Grand Challenges is a family of initiatives fostering innovation to solve key global health and development problems. Each initiative is an experiment in the use of challenges to focus innovation on making an impact. Individual challenges address some of the same problems, but from differing perspectives.
Kerry Selvester of Associação Académica de Nutrição e Segurança Alimentar (ANSA) in Mozambique will develop an interactive online mapping and data visualization tool to identify high-risk and under-served populations in Mozambique to improve the outcome of health campaigns. They will generate high-resolution maps of existing geospatial datasets such as estimated travel time to health facilities and use remote sensing and make predictions using machine learning approaches to generate new maps of other health-related indicators. They will also build an interactive dashboard for users to quickly and easily access the available data for health campaigns. They have performed a pilot of the approach to help the Ministry of Health better target their limited resources for tackling the spread of COVID-19. They will evaluate the results of this pilot and identify additional datasets and data gaps that will be collected and consolidated, and the visualization tool will also be updated and its value qualitatively assessed.
Anna Winters of Akros Inc. in the U.S. will adapt an existing web-based mapping tool currently in use in several low-middle income countries that guides and maps the progress of health-related campaigns, to incorporate human movement and thereby improve campaign coverage. Although the existing tool maps populations at the level of individual households, it fails to incorporate spatial-temporal population changes caused by permanent relocation, seasonal migration, and short-term movements, which are more difficult to track. To address this, they will incorporate published human movement models and explore agent-based models, mathematical models, and cell-phone-data based models, along with survey data, to build their own model that can predict the location of individuals at specific times, and to identify people most at risk of being missed. This will enable better forward planning for predicted movements. They will evaluate their approach using the malaria Indoor Residual Spraying campaign in Luapula Province in Zambia.
Mark Adams of Population Services International in the U.S. will develop a "Digital Gateway" that provides health campaign managers easy access to a range of datasets to improve the planning and performance, and lower the cost, of health campaigns. Health campaigns generate data, such as population estimates, locations of health clinics, and mobile phone data, that could make planning new campaigns much more efficient, but these data tend to be difficult to access. They will develop a mechanism that ensures campaign managers share key information from past and current campaigns and develop a digitized algorithm that identifies data requirements for campaign planning. The Digital Gateway will be designed to host the data requirements for the algorithm, store and visualize datasets from past campaigns, display links to access outside data sources, and provide help to navigate the data for campaign managers to identify and apply the most relevant information when planning their own campaigns.
Nyasatu Ntshalintshali of the Clinton Health Access Initiative in the U.S. will use a benchmarking approach to guide individuals and teams delivering mass drug administration (MDA) campaigns in low-middle income regions in order to improve coverage. Individual and team behavior during health campaign planning and implementation have been identified as major causes of variation in campaign quality and performance. To encourage behavioral changes, they will use a benchmarking technique that shows individuals that their peers behave in the desired way. To be effective, the benchmark needs to be carefully targeted to the individual. They will update an existing MDA tracking application to display benchmarking prompts delivered in real-time, and evaluate its effect on the quality and coverage of MDA in a randomized control trial with 12 campaign outreach teams in the Kingdom of Eswatini.
Qingfeng Li of Johns Hopkins Bloomberg School of Public Health in the U.S. will develop a computational simulation tool to optimize the design of health campaigns in low-income settings. Health campaigns are complex events involving multiple, interconnected components, such as families and socioeconomic contexts, as well as being time restricted and targeting specific populations. Their tool uses geospatial measures and community maps, and it includes an automated algorithm to test different design strategies to identify the optimal design. They will develop their tool and test whether it can improve the vaccination rate from the current average of 30% to 50% for a selected immunization campaign in Ethiopia, where rates are low and uneven across the country.
Robert Miros of 3rd Stone Design, Inc. in the U.S. will adapt their portable vaccine refrigerator, which is battery powered and can be monitored remotely, to maximize the charge life so that it can support vaccination campaigns in low- and middle-income countries. Vaccines are normally stored in ice boxes and manually tracked, both of which are unreliable and can cause spoiling. Their vaccine refrigerator integrates thermo-electricity cooling and battery power together with an algorithm that can sense temperature and adjust it according to the available power. They will develop a new algorithm that can correlate charge times with vaccination routes, and integrate upgraded battery technologies to prolong charge life. This will then be field-tested in a selected vaccination campaign.
Coite Manuel of Food Chain LLC. in the U.S. will develop a web-based tool that takes existing road network and population data for any country, divides it into regions, and identifies the closest health facility based on type and time of travel to improve health campaign planning and better monitor population health. Current approaches to map catchment areas for health facilities use administrative boundaries or population statistics, which often don't reflect where people actually go. Once the site catchment areas have been defined, they will overlay heatmaps to identify vulnerable areas for targeting during health campaigns. They will use opensource tools and free or low-cost technologies such as MapBox, and partner with at least one country to help develop and test their technology.
David Hammel of Balcony Labs Inc. in the U.S. together with their partner Direct-Relief will develop a communication tool that combines smartphone messaging with geographical information systems to enable health campaign managers to communicate directly with health workers and civilians in a specific region of interest to improve the impact of their campaigns. The tool enables managers to send messages such as alerts, instructions, or surveys, directly to target areas, as well as collect geo-specific information in real-time for updating campaign designs and evaluating outcome. Once they have developed the dashboard and mobile software module, in partnership with a leading NGO, they will evaluate its performance in 2-3 campaigns incorporating 100,000 mobile phone users.
Pascal Geldsetzer of Stanford University in the U.S. will develop a computational tool to support health campaigns in low- and middle-income countries that can predict the number and location of the people that need targeting. They will use freely available databases, including the Demographic and Health Surveys (DHS), covering over 10 million households and high-resolution population estimates to estimate the percentage of children under 5-years-old who are un- or under-vaccinated within each 30m by 30m area. They will then use algorithms to determine the optimal numbers and locations of campaign posts needed to ensure a desired vaccination coverage within a given budget. This work will lay the foundation for developing a simple Android application to help managers increase the impact of health campaigns.
Hannah Wild of Stanford University in the U.S. will develop a modelling-based approach that uses remote sensing and geospatial analysis to include neglected and high-risk nomadic populations in health databases and for campaign planning. Nomadic pastoralists are some of the poorest populations but they are often missed by health services and campaigns because they are difficult to track. They will design algorithms and methods that use satellite imagery and open access geospatial data to capture population movements over time, which will be validated in the field. They will also develop the technical infrastructure to scale these methods and produce automated packages and best-practice guidelines so they can be implemented in a variety of low-resource settings.