation model (AOGCM) – EC-EARTH [23]. Three RCP scenarios were used – RCP2.6 RCP4.5 and RCP8.5 – which represent low, moderate and high levels of warming respectively. All three regional simulations were made within the African branch of the Coordinated Regional Downscaling Experiment (CORDEX), and cover the whole African continent at about a 50 km (0.44°) resolution. A smaller sub-domain in eastern Africa, with an area of 1470 (35 x 42) grid boxes, was selected for the study.
Task 3.4 Scenario integration
This task was completed towards the end of the project in the form of the HF Atlas, an online system for analysing and visualising integrations of data and models in order to assess the projected risks of the three VBDs targeted by HF. A summary of activities related to this task, undertaken by UDUR, UNILV, ICTP and ILRI, is provided below for each of the three VBDs.
RVF
How peak-year frequency in RVF incidence changes in comparison to the historical period for both immature and mature livestock was tested. The two regions previously highlighted as hotspots for vector abundance (central/western Kenya and Rwanda) showed significant increases in incidence while most other areas show a significant decrease. These results appear to correlate with suitable mosquito habitat based largely on temperature. As the entire region heats up throughout the century, mosquito abundance is effectively forced into cooler, more highland regions in order to survive. It should be noted that this is a feature of the chosen mosquito-survival scheme implemented in the model, which would benefit from further rigorous analysis due to its evident sensitivity. The decrease in relatively warmer, lower areas was seen in both vectors in the late century. Aedes start to decrease significantly even earlier as large parts of the region also become wetter, but this does not affect Culex proliferation. These trends in vector population dynamics propagate through to the transmission characteristics of the model.
The change in incidence peaks for immature and mature livestock were similar (with significant increases in highland regions) but the increase around Rwanda was less pronounced for mature livestock. Since livestock characteristics were the same in all but case fatality ratio this suggests that livestock immunity has a role to play. Since infected immature livestock either die from the disease (with a probability of approximately 70%) or mature into adults at approximately 2 months, the effects of immunity are negligible. Hence we can think of these immature livestock dynamics as representing an entirely susceptible population.
Information from this mapping exercise formed the basis of the hazard layer for the HF Atlas.
Schistosomiasis
UDUR undertook analysis of different scenarios using the dynamic agent-based model developed in Task 3.3b. Model output was used as the basis for disease risk mapping in the HF Atlas.
Malaria
Different malaria simulations were carried out within HF for the historical period (1980 to 2005).
The impact of future climate change on simulated malaria distribution was tested for LMM-VECTRI based on the super climate ensemble for two emission scenarios (RCP4.5 or RCP4p5 and RCP8.5 or RCP8p5) and for different time slices (2020s, 2050s, 2080s). The results of the super ensemble generally agree with former results obtained during the ISI-MIP and QWeCI EU projects. Those former results were based on a greater malaria model ensemble (including MARA, MIASMA and UMEA) using fewer climate model inputs as drivers (five GCMs were used whereas here we combined 23 different GCMs and RCMs). The climate seems to become increasingly suitable for malaria transmission over the highlands of eastern Africa. The LMM alone tends to simulate similar changes which are, however, smaller in magnitude with respect to the super ensemble and VECTRI.
As with RVF and Schistosomiasis, the HF Atlas provided a means of integrating projected disease hazard with indicators of vulnerability.
WORK PACKAGE 4 - Disease risk and vulnerability mapping
Lead beneficiary: University of Rwanda (UR)
Task 4.1 Identification of current vulnerability hotspots for the 3 target diseases
PLUS, ILRI and UR together adopted a risk and vulnerability framework that viewed humans and their environment as part of coupled socio-ecological systems (SES). Based on a comprehensive literature review and expert consultations, a list of indicators and related data for social vulnerability was agreed for malaria, RVF and schistosomiasis. Building on a methodology developed by PLUS, spatial vulnerability units/regions (geons) were modelled, representing the social vulnerability for the three target diseases (deliverable D4.1).
Task 4.2 Assessment and provision of downscaled climate change projections for the study region from past and on-going projects
An ensemble of statistically downscaled future projections of daily rainfall and temperature for 11 CMIP5 General Circulation Models (GCMs) using the historical and the RCP4.5and RCP8.5 scenarios was provided through this task. The output covered the 140-year period 1960-2099 at 50 km resolution for the eastern Africa region. This ensemble of runs allowed for an exploration of the range of uncertainty introduced by the different GCMs and due to the different emission scenarios. The results were used as input into the different disease models used within HF.
Task 4.3 Provision of land-use change scenarios
To assess the effects of land use change (LUC) on the malaria risk, data from four of the five available Earth System Models (ESMs) that contributed to the LUC experiment of the CMPI5 were used to drive a spatially explicit, dynamical malaria model. Previous efforts to incorporate LUC in dynamical disease models have been extremely limited, and have mostly examined the indirect effect by which LUC impacts temperature and precipitation, which in turn impacts malaria. This study used the suite of ESMs that participated in the CMIP5 process and conducted twin investigations with and without LUC. Results from the task provide the first multi-model assessment of the potential impact of LUC on malaria transmission in Africa via its impact on climate.
To isolate the impact of LUC on climate, the five CMIP5 groups performed two sets of simulations for the period 2006-2100 using the same forcing as for RCP2.6 and RCP8.5 but with the land-use invariant over time. Daily precipitation and temperature output from four of the five ESMs was used to drive the VECTRI model for integrations to 2100 for the model’s historical run (1960-2005) and the two RCPs (2005-2100). A malaria model was run for each available ESM member (three CanESM, two MPI and one member for IPSL and MIROC).
The impact of LUC on climate was assessed in terms of two variables only, the 2 metre temperature and precipitation, as these are key inputs to the malaria model. The first notable aspect of the impact of LUC on precipitation is that, for each ESM, the response is considerably different between the two RCPs. This is in distinct contrast to the changes that occur between the present and future, where the higher greenhouse gas concentrations essentially amplify the magnitude of the spatial changes observed in the lower emissions scenarios (see for example section 2 of HF D3.4). This highlights the fact that the LUC maps are considerably different spatially between the two scenarios. The magnitudes of the changes are relatively small and considerably smaller than the changes observed due to greenhouse gas emissions.
As the impact of LUC on climate is minor in three of the four ESMs, LUC was expected to have a limited indirect impact on malaria in those models. Using the metric of the malaria prevalence (parasite ratio, PR) and length of the transmission season (LTS) this was indeed seen to be the case. In the IPSL, MPI and CanESM models, the change in prevalence was less than 5% everywhere within Africa, and the LTS were mostly less than 10 days. Only the MIROC model produced substantial changes, with prevalence increasing strongly in the southeast of Africa over the northern part of Mozambique and southern Tanzania. Over the Sahel region, the increases in temperature led to decreases in prevalence. This is because above about 30°C mean temperature, malaria transmission starts to decrease with temperature.
In conclusion, this task was the first attempt to isolate and assess the potential indirect impact of LUC on malaria transmission using a dynamical modelling framework, and multiple ESMs based on the latest projection of potential LUC that contributed to the latest round of the IPCC climate assessment. The study showed that in terms of indirect impact, whereby LUC impacts local climate, which can then alter transmission, the effect was very limited in three of the four available models. In one model, the impact was more significant.
The study also highlighted the uncertainty in projecting such extremely model dependent impacts.
Task 4.4 Provision of socio-economic change scenarios
The original intention was to integrate socio-economic indicators and their respective quantities, and future projected values for these, with spatially explicit output from dynamic disease models in order to generate risk maps for the three VBDs. Projected future changes in socio-economic indicators were to be based on IPCC scenarios. Two main activities were carried out in relation to this original intention. A joint scenario workshop was organised by ILRI from 5-7 November 2012. This was organised together with colleagues from CCAFS, where regionally based scenarios were adapted to the specific disease context. These scenarios are qualitative and help to better understand future development pathways in the VBD context. Additionally, how levels for key demographic and economic social vulnerability indicators might be projected for the future and integrated within the HF Atlas was discussed (deliverable D4.4).
Task 4.5 High resolution regional climate projections for eastern Africa study area
A set of high resolution Regional Climate Model (RCM) simulations for eastern Africa was created for use by HF consortium members, in particular those directly involved in projecting future environmental change-driven disease outcomes. Simulations covered the period 1970 to 2100 and generally followed the WCRP CORDEX experiment protocol (http://wcrp-cordex.ipsl.jussieu.fr/). While the majority of Africa CORDEX simulations employed an RCM grid covering the entire African continent at a spatial resolution of 50km, special resolution of the RCM was increased and targeted eastern Africa in order to provide finely resolved spatial detail in simulated present and future climate variables. These simulated data were then input to subsequent assessments of disease risk.
Task 4.6 Risk and vulnerability mapping of disease morbidity and related impacts
The primary aim of Task 4.6 was to produce spatial assessments of current and projected future risks of, and vulnerability to, the three target VBDs in the EAC region. Future projections of risk were based on environmental change data-driven output from dynamic disease models and their subsequent integration with vulnerability indices. Achieving this aim also involved development of the HF Atlas though a highly consultative and reflective process.
Mr Jean Pierre Bizimana was recruited as PhD student for HF in October 2011. The title of his PhD, submitted before project end, was ‘Climate Variability and Malaria in Rwanda: Spatial Assessment of Social Vulnerability at Different Scale Levels’. The specific objectives of the research were: to conceptualise social vulnerability to malaria in the context of environmental and socioeconomic changes; to identify suitable indicators of social vulnerability to malaria; to apply an indicator framework for spatial assessment and explicitly model the social vulnerability to malaria; and to examine the extent to which malaria incidence is the interplay of climate variability and socioeconomic conditions of highland communities in Rwanda.
WORK PACKAGE 5 - Adaptation and support tools: development of decision support tools
Lead beneficiary: Stockholm Environment Institute (SEI)
Task 5.1 Identification and assessment of environmental management and climate change adaptation strategies
An inventory of all strategies relevant to the project – health, environmental, water and sanitation – was created and a report summarising the findings was produced. A total of 50 country documents were posted to the internal project website.
Task 5.2 Identifying key stakeholders for strengthening human and animal health systems
A database of relevant stakeholders and domain experts in the EAC included representatives of numerous sectors (ministries of environment, health, water and irrigation; environmental management agencies; climate change directorates; civil society; and donor agencies). The development of the database involved multiple visits to all five countries comprising the EAC, thus raising awareness of HF. The database was also instrumental in disseminating project outputs and identifying the appropriate mix of participants for the RVF and malaria/schistosomiasis workshops that provided a forum for discussing the architecture and content of the decision support facilities (frameworks and tools) that were important products of WP5 and the project in general.
Task 5.3 Use and assessment of current environmental management and climate change adaptation tools
This task assessed environmental management and climate change adaptation tools that address the environment-related vulnerability of human health. Relevant literature on environmental health impact assessment tools and their application, including those developed specifically for addressing environmental health concerns, was evaluated. Feedback on recent and current use of assessment tools was also sought from domain experts in the EAC. The task informed later stages in the research project when decision support tools (DSTs) and frameworks (DSFs) were in the process of being discussed and developed.
Task 5.4 Refining the RVF Decision Support Tool
Activities carried out as part of this task refined the RVF DSF that had been developed earlier by domain experts involved in attempts to mitigate the impact of RVF in eastern Africa. Three RVF workshops were organised by project members during 2014: 24-25 February (Nairobi, Kenya); 30 September – 1 October (Naivasha Kenya); and 12-13 November (Dar e Salaam, Tanzania). Discussions at the workshop enabled the DSF to be modified in line with up-to-date information, including data generated by HF, and recent past experience, with the amended DSF forming part of deliverable D5.4.
Task 5.5 Developing decision support framework
This task resulted in the production of risk-based DSFs for malaria and schistosomiasis. These two DSFs address changing risk for the two diseases in light of environmental, including climate, change. The DSFs were developed in a workshop held in Nairobi in November 2014 that brought together HF researchers and decision-makers involved in managing the response to these two diseases.
WORK PACKAGE 6 - Stakeholder engagement and empowerment through sharing of knowledge and training
Lead beneficiary: Trinity College Dublin (TCD)
Task 6.1 Identification – and if necessary grouping - of stakeholders and target end users of the HF project
A number of partners contributed to identifying and grouping potential stakeholders and target end users of HF, including TCD, PLUS, NUS and CH. For example, PLUS provided contacts for relevant stakeholders in eastern Africa, within the EC and internationally. The specific focus provided by PLUS was on the identification of stakeholders in the domain of Earth Observation and GIS, including initiatives such as GEO, Copernicus and the WHO. CH helped identify key stakeholders in government and private institutions in Uganda. Information on stakeholders was also harvested from WP5, Task 5.2 ‘Identifying stakeholders for strengthening human and animal health systems’ and other relevant WPs. Project deliverable D5.2 ‘Identification and engagement of key stakeholders’ illustrates the grouping of stakeholders and maps the linkages and flows of information between various groups.
All identified stakeholders were input to a database with fields comprising: type of stakeholder; contact details; and preferred communication language (English or Kiswahili). A protocol was devised by AquaTT for updating and managing the database, and throughout the project the database continued to grow with the addition of the contact details of relevant stakeholders.
The stakeholder database was actively employed when carrying out dissemination activities to ensure those with an interest in HF outputs were kept up-to-date, to increase impact and to facilitate translation of research findings into meaningful actions on the ground. The database was also utilised to invite the participation of relevant stakeholders for the project’s various stakeholder meetings and engagement workshops.
Task 6.2 General dissemination
Over the course of the project, HF has employed a range of methods to disseminate project results and outputs to stakeholders and end users. Dissemination of information to project members was facilitated through the use of Basecamp, the project intranet site. A publically accessible website, available in both English and Kiswahili, was developed and maintained throughout the project providing regular updates on project news and outputs. The website has proven to be highly popular having received ~40,000 hits since its launch midway through the first year of the project. AquaTT will continue to maintain this website for up to five years after the completion of the project to ensure the continued and uninterrupted dissemination of HF outputs.
In addition to the website, a number of other dissemination activities has taken place including the production and distribution of:
- Seven contractual HF news releases (available in both English and Kiswahili) and additional news releases disseminated through such channels as AlphaGalileo (http://www.alphagalileo.org) the EC’s CORDIS News services (http://cordis.europa.eu/news/) CORDIS WIRE, Twitter, LinkedIn and AquaTT’s Training News.
- Seven HF newsletters (non-contractual)
A non-technical, tri-fold project brochure was developed for general release outlining the project and English and Kiswahili versions are available to download from the project website.
A short promotional video (c.90 seconds) that summarises HF was also developed. The video, hosted by VIMEO and aimed at raising awareness of the project, can be viewed at www.vimeo.com/70318624 or on the project website.
Two promotional items that include the HF logo and website address were designed to increase the brand awareness of the project: a key-ring with bottle opener and torch, and post-it notes. These have been distributed at a number of stakeholder engagement workshops and events over the course of the project. These items will continue to be disseminated after the completion of the project, for example, at the HF session at the ‘Impact of Environmental Changes on Infectious Disease’ Conference which will be held from the 23 – 25 March 2014 in Sitges, Spain.
HF also contributed to the completion of a short documentary produced by the charity Africa Turns Green. The documentary “Health and climate change in Africa” was released in August 2014 and is available to view on the project website:
http://www.healthyfutures.eu/index.php?option=com_k2&view=item&layout=item&id=176&Itemid=276&lang=en
and on YouTube:
https://www.youtube.com/watch?v=oYL4Nc-qnKE
The video focuses on the collaborations between HF and its sister project, QWeCI, and outlines how both projects have worked to increase capacity in Africa and reduce the impacts of the target VBDs in the study region. A number of representatives from HF partners (UDUR, UNILIV and ILRI) were featured in the documentary, while Dr Laragh Larsen and Professor David Taylor were consulted over the content of the script. This documentary has been used during stakeholder engagement events and will be employed in future to inform the general public about HF and about project outcomes.
HF has been included as a case study in the World Health Organization/the World Meteorological Organization (WHO/WMO) special publication on Climate Services for Health. The submission focuses on the HF Atlas, and is part of a small group of accepted pieces from more than 75 submissions globally. The final expected outputs of the Climate Services for Health project comprise an online e-book, interactive web repository and a hardcopy publication in the six UN languages.
Task 6.3 Targeted dissemination
Four project updates have been distributed to all stakeholders listed in the stakeholder database. These project updates summarise the work that was carried out each year, highlighting noteworthy achievements. These short briefs provided readers with the means to obtain further information if desired.
Task 6.4 Communication and feedback
A ‘FAQ & Feedback’ section on the project website:
http://www.healthyfutures.eu/index.php?option=com_k2&view=item&layout=item&id=133&Itemid=265&lang=en
provided a channel for communication and feedback. Furthermore, all stakeholder engagement workshops carried out during the project facilitated an interactive forum for two-way communication and feedback.
Task 6.5 Compilation of all knowledge generated from the project and translation (appropriate level, terminology, language) for end users
All content on the HF website is available in both English and Kiswahili in order to be accessible to a wide range of target stakeholders. The same applies to all contractual news releases, and the project factsheet and brochure.
Within several work packages and tasks, user manuals and guidance documents on technical issues were produced, particularly those provided by PLUS on the metadata portal and the HF Atlas.
Task 6.6 Scientific publications of research and presentations at scientific conferences
A full list of peer reviewed publications can be found in deliverable D6.15. HF research, results and outputs have been presented in a variety of formats and capacities since the project’s start.
Task 6.7 Engagement and training workshops (for refinement of Decision Support tools developed in WP5)
A number of engagement and training workshops for the refinement of DSTs developed in WP5 have been held.
1. One-day stakeholders meeting on malaria and climate change, Kampala, Uganda, 25 October 2013
CH (part of the Ministry of Health (MoH) Uganda) in collaboration with ICTP, NUS and SEI organised a one-day meeting themed ‘Malaria seasonal forecasting and Climate Change adaptation’. The meeting, held in Kampala, Uganda, was attended by a variety of stakeholders, including those from the health sector, academia, researchers and malaria-related institutions. Key attendees included the Director General of Uganda National Health Research Organisation (UNHRO) Dr Sam Okware and Dr Myers Lugemwa of the Uganda National Malaria Control Programme. The meeting reviewed available malaria and climate data, current malaria forecasting by ICTP and the European Centre for Medium Range Weather Forecasts (ECMWF) and possible DSTs and approaches. One of the main aims of the meeting was the sharing of current knowledge on malaria and climate modelling using available data from Uganda. In his opening remarks to the meeting Dr Sam Okware observed that climate change affects health and the whole health system, noting that most diseases in the country are climate-sensitive, especially vector-borne and water-related diseases such as malaria and diarrhoea. Subsequent presentations at the meeting covered broad areas of data collection and quality, especially related to malaria data, malaria early warning systems and forecasting, the climate change national adaptation programme of action in the health sector in Uganda and DST. The workshop highlighted the need for the different stakeholders to use the outputs of the research to improve preparedness, quantify disease impacts, and above all to enable the decision-making/disease outbreak forecasting and response processes. Recommendations results from the meeting included: the development of a hands-on module by ICTP and supported by MoH, Uganda, for training in the use of the malaria forecasting system output in the operational planning environment; the evaluation of the potential for directly using the VECTRI modelling system for intervention planning and also investigating potential adaptation strategies with regard to sensitivity to climate change; the identification and exploration of key areas of collaboration in vulnerability of health to climate change with a view to improving upon current adaptation strategies; undertaking further research on how to extend climate and disease modelling and forecasting to other diseases. A full report on the meeting including the presentations given is available to download from the project website.
2. Stakeholder Engagement Workshop, Nairobi, Kenya, 24-25 February 2014
A second HF stakeholder workshop was held from 24-25 February 2014 at the International Livestock Research Institute (ILRI), Nairobi, Kenya. The aims of the workshop were to raise the profile of HF research among selected stakeholders in eastern Africa and to further the process of enabling the development and effective uptake of outputs from the project. Specifically, the workshop had the following objectives: to communicate the project research outputs (modelling, risk mapping etc.) that were currently in progress for each of the three target VBDs; to collect feedback on the value of the research outputs to date and on how to support their uptake by decision makers, for example, integration of modelling results into DSTs; to develop case studies for the application of DSTs and DSFs; to test different decision-making methods through the use of case study examples to compare the value of various intervention strategies; and, to define the outputs and needs for further stakeholder engagement in HF and beyond.
Participants in the workshop included a number of HF partners, external members of the HF Expert Review Panel (ERP) and representatives from government ministries and research institutes in eastern Africa. During the first day of the workshop, plenary presentations were made by project partners to introduce the project and provide updates on case studies of the three diseases. The presentation included: an overview of HF; malaria projections for 2050 and how climate information can be integrated into health planning; sources of climate data for national decision-making; guidance that models can provide for schistosomiasis control; and a DST for RVF. On the second day of the workshop, participants carried on working in three groups, each focused on one of the diseases using the case studies presented during day one. The groups tested two of the methods or ‘engines’ of the adaptation decision explorer (ADx) tool: the voting method and Analytic Hierarchy Process. ADx was being trialled at the workshop as a possible component of a DST.
This stakeholder engagement workshop was a positive step forward in communicating the latest research from the project to stakeholders, in receiving feedback about what specific concerns decision makers have, and in sharing insights into how those concerns could be best addressed. However, weaknesses in the ADx DST became apparent, as did a divide between time frame of main interest to HF (up to century scale) and the much shorter policy-oriented time scales that most concerned many decision-makers present at the workshop. One specific request to emerge from the workshop was for concise, user-friendly summaries of the likely risks of the three target diseases under changing conditions predicted for coming decades. These summaries were subsequently produced by members of the HF consortium and are available on the HF website. Overall, this stakeholder engagement workshop proved to be a valuable exercise in identifying how the outputs of HF could be designed to meet the needs of decision makers in the EAC. More information on this workshop can be found on the project website.
3. Decision-Makers Workshop for Malaria and Schistosomiasis, Nairobi, Kenya, 18-20 November 2014
The third stakeholder engagement workshop for the refinement of DSTs was held in Nairobi from 18-20 November 2014. The objective of the meeting was to develop two separate risk-based DSFs for schistosomiasis and malaria that could assist decision-makers in the EAC by serving as guides to responses to these two diseases in the face of climate change. HF outputs include spatially explicit assessments of risks of the three target VBDs in eastern Africa as a result of projected changes in environmental conditions. These assessments are based on combined hazard and vulnerability data. This HF workshop reviewed the schistosomiasis and malaria research outputs of the project in the context of the general state of knowledge for these diseases and assessed the implications of the findings for future decision making and action. Risk-based DSFs are decision-making guides developed by decision-makers. This approach was pioneered in applications to RVF epidemic response; DSFs are developed through participatory consultations that bring together health decision-makers, control programme implementers and researchers. They systematically break down the decision-making process into sets of steps by defining decision points and action categories where decisions need to be taken. They are not prescriptive. Instead DSFs act as aide memoires to remind decision-makers of areas where action should be considered at appropriate points in time. DSFs are specifically intended to reduce the impact of uncertainty and risk in decision-making. DSFs are living documents that are part of an adaptive management process.
This workshop began in plenary session with a brief technical overview leading to a discussion to distil key action points coming from the research that should be captured in the DSFs. Next the meeting discussed the objectives of the two frameworks and the time frame that they should cover. Thereafter the meeting broke into two parallel sessions: one on schistosomiasis and the other on malaria. These parallel sessions built the timelines describing the process of epidemiological change, identified the decision points and action categories and articulated the DSFs in detail. Many of the key action points arising from the technical discussion were captured in the statements of objectives and assumptions that guided the detailed construction of the DSFs.
In the case of malaria, major international control programmes are in place and the DSF is intended to be complementary to existing policy and strategy documents. The DSF accommodates the future effects on malaria epidemiology of three inter-related phenomena: climate, environment and land use change, and population migration and displacement.
In the case of schistosomiasis, international control efforts are lacking and the institutional context of schistosomiasis control is not well developed. The schistosomiasis DSF is intended to complement to regional and national policies and strategies, and to catalyse appropriate institutional development.
Task 6.8 Africa-EU Workshop
The original intention was for the University of Rwanda to host a workshop, focusing on environmental change and VBDs, in the final year of the project to coincide with the release of the IPCC’s 5th Assessment reports, particularly the report from Working Group II dealing with climate change impacts, including projected health effects. Unfortunately organisational delays, including delays releasing information about the meeting, key note speakers etc, led to a low level of international interest in the meeting and forced a decision, taken at the last (the 6th) partners’ meeting in September 2014 to cancel the workshop.
Permission was sought and fortunately obtained from the project officer in the EC to organise and host a HF-focused session at the ‘Impact of Environmental Changes on Infectious Disease’ (IECID) Conference, Sitges, Spain, 23-25 March 2015. The session attracted a large, engaged audience and was deemed a major success. Papers delivered in the session are forming the core of a special issue of the journal Geospatial Health, which is now in the process of being produced. In addition to the HF-focused session at the IECID Conference in Sitges, Spain, HF project members organised a session (CL2.5 Climate and infectious disease interactions) at the European Geophysical Union (EGU) annual meeting in Vienna, Austria, in April 2012 (i.e. during the second year of the project). During the following year, HF project members also jointly coorganised and hosted a symposium, along with its sister project QWeCI, as part of the 4th Annual East African Health and Scientific Conference that took place in Kigali, Rwanda from 27 to 29 March 2013.
HF Symposium, 4th EAC Health And Scientific Annual Conference, Kigali, Rwanda, 27 to 29 March 2014
The conference programmed comprised four sub-themes (Maternal and Child Health, Non Communicable Diseases and Trauma, Health Systems Strengthening, and Quality of Health Care) and four symposia: HIV and AIDS; Integrated disease surveillance and disaster preparedness; Tobacco Control; and the HF/QWeCI symposium on ‘Environment and Health in Africa: climate and vector-borne diseases’. Facilitated by NUR, the symposium was well-attended and brought together researchers at the cutting edge of efforts to understand the relationships between health and environment, and in particular the links between climate and VBDs in Africa. There were a total of 17 presentations in this session, including the plenary by Dr. Margaran Bagayoko (Protection of Human Environment Programme, World Health Organization Regional Office for Africa) whose talk centred on the potential of climate-based early warning systems for improved management of VBDs in Africa. The presentations that followed were wide ranging and delivered by speakers who had travelled from other parts of Africa and from Europe and Asia (the full programme is published in the third issue of the HF newsletter available on the HF website).
WORK PACKAGE 7 - Enhancing research synergy & application
Lead beneficiary: AquaTT
Task 7.1 Identify relevant projects and conferences
Over the course of the project, a list of relevant projects, events and conferences has been uploaded to the project intranet site and maintained. This list has been managed by AquaTT, which has worked to identify and add relevant information on an ongoing basis. Other consortium members have also contributed to the list by informing AquaTT of any relevant events and projects they have identified.
In addition to the Basecamp file, AquaTT has added information on all events and conferences to the HF website on both the homepage, under ‘Upcoming Events’, and to the dedicated ‘Events’ page, which displays all events on a Google calendar and provides the ability to search for events and conferences by date. There is also a dedicated ‘Search’ page that allows users to search using particular keywords.
The project website also contains a ‘Useful Links’ page. This page contains a list of relevant projects identified and links to those project websites as well as links to further Information. Currently there are eighteen links to relevant research projects, four links to relevant tools and four links to further Information.
Task 7.2 Establish an Expert Review Panel
The ERP was established at the beginning of the project in order to provide guidance and direction to HF project members. The ERP advised and evaluated HF activities using a robust system of internal and external controls. Focusing on providing advice on project direction, the ERP did not have any authority to vote on project matters, nor had it a legal responsibility in the operation of the project or the partner organisations.
The ERP initially consisted of eight panel members comprising of five experts in the research field who were not members of HF and three members of HF, one of whom was based in Africa and two of whom were originally based in Europe (one of the two subsequently moved to Asia, but continued to attend ERP meetings). The composition of the ERP was designed to ensure some independent scrutiny of the operation and aims of the project, while also facilitating communication between external members of the ERP and the wider HF consortium.
The members of the ERP were:
- John B. Malone, Professor, School of Veterinary Medicine, Baton Rouge, LA, USA (Chair of Expert Review Panel)
- Madeleine Thomson, Senior Research Scientist (Climate Information for Public Health), Africa Program, International Research Institute for Climate and Society, Columbia Earth Institute, USA
- Simon Brooker, Reader in Tropical Epidemiology in the Department of Infectious and Tropical Diseases at the London School of Hygiene and Tropical Medicine & Welcome Trust Research Fellow, KEMRI, Kenya
- Timothy Wesonga, East African Community Secretariat (EAC) Senior Livestock and Fisheries Officer, Tanzania
- Dr. Maurice Owuor Ope, East African Community Secretariat (EAC) Disease Surveillance and Epidemiology Officer, Tanzania
- Jan Semenza, European Centre for Disease Prevention and Control (ECDC), Sweden
- Paul Lowen/Ciara Egan (AquaTT), Ireland
- David Taylor (National University of Singapore), Singapore (originally Trinity College Dublin, Ireland)
- Theophile Niyonzima (NUR), Rwanda
Task 7.3 Formal peer-review and synergy events
Over the course of the project four ERP meetings were held. Each meeting was organised to coincide with Partner meetings and other project workshops in order to provide ERP members the opportunity to meet with members of the HF consortium to discuss the project.
The first meeting took place on 13 October 2011 in Kampala, Uganda. This meeting established a relationship between the ERP members and HEALTHY FUTURERS partners. The ERP members reviewed and discussed the work done to date within each work package and provided a number of recommendations to be incorporated by partners.
The second meeting took place on 10 May 2012 in Arusha, Tanzania. The ERP was bolstered by the presence of Dr. Maurice Ope, an expert in disease surveillance and epidemiology in eastern Africa, representing the Secretariat of the EAC. Maurice made a highly valued contribution to the ERP and was able to brief other members of the ERP on the fourth Annual East African Health and Scientific Conference that was scheduled to take place in Kigali in March 2012. As a result of this synergy, HF organised and ran a session at the conference on the theme of ‘Environment and health’.
The third ERP meeting was held on the 1 May 2013 in Trieste, Italy. Again the ERP met and reviewed the progress made in each work package including how the previous recommendations were incorporated by partners.
The fourth and final ERP meeting took place on the 28 February 2014 in Nairobi, Kenya. The ERP provided invaluable advice during this meeting on the strengths and weaknesses of the research approaches taken within the project, particularly in relation to the development of a DST.
Task 7.4 Report on meeting outcomes
Reports from the each of the ERP meetings are available to download from the project website: http://www.healthyfutures.eu/index.php?option=com_k2&view=item&layout=item&id=140&Itemid=267&lang=en
and have been disseminated to key stakeholders.
A report on the collaborations and synergies developed over the course of the project was also produced, and provides a comprehensive overview of the main collaborations established over the course of the project. The report has been disseminated to key stakeholders and is available to download from the HF website: http://www.healthyfutures.eu/images/HEALTHY_FUTURES_Synergies_and_Collaborations.pdf
Potential Impact:
HF with its integrated holistic approach has considerably enhanced current understanding and dynamic modelling of links between the three target, water related VBDs (malaria, schistosomiasis and RVF) and environmental change drivers and effects. The project has also raised awareness of potential future environmental change impacts on health among decision makers in the countries comprising the East African Community (EAC), and among the secretariat for the EAC. The potential impact of HF is described in relation to the expected impacts identified in the project proposal:
(1) More accurate and reliable predictions for the distribution of three water related vector-borne diseases (VBDs) in Africa
HF collected a wide range of existing and new data and collated these data within a project database (‘Geoportal’) (WP 2). These data comprise historical, socio-economic, migration, settlement and conflict data; earth sciences data; direct climate observations and the results of regional downscaling of global reanalyses; and information on disease-environment relationships. The Geoportal acts as a main repository for data and information on the current and future impact of environmental and socio-economic change on VBDs in eastern Africa, specifically the countries comprising the East Africa Community (EAC), and is an important resource for disease planning, further research and teaching and training in the study area and farther afield. The Geoportal provides a standardised format that enables communication between databases. The link, accessible to all interested parties, is: http://41.204.190.50/geoportal/catalog/main/home.page
The HF Atlas, intended as a resource to support teaching, research and health planning, utilises the data collected and collated through HF to highlight hotspots of projected disease risk in the region. This interactive, web-based mapping and DST, aims to provide meaningful and guided access to information on environmental change, and disease hazard, vulnerability and risk relating to the three target VBDs in eastern Africa. Changes in projected disease hazard and risk can be visualised for different time periods up to the end of the current century and for different geographical scales. The Atlas also compiles searchable information on historical disease outbreaks in the study area. An important aspect of the Atlas is use of climate models and future scenarios referred to in the latest (the 5th) IPCC Assessment Report.
HF did not carry out any disease modelling for the European region owing to budget constraints. However, environmental change impacts have a transboundary nature, and the distributions of vectors and related diseases are variable. In a highly globalised and dynamic world, VBD surveillance and early warning systems targeted at Africa have therefore relevance for Europe. Moreover, research carried out through HF is, with further financing, extendable to neighbouring parts of continental Africa, and to continental Europe to the north.
(2) Strengthening of the early warning, surveillance and monitoring systems for vector-borne diseases
A key outcome of HF has been release of the DSF designed specifically to strengthen existing early warning, surveillance and monitoring systems in eastern Africa for the three target VBDs and developed in consultation with decision makers in the East African Community (EAC).
(3) Support to policies on climate change and health
HF identified and deployed the most appropriate environmental change adaptation strategies and decision support approaches (including relevant frameworks and information platforms) for adoption in the project. One of these tools is the RVF DSF, which has been established and field-tested in the study area. The RVF DSF has been designed to guide timely, evidenced based decision-making in the control of RVF, and decomposes the RVF epidemic cycle into explicit steps that are then matched against specified actions (including interventions). Many of the steps are explicitly time dependent. The integration of inputs from multiple partners, decision-makers and experts (through a series of stakeholder engagement workshops hosted in the study region) was designed to ensure ownership of the product, and relevance to the decision-making challenges that have been experienced during previous RVF outbreaks.
Refinement of the RVF DSF was based on new outputs from research, particularly those evaluating the impacts of climate change on RVF epidemiology and the incorporation of human health interventions into the framework in line with One Health principles. The RVF DSF makes clear links to policy. For example, during an inter-epidemic period the RVF intervention for animal health disease prevention is to develop a clear policy on vaccination against RVF, including during inter-epidemic periods, when risk of RVF outbreak is high and in the face of an outbreak; during a pre-outbreak phase the RVF intervention for human health and animal health is to conduct a rapid risk assessment incorporating the level of vaccination in the area and informing decision and policy-makers of results.
DSFs have also been developed for malaria and schistosomiasis, also in consultation with domain experts in the EAC. Guidance notes covering projected changes in disease rises for the three target VBDs over coming decades, driven by environment, including climate, change, have also been made publically available.
In recognition of the project’s policy relevance in the field of environment and health, HF has been included as a case study in the World Health Organization/the World Meteorological Organization (WHO/WMO) special publication on Climate Services for Health. The case study focuses on the HF Atlas, and is part of a small group of accepted pieces from more than 75 submissions globally. The final expected outputs of the Climate Services for Health project comprise an online e-book, interactive web repository and a hardcopy publication in the six UN languages, produced by the WHO/WMO.
(4) Promotion of sustainable management of the natural and human environment and its resources by advancing our knowledge on the interactions between the biosphere, ecosystems and human activities
Water-related VBDs are transboundary in their distribution and effects. An integrated (i.e. regional and global), multi-national response is therefore required. The transmission and outbreaks of VBDs are also not only related to health and to environment. Their occurrence and severity are also products of socio- economic conditions, and the abilities of health and veterinary services to respond to early warnings of possible outbreaks. Joint action, across international boundaries, therefore represents the only effective way to ensure secure livelihoods and the health security of citizens. HF was guided by the concept that the stability of livelihoods, i.e. their strong resilience, crucially depends on maintaining environmental integrity. In this context, the consortium partners saw a strong link between animal and human health and environmental sustainability.
(5) Developing new technologies, tools and services, in order to address in an integrated way global environmental issues [...tools and on technologies for monitoring, prevention and mitigation of environmental pressures and risks...]
The HF project has led to the development and application of the following new technologies, tools and services:
- A publically accessible Geoportal comprising information drawn from a range of primary and secondary sources, including: historical, socio-economic, migration, settlement and conflict data; earth sciences data; direct climate observations and the results of regional downscaling of global reanalyses; and information on disease-environment relationships. The Geoportal is at: http://41.204.190.50/geoportal/catalog/main/home.page
- Improved understanding of environment-disease dependencies in eastern Africa, and of the factors that act to confound any relationships
- A new generation of dynamic disease models for the three target, water related VBDs
- The HF Atlas (http://zgis186.geo.sbg.ac.at/hf_atlas/) an interactive, web-based mapping and visualisation tool
- DSFs for the three target VBDs developed in consultation with relevant decision makers in the EAC region
- Environmental change and health in eastern Africa is now regarded as an important case study in discussions concerning the nature and effectiveness of climate services provision for health by the World Health Organization/the World Meteorological Organization (WHO/WMO). The case study focuses on the HF Atlas, and is part of a small group of accepted pieces from more than 75 submissions globally that will form the basis of an online e-book, interactive web repository and a hardcopy publication in the six UN languages produced by the WHO/WMO.
(6) Addressing 'Environmental safety and welfare' as an overarching theme
By contributing to enhanced understanding of links between environmental, including climate, change and the emergence and spread of the three target VBDs, by improving related outbreak early warning systems and by having a relevance beyond the geographic boundaries of the study area – including raising awareness of the potential health and economic impacts of climate change in the developing world - HF addressed environmental safety and welfare as an overarching theme.
Moreover, by communicating the potential links between environment, including climate, change and animal and human health, HF added further weight to arguments in favour of the need to mitigate environmental/climate change impacts, through the reduction of greenhouse gas emissions and increased environmental protection and wise use of resources, and through supporting adaptation measures among the most vulnerable human populations.
(7) Assessing, reducing and preventing tensions and conflicts related to the depletion of natural resources and environmental services [...of all means including technologies, earth observation, modelling and socio- economic research approaches...]
The HF project utilised Earth observation and environmental change technologies, in combination with existing data, to assess, monitor and model dynamic environmental conditions – and their potential implications for VBD burdens – in eastern Africa. The multi-disciplinary and consultative approach, adopted in the results generated, are fully compatible with ensuring sustainable development and the wise exploitation of natural resources in the study region.
Moreover, HF obtained input from decision-makers associated with the VBDs in the EAC region through a series of stakeholder engagement workshops. This series was an important foundation for the development of DSFs (WP 5). By working with the EAC, which is mandated to improve the wellbeing of citizens in eastern Africa, HF contributed to easing tensions within the region, and to the maintenance of good relationships within and between countries comprising the EAC.
(8) Consideration given to the various geographical, sectoral and cultural differences which exist within Africa, and broader socio-economic factors
The HF study area incorporated the eastern African countries of Burundi, Kenya, Rwanda, Tanzania and Uganda. The boundaries of the study area therefore equate to those of the EAC. The EAC is the regional intergovernmental organisation of the republics of Burundi, Kenya, Rwanda and Uganda and the United Republic of Tanzania. Regional cooperation towards a completely healthy society within member states is a key, stated aim of the EAC.
The choice of eastern Africa as a study area acknowledges the necessity of willingness at all levels of society – including supra-national organisations such as the EAC - to implement the findings and increased understanding that accrue from scientific research in order to mitigate effectively the negative health and economic effects of environmental, including climate, change. The cooperation of supra-national organisations is required because of the transboundary dimensions of environmental change causes and effects, the latter including health impacts. The study area also serves as a useful model in which to examine interlinkages between environment and water-related VBDs in Africa. A range of altitude- and latitude-related environmental conditions, including extensive highland areas and a lowland coastal plain, and a range of humidity, from more or less permanently humid highlands to arid and semi-arid plains, are accommodated within the study area. Moreover, large freshwater bodies and extensive wetlands, some of which are only now being exploited for food production, are also present.
Eastern Africa can also be viewed as a model of the African continent as a whole, in terms of human life. Environmental diversity in the study area is matched by a rich variety of human populations, languages, cultures and religions, and unevenness in the distribution of settlements, economic activities and investment in health. Highlands in the study area support disproportionately high densities of human populations. Many of the people in the study area are living in rural areas (rural population densities in some parts of the study area are among the highest in the world) as subsistence farmers. As such they are often distant from health and veterinary services. Parts of the study area have also experienced political instability in the relatively recent past, one consequence of which is the presence of temporary camps for displaced people, while urbanisation has also become a major factor. These are all factors that will influence the health effects of environmental change.
Within the project itself, there has been a strong representation of Africa-based partners, including in positions of responsibility, and stakeholders in the research. Several Work Packages were jointly-led by representatives from Africa-based institutional members of the HF consortium, while the Expert Review Panel was chaired by the representative of NUR (now University of Rwanda) in the project. Moreover, three of the six partners’ meetings were held in eastern Africa – including in Arusha, Tanzania, in order to facilitate attendance by a representative of the EAC, as was one of the two research symposia and one of the two Early Stage Researcher Generic Skills and Networking workshops. All three meetings with stakeholders were held in eastern Africa. Two of the four PhD scholarships funded through the project when to Africa-based students. Finally news releases and summaries of HF research and results were made available in both English and Kiswahili, while project updates were frequently sent to stakeholders in the research. More than 30% of the c. 40,000 hits on the project website (www.healthyfutures.eu) have originated in Africa – with Kenya (over 6000 hits) and Rwanda (over 4500 hits) showing the greatest online interest in the project, according to the number of hits.
(9) Training activities and exchange of staff
HF had a strong commitment to training activities. There were two Early Stage Researcher (ESR) Generic Skills and Networking workshops held:
- The first took place in Arusha, Tanzania in May 2012. Eight ESRs funded through HF participated in the workshop, which was also attended by seven research trainers.
- The second took place at the ICTP, Trieste, Italy in April 2013, and included attendance at the Spring School on Modelling Tools and Capacity Building in Climate and Public Health in the ICTP. The content of this school was of particular relevance to the work being carried out in the HF project. Six ESRs from five of the HF partner institutions participated in the two-week school, which was attended by more than fifty people in total (the vast majority from developing world countries). In addition, several other members of the HF team contributed to teaching at the school.
Four PhDs, funded through HF, has led to the publication of several papers in high impact, internationally refereed journals, in addition to theses. The titles of the PhD thesis research and the students concerned are:
a. ‘Modelling the effects of temperature changes on Schistosoma mansoni transmission’ completed by Nicky McCreesh at Durham University.
b. ‘Integrated spatial indicators for modeling, exploring and visualising vulnerability to vector-borne diseases’ completed by Michael Hagenlocher at PLUS.
c. ‘A simulation model of Rift Valley fever transmission in Kenya’ completed by John Gachohi at the ILRI.
d. ‘Climate change and malaria in Rwanda: Spatial assessment of social vulnerability at different scale levels’ completed by Jean Pierre Bizimana at the University of Rwanda (UR) - The first ever PhD to be awarded in Science by the UR.
(10) Innovative management and governance tools and adaptive technologies suitable for the relevant authorities and stakeholders in Africa
As mentioned above, HF developed and implemented innovative management and governance tools through its online Geoportal and DSTs. These were designed and implemented together with key local partners – essential players in sustaining the technologies after the project end - but also through the integration of regional and local stakeholders through stakeholder engagement workshops. Furthermore, environmental change and health in eastern Africa was chosen as an important case study for Climate services for health by the World Health Organization/the World Meteorological Organization (WHO/WMO). The case study focuses on the HF Atlas, and is part of a small group of accepted pieces from more than 75 submissions globally. The final expected outputs of the Climate Services for Health project comprise an online e-book, interactive web repository and a hardcopy publication in the six UN languages.
(11) Integration of local stakeholders, and/or regional actors, and the necessary networking; clustering and coordination activities between the relevant selected projects
HF partners consulted actively with local and regional stakeholders throughout the project. Besides strategic involvement of key local stakeholders in a number of engagement and training workshops for the refinement of DSTs, partners also networked with regional actors. For example, partners from TCD travelled to Burundi to investigate the country’s archival collection and to establish better links with the country, meeting representatives from Ministère de la Santé Publique et de la Lutte contre le Sida (Minisanté) (Ministry of Public Health and Fight against AIDS) and the Programme National Intégré de Lutte contre les Maladies Tropicales Négligées et la Cécité (PNIMTNC) (National Integrated Program for Neglected Tropical Diseases and Blindness Control) amongst others. Burundi is the only country member of the EAC that had no institutional representatives in the HF project.
Furthermore, HF with its sister project QWECI (which focused on western Africa) jointly coordinated a symposium at the 4th annual East African Community Health & Scientific Conference, Kigali, Rwanda in 2013. Both projects also appeared together in the video documentary entitled Health and Climate Change in Africa, produced by Africa Turns Green, a charity that showcases the work of African green entrepreneurs who are protecting their environment.
As previously mentioned, within the project itself, there has been a strong representation of Africa-based partners, including in positions of responsibility, and stakeholders in the research. Several Work Packages were jointly-led by representatives from Africa-based institutional members of the HF consortium, while the Expert Review Panel was chaired by the representative of NUR (now University of Rwanda) in the project. Moreover, three of the six partners’ meetings were held in eastern Africa – including in Arusha, Tanzania, in order to facilitate attendance by a representative of the EAC, as was one of the two research symposia and one of the two Early Stage Researcher training and networking workshops. All three meetings with stakeholders were held in eastern Africa. Two of the four PhD scholarships funded through the project when to Africa-based students. Finally news releases and summaries of HF research and results were made available in both English and Kiswahili, while project updates were frequently sent to stakeholders in the research. More than 30% of the c. 40,000 hits on the project website (www.healthyfutures.eu) have originated in Africa – with Kenya (over 6000 hits) and Rwanda (over 4500 hits) showing the greatest online interest in the project, according to the number of hits.
Main Dissemination Activities and the Exploitation of Results
HF results were exploited through the prompt publication of research findings in a diversity of fora, including high impact scientific journals, webpages, frequent news releases, workshops and international conferences. HF deliverables were disseminated in English and, where relevant, material was translated into Kiswahili, which is also widely spoken in the study area. The variety of platforms and forms of media used – and the hosting of deliverables on an open access, project website - maximised the impact of the results of the project. Main dissemination activities included:
- HF website - has to date received over 40,000 hits and is available in both English and Kiswahili (http://www.healthyfutures.eu/)
- Seven contractual HF news releases - available in both English and Kiswahili and additional news releases disseminated through such channels as AlphaGalileo (http://www.alphagalileo.org) the EC’s CORDIS News services (http://cordis.europa.eu/news/) CORDIS WIRE, Twitter, LinkedIn and AquaTT’s Training News (also available on project website)
- Project factsheet - available in both English and Kiswahili (also available on project website)
- Seven HF newsletters - non-contractual (also available on project website)
- Four project updates - distributed to all stakeholders listed in the Stakeholder Database (also available on project website)
- A non-technical, tri-fold project brochure - available in both English and Kiswahili developed for general release outlining the project (also available on project website)
- A short promotional video - (c.90 seconds) summarising the HF project and hosted by VIMEO aimed at raising awareness of the project, viewable at www.vimeo.com/70318624 or on the project website
- Two promotional items: a key-ring with bottle opener and torch, and post-it notes - which include the HF logo and website address designed to increase the brand awareness of the project. These have been distributed at a number of stakeholder engagement workshops and events over the course of the project.
- A short documentary “Health and climate change in Africa” - produced by the charity, Africa Turns Green and HF contributed to its completion. The documentary was released in August 2014 and is available to view on the project website (http://www.healthyfutures.eu/index.php?option=com_k2&view=item&layout=item&id=176&Itemid=276&lang=en) and on YouTube (https://www.youtube.com/watch?v=oYL4Nc-qnKE). The video focuses on the collaborations between HF and its sister project, QWeCI, and outlines how both projects have worked to increase capacity in Africa and reduce the impacts of the target VBDs in the study region.
- HF Symposium - Jointly coordinated by members of HF and its sister project QWeCI and held at the 4th Annual East African Health and Scientific Conference in Kigali, Rwanda in March 2013. The conference programmed comprised four sub-themes (Maternal and Child Health, Non Communicable Diseases and Trauma, Health Systems Strengthening, and Quality of Health Care) and four symposia: HIV and AIDS; Integrated disease surveillance and disaster preparedness; Tobacco Control; and the HF/QWeCI symposium on ‘Environment and Health in Africa: climate and vector-borne diseases’. The symposium was well-attended and brought together researchers at the cutting edge of efforts to understand the relationships between health and environment, and in particular the links between climate and VBDs in Africa. There were a total of 17 presentations in this session, including the plenary by Dr. Margaran Bagayoko (Protection of Human Environment Programme, World Health Organization Regional Office for Africa) whose talk centred on the potential of climate-based early warning systems for improved management of VBDs in Africa. The presentations that followed ranged in topics related to the symposium title, from speakers who had travelled from other parts of Africa and from Europe and Asia.
- HF-focused and organised/hosted session - at the ‘Impact of Environmental Changes on Infectious Disease’ (IECID) Conference, Sitges, Spain in March 2015
- Scientific Papers - 19 papers published in scientific journals with one more awaiting decision. Two submissions in a WHO-WMO special publication on climate services and health. One special issue of the journal Geospatial Health is currently in preparation.
- EC Promotion - HF has featured prominently in two widely circulated magazines produced by the EC: Horizon (“Predicting disease outbreaks in East Africa”) and Research EU (“Anticipating climate change, tackling disease”) and most recently through the EC Horizon 2020 website (http://ec.europa.eu/programmes/horizon2020/en/news/mapping-effects-climate-change-deadly-diseases).
- Oral presentations of HF results and networking - partners have carried out dissemination activities through attendance (and in a lot of cases presentation of HF results) at many key conferences and meetings, details of which have been provided in the participant portal.
- Case study in a special publication on Climate and Health (WHO/WMO) – HF has been included as a case study in the World Health Organization/the World Meteorological Organization (WHO/WMO) special publication on Climate Services for Health. The case study focuses on the HF Atlas, and is part of a small group of accepted pieces from more than 75 submissions globally. The final expected outputs of the Climate Services for Health project comprise an online e-book, interactive web repository and a hardcopy publication in the six UN languages, produced by the WHO/WMO.
List of Websites:
The address of the public website is: www.healthyfutures.eu
The following is a list of the email contact details and affiliations of HF project partners:
[Partner number (Beneficiary Short name, Country) - Contact person]
P1 (TCD, Ireland) - Laragh Larsen llarsen@tcd.ie/Gayle McGlynn mcglyng@tcd.ie
P2 (ICTP, Italy) - Adrian Tompkins tompkins@ictp.it/Diro Gulilat Tefera gtefera@ictp.it/Felipe de Jesús Colón-González fcolon_g@ictp.it/Riccardo Biondi rbiondi@ictp.it/Susanne Henningsen hennings@ictp.it
P3 (PLUS, Austria) - Peter Zeil peter.zeil@sbg.ac.at/Stefan Kienberger stefan.kienberger@sbg.ac.at/Michael Hagenlocher michael.hagenlocher@sbg.ac.at
P4 (SMHI, Sweden) - Colin Jones Colin.Jones@smhi.se/Grigory Nikulin grigory.nikulin@smhi.se/Monica Wallgren Monica.Wallgren@smhi.se
P5 (UoN, Kenya) - Winnie Mitullah wvmitullah@swiftkenya.com/Eric Othieno Nyanjom othieno_n@yahoo.co.uk/Fredrick Obonyo Mukanga fredmukanga@yahoo.com
P6 (AQUATT, Ireland) - David Murphy david@aquatt.ie/Olivia Daly olivia@aquatt.ie
P7 (ILRI, Kenya) - An Notenbaert A.notenbaert@cgiar.org/Jeffery Mariner j.mariner@cgiar.org/Bernard Bett b.bett@cgiar.org/Wachira Theuri w.theuri@cigar.org/Jusper Kiplimo jusronohk@gmail.com/Nancy Ajima n.ajima@cgiar.og/John Gachohi j.gachohi@cgiar.org
P8 (UR, Rwanda) - Theophile Niyonzima tniyonzima@nur.ac.rw/Jean Pierre Bizimana jpbizimana@nur.ac.rw/Jean Damascene Mazimpaka jdmazimpaka@nur.ac.rw/Rachel Murekatete rmurekatete@nur.ac.rw/Caritas Thereza Gasengayire tgasengayire@nur.ac.rw
P9 (SEI Tanz/York,Tanzania/United Kingdom) - Stacey Noel stacey.noel@sei.se/Neela Matin neela.mation@sei.se/Victor Kongo victor.kongo@sei.se/Richard Taylor richardtaylor.sei@googlemail.com/Sukaina Bharwani sukaina.bharwani@sei.se
P10 (Community Health, Uganda) - Agaba E. Friday agabafriday@hotmail.com/Didacus B. Namanya didamanya@yahoo.com
P11 (KEMRI, Kenya) - Andrew Githeko githeko@yahoo.com/Diana Karanja Diana@cohesu.com
P13 (UCT, South Africa) - Bruce Hewitson: hewitson@csag.uct.ac.za/Lisa Coop: lcoop@csag.uct.ac.za
P14 (UDUR, United Kingdom) - Mark Booth mark.booth@durham.ac.uk/Nicky McCreesh Nicky.mccreesh@durham.ac.uk/Dajana Dzanovic dajana.dzanovic@durham.ac.uk/Gary Mitchell Gary.Mitchell@durham.ac.uk
P15 (UNILIV, United Kingdom) - Andy Morse A.P.Morse@liverpool.ac.uk/Cyril Caminade Cyril.Caminade@liverpool.ac.uk/Anne Jones Anne.Jones@liverpool.ac.uk/Scott McGee S.Mcgee@liverpool.ac.uk
P16 (NUS, Singapore) - David Taylor geodmt@nus.edu.sg