ITGH: Smart Sanitation

This project addresses the urgent challenge of poor sanitation which contributes to disease transmission, particularly in developing countries that are resource constrained. The solution utilizes data from sensors in public toilets, sewage and water systems, in combination with ecological and environmental changes from earth observation data to detect disease-causing pathogens in communities and disease outbreaks. The data is used to train AI models to predict potential outbreaks, equipping healthcare workers, policymakers, and community members with critical information necessary for developing and implementing mitigation strategies to fortify community health security. The system is connected to a centralized resource sharing platform to visualize, synthesize, and utilize data, helping to increase disease surveillance.

In South Africa, about 2.6 million households do not have access to safe drinking water and 400,000 households do not have a toilet facility. These circumstances lead to a significant proportion of young children and vulnerable individuals to die of preventable diseases - one in four children before age five and one in 10 before age 15. The high prevalence of these water and hygiene-related illnesses such as diarrhoea, cholera, intestinal worms and typhoid contribute to malnutrition and cognitive deficiencies, which erodes health and economic outcomes in communities, as the full potential of affected individuals are not realized. These are also fertile conditions for incubation of disease-causing pathogens which could lead to outbreaks. 

Globally, approximately 3.8 trillion litres of waste are left untreated or lost and 892 million people in the world are still practicing open defecation (UNICEF/WHO JMP, 2017). In India alone, it is estimated that as a direct result of poor sanitation, about $54bn of GDP is lost annually as well as 200,000 human lives. These figures are not insignificant and improvements in these conditions could lead to improved outcomes for human life.

Pathogen detecting sensors are placed in public toilets, sewage systems, and water sources in pilot communities. Signals from the sensors are transmitted over local cellular networks and satellite communication channels to remote management systems. Earth observation data from two satellites provided by the European Space Agency assesses environmental and ecological changes including water chemistry, conditions at dumping sites, temperature changes, weather patterns, and water levels. These observations are used to model the impact of weather on water levels and how changes to these conditions might help disease propagate in communities.  Basic edge-analysis change impacts are also conducted to understand how the topography affects the water flow and any associated potential for contamination. 

The combination of data from sensors and Earth observation data from the satellites are used not only to identify the presence of disease-causing pathogens in water bodies in communities, but to also develop and train AI models that predict possible disease outbreaks based on those observations. The data and findings from the analyses are published to a centralized platform that is accessible to health practitioners, equipping them with the knowledge required to make rapid decisions aimed at controlling the spread of any disease outbreaks.

The eThekwini Municipal Area or Durban, South Africa is the site of the pilot and its residents will be the primary beneficiaries of this implementation. About one million people or 26 percent of the population currently live in informal settlements with poor sanitation conditions (eThekwini Municipality, 2019). There are about 2500 Community Ablution Blocks - shared sanitation and water facilities - and about 85,000 dry toilets or pit latrines. Most of the informal settlements do not have proper stormwater drainage systems and as a result, water collects on the surface with all the greywater that residents disposed of, using the bucket system when it rains. 

Isolating the core issue of poor sanitation and designing possible solutions to address them came from substantial interactions with locals in the community, including researchers at the University of KwaZulu-Natal's School of Built Environment and Development Studies, who have conducted extensive research in the space. Clinical data provided evidence of the impact of preventable diseases on the local population as a result of poor sanitation. Economic data revealed the impact on individual productivity and household income. Taken together and with input from community members, it became clear that the issue of poor sanitation presented significant risk to the local population. Knowledge of the local context including cultural and behavioral characteristics, imparted by local groups also helped to inform the design of the solution. 

By identifying the presence of disease-causing pathogens in the community and being able to predict possible outbreaks, we can equip policymakers and health practitioners with critical information needed to make rapid and informed decisions. 

Our solution utilizes technology-based solutions to strengthen disease surveillance and equip healthcare workers with information necessary to take actionable steps to strengthen health security and mitigate the spread of disease in underserved and "at-risk" populations.

Utilization of space technology and IoT to monitor ecological changes, pathogen detection, and in combination with data from community input helps with developing predictive systems to strengthen disease surveillance.

Parsing data for use by healthcare workers and policymakers to make rapid and informed decisions that can help mitigate disease prevalence and burden, equips healthcare workers with knowledge to develop effective actionable containment plans.

A feasibility study commissioned by the European Space Agency has been conducted and the first pilot in the eThekwini Municipality is scheduled to be implemented. This pilot stage will benefit residents of two informal settlements - Quarry Road and Thandanani, with approximately 1450 households. Following the pilot, we expect to rapidly scale up the solution, first throughout the broader municipality which has a population of approximately three million people and then the country of South Africa.

The solution combines repurposed space technology to conduct ecological and environmental observations which is then combined with data from IoT sensors - acoustic in public toilets, from fecal sludge in sewage systems, and in water systems to detect the presence of disease-causing pathogens. Using these datasets, machine learning models and AI can be developed and trained to predict potential community disease outbreaks, when the conditions that are conducive to this phenomena converge. The data and results from the analyses are maintained in a global, centralized, and accessible platform with no government intervention, which is an important feature for communicating vital and valid information.  

The combination of sanitation systems data and earth observation data to predict disease outcome is not currently practiced, yet we know that environmental and ecological changes may create the conditions necessary for diseases to incubate and propagate. Analyzing fecal waste in community sewage systems also eliminates violating individual privacy. The availability of both ecological and fecal analysis data presents opportunities for researchers and health practitioners to utilize in their various approaches to understanding the nature of disease spread and their effects in communities. 

In the one year pilot stage, residents of informal settlements in the eThekwini Municipality in South Africa will benefit, before being extended to the rest of the country. Specifically, two settlements: Thandanani and Quarry Road will be recipients of the first implementation, approximating 1450 households. That number is expected to rise significantly in five years, with a planned second implementation in India, where there are 344 million people without regular access to toilets in the country, according to 2017 statistics from the World Health Organization and UNICEF. We are working to identify a site for that implementation. Given the country's vast population and its densely populated cities, we can project that the impact of improving sanitation conditions will be significant. 

The Boston Scientific Group issued a Smart Sanitation market forecast in 2018 stating that the market would be worth $6bn. Of this, over 50% is in emerging or third world markets. Woodco Bioscience estimates that 20% of the emerging market is likely to add health monitoring to Smart Sanitation solutions. 20% of the emerging market represents 10% of the overall smart sanitation market, hence we estimate this as the addressable market for a combined sanitation and health monitoring solution. Following a successful pilot demonstration, key markets include South Africa, India, Bangladesh, Nepal and Senegal serving between 29  and 32 million people. This is based on expected tenders and focuses on population centers with existing sanitation issues.

In addition to continuous monitoring of data via the smart management system, which will keep track of KPIs, the Institute for Technology & Global Health will conduct a randomized control trial to empirically measure the effects of the pilot implementation. 

KPIs include the number of residents in settlements with diarrheal disease and other preventable diseases, variation in household income over time, number of employment days, viral load in waste water, dumping activity, and usage of public toilets, and number of health workers who access our system.

Data collection and analysis is built into the evaluation framework, giving insight into performance of the technology and will allow us to be iterative with the solution.  

The core team is comprised of 14 individuals across the various organizations that are contributing to the solution: Institute for Technology & Global Health - 3, University of KwaZulu-Natal - 3, Woodco Bioscience - 2, Mindeed - 2, Toilet Board Coalition - 1, European Space Agency - 1

The team is composed of both local and international experts across a range of disciplines including global and population health, computer science including machine learning and artificial intelligence, development and public economics, public policy, and behavioral science. Importantly, identification of the core issues came from local researchers, whose input was instrumental to the development of the solution. All of the "ground truth" data from surveys and focus groups came from local researchers at the University of KwaZulu-Natal in the eThekwini Municipality, who themselves are residents. Team members have extensive research backgrounds stemming from their doctoral work and others with technology and industry experience. This combination of skillsets and experiences very well positions the team to effectively manage the development, implementation, and evaluation of the solution, while using the data to iterate its design and processes.

This project was selected by the World Health Organization (WHO) and International Telecommunications Union (ITU) as a 2021 Focus Group on Artificial Intelligence for Public Health. Being able to draw upon the support and expertise of experts across the two organizations is a vital asset. 

Equitable distribution across a range of characteristics is inherent to the team, including gender, race, age, and delegation of responsibility. Importantly, implementation of the solution is driven by the local community of residents who themselves are diverse across age and gender. Their input and information on local conditions helped to inform the solution. The broader team is comprised of scholars and industry experts who also vary across age, race, gender, and socioeconomic backgrounds. Although I - Khahlil Louisy - am in the driver's seat for this project and a researcher at Harvard University, based in Cambridge, MA, I draw from my experiences growing up in an underdeveloped island in the Caribbean. My interest in economic development and public health comes from my experiences in communities that are heavily constrained by resources. Also significantly is the lateral structure of the team, which distributes authority equally. 

The Institute for Technology and Global Health is a non-profit organization that is committed to diversity, inclusion, and equity. This is reflected in the organization's leadership team which has a current 50-50 male to female make up, and diverse across race, age, and socio-political beliefs.  

The Solve Challenge and the Solve community's focus on the development of high impact, technology-enabled solutions aimed at addressing the world's most pressing challenges, is perfectly aligned with the mission of our organization - the Institute for Technology & Global Health (ITGH). At ITGH, we actively pursue avenues through which the application of technology can lead to improved public health outcomes and is driven by a robust research agenda. But solving (pun intended) the world's pressing challenges, requires active collaboration among those bold enough to tackle them. That is because the challenges are complex and require weaving together methods drawn from interdisciplinary approaches and expertise. This is one of the biggest draws to the Solve challenge - being part of a vibrant community of passionate, mission-driven organizations and individuals who are as relentless as we are in our pursuit of developing solutions capable of improving and transforming human lives. 

We believe that the Solve network coupled with our drive, can get us there. Mentorship and insights from organizations and experts who have been bridging the gap between technology and government is vital. In our work, this has been a persistent issue, governments are not well-equipped to handle the adoption of new technologies efficiently. This is a significant challenge and one that we will need assistance with. 

We will need experts across the domains of human resources, business development, financial, and legal to help with structuring the collaboration between non-profit, for-profit, and government. What is the ideal business structure that can drive change while maximizing returns? How do we pitch such a structure to investors who may not be familiar with these operational processes? What are the legal and regulatory hurdles that must be navigated? How should we navigate global supply and distribution channels? These are all questions which we are actively dissecting and having experts to provide guidance would be beneficial. 

While we have an incredible team of research and industry experts, having partners with substantial experience working with governments would be invaluable. Organizations like the Bill and Melinda Gates Foundation have made tremendous headway on issues of global sanitation and have deep expertise working with governments. Their insights from previous implementations, research, and experience in the space could help us to further shape our approach to sanitation challenges. 

As we are an organization driven by research to demonstrate impact across economic, health, and social dimensions, the Alfred P. Sloan Foundation would be an ideal partner because of our alignment on the economic impact of technological development and their applications to solving public challenges.  

Other partners of interest are Atlassian for their expertise in custom software development and APCO Worldwide who can help us with the development of a modern, agile, mission-driven organization.

While the site of our pilot project is South Africa, we do plan to expand the offering to the United States. Data and learnings from the pilot phase will help us improve the solution. Indeed some U.S. States struggle with poor sanitation and water conditions. As cities across the country move toward becoming "Smart Cities," integrating solutions such as the "floating lab" into community sewage and water systems can help with early detection of disease outbreaks. 

Additionally, one component of our solution involves a data and resource sharing platform, which will consist of a global database of data collected and analyzed by our solution. As movement across geographic boundaries continue to increase, so too is the potential for the spread of disease. This was made evident by the coronavirus global pandemic. If U.S. health workers and policymakers can be immediately notified of an outbreak in another part of the world from data that is not subject to any interference by governments, they can take action to reduce the spread of diseases to protect their communities. 

Data from 2019 revealed a recorded 89,285 refugees in South Africa for that year alone, who emigrated from some of the world's poorest countries - The Democratic Republic of Congo, Somalia, Bangladesh, and Pakistan, among others. Many of these refugees move to informal settlements and participate in the informal sector. The initial implementation of our solution is specifically focused on addressing the sanitation needs of residents of the informal settlements in the eThekwini Municipality in South Africa. These areas are serviced by Communal Ablution Blocks and Pit Latrines, with limited infrastructure to protect water quality. As a result, the community is plagued by communicable and preventable diseases from poor sanitation conditions. 

Winnings from the Andan Prize will be applied to the acquisition of a sufficient number of sensors that can continuously monitor the presence of pathogens in sewage and water systems. As explained in our Theory of Change, reducing disease burden from preventable diseases could lead to increased individual productivity as they are no longer constrained by illness and therefore, have a positive impact on economic outcomes. The sensor data combined with satellite data will allow us to develop predictive models that can detect potential outbreaks prior to them taking place. This information equips healthcare workers, policymakers, and community members with knowledge to take actionable steps aimed at mitigating the spread of diseases in especially fragile populations.

In many informal settlements, woman are often the ones tasked with household chores, including traveling to water sources to wash clothes, gather water for cooking, drinking, and bathing. These are usually the sites of infection because of unsanitary conditions. Unfortunately, these infections are then transmitted to the rest of the household. 

The effects of illness among women on households are significant. Women not only are responsible for maintaining households, but also caring for children, which means that when they are incapacitated as a result of illness, there are implications for other household members. Diarrhoeal disease, which is common in the informal settlements, also has implications in particular for pregnant women, who are more susceptible to illness. The risk and consequences of dehydration for pregnant women are more severe than for the general population. Diarrhoeal disease can result in dehydration, which may impact amniotic fluid composition, foetal development, or induce preterm labor. While our solution does not target women specifically, the technology has potential to enhance women and girls’ quality of life by: (1) reducing rates of illness that may cause girls to stay home from school and, as a result, reducing the need for mothers to stay home to take care of their child and/or the impediment to completing the necessary tasks around the home, (2) empower women by including their insights into the implementation of this technology within their own communities via participant-driven research, potentially offering them an opportunity to voice actionable opinions and go against gender norms. 

Our solution is powered by AI. It involves a complete system to collect, treat, and process human waste, managed by a single ICT system that combines "ground truth" data from surveys and sensors of ecological, biological, and pathogen data with AI and machine learning to predict disease outbreaks. To monitor our impact, we aim to create and synthesize novel data sources, including sensors for the autonomous detection of waterborne pathogens, terrestrial environment, satellite Earth observation, GNSS data, and toilet usage data; this data can then be systematized in a sharing platform for analysis to increase the understanding of linkages between different environmental factors through AI and machine learning.