Child Growth Monitor

The world is moving backwards regarding the elimination of hunger and malnutrition.¹   The situation of children is alarming, nearly 200 million children globally are undernourished according to The State of Food Security and Nutrition report in 2022.²

Young children and pregnant or breastfeeding women are the vulnerable to malnutrition. Their bodies have a greater need for nutrients, such as vitamins and minerals, and are more susceptible to the harmful consequences of deficiencies. According to WHO, children are at the highest risk of dying from starvation. They become undernourished faster than adults. Severely wasted children are 11 times more likely to die than those with a healthy weight. Undernourished children catch infections more easily and have a harder time recovering because their immune systems are impaired.³

An estimated 45 million children under the age of five suffered from wasting in 2020, the deadliest form of malnutrition, increasing children’s risk of death up to 12 times. In 2020, 149.2 million children under five years of age were too short for their age (stunting).

Stunting causes irreversible physical and mental damage to children. A stunted child is too short for their age, does not fully develop and stunting reflects chronic undernutrition during the most critical periods of growth and development in early life. Wasting, in turn, refers to a child who is too thin for his or her height. Wasting is the result of recent rapid weight loss or the failure to gain weight. A child who is severely wasted has a much higher risk of death, but treatment is possible.4

A major bottleneck in treating malnutrition is its detection through correct measurement and monitoring. Traditional child growth measurements are at risk of being flawed and can lead to fatalities.

The challenges to ending hunger, food insecurity, and all forms of malnutrition keep growing and continue to be of great concern – with the Ukraine conflict, the Covid-19 pandemic and climate change leading to an aggravation of the global hunger crisis. Conforming to FAO, the number of malnourished people, especially women, and children, may further increase and impede any progress in achieving the 2030 global nutrition targets.

 _________________________

1. Source: https://www.fao.org/publicatio...

2. Source: The state of food security and nutrition in the world 2021 (who.int)

3. Source: https://www.who.int/news-room/...

4. Source: https://data.unicef.org/topic/...

Traditional tools and methods to measure malnutrition (MUAC¹ tape, scale, height board) require expertise and good equipment. Traditional tools are error-prone (for instance, MUAC tape only has a sensitivity of 20%, producing a lot of false negatives), and stressful for the child. Even when measured correctly, the process of recording and communicating data is susceptible for errors and manipulation. The identification of malnourished children via traditional tools and methods is impeded by the following factors:

• Height boards and scales are heavy and hard to handle in remote places
• Affordable equipment and skilled experts are often not available in some countries of the Global South.

As a result, currently, only 35% of children are measured correctly, meaning almost two-thirds are not measured correctly and will not receive proper treatment.

Data-driven technologies – like the Child Growth Monitor (CGM) – can provide better quality in less time. -CGM supports untrained users to conduct easily and quickly good quality anthropometric measurements.

The CGM app uses image data collected via smartphones and Artificial Intelligence to measure any form of malnutrition. In doing so, the CGM app aims to provide better quality measurements, higher quantity, real-time results, and the highest standard of data protection (GDPR). Moreover, the app will be easy-to-use for its main target group: health workers and nutritionists. A user-centric approach guarantees that the app fits their needs. The development is conducted in close collaboration with users (local community workers/health workers). It is designed for different use cases, such as humanitarian contexts or development aid (nutrition surveys, continuous growth monitoring).

Currently, the app can predict stunting quite at a high level.

_____________________________

1. MUAC: Middle-Upper-Arm Circumference

In order to address malnutrition, people working in the nutrition sector need tools they can rely on. Once malnutrition is identified, partners such as health workers, caretakers, iNGOs and governments can use the data to provide treatment more efficiently and effectively. But only what is measured correctly can be managed correctly. Correct measurement is the first step in treating children who are affected by insufficient nutrient and energy intake, poor digestion and nutrient absorption, and frequent or prolonged illness.  

The four main target groups of CGM are: (1) children from 6-59 months as the main project participants in India, Nepal, Bangladesh, Uganda, Zimbabwe, Ethiopia, and possibly other countries. Once the app is fully developed and rolled out in countries with high prevalence of malnutrition our goal is to measure children in the millions.

Other target groups are: (2) Health care workers, caretakers as users of the app (3) community members, parents, caretakers as participants in measurement and follow-up health activities and (4) Government and other stakeholders at higher levels, as potential user organizations and multiplicators with many users.

Welthungerhilfe is one of the largest private aid organizations in Germany. It was one of the first global initiatives in the fight against hunger when it was founded in 1962 as the German chapter of the Freedom From Hunger campaign, which was led by the UN Food and Agriculture Organization (FAO). The organization’s vision is a world in which all people can exercise their right to a self-determined life in dignity and justice, free from hunger and poverty. Welthungerhilfe follows the principle of providing help for self-help to sustainably improve people’s living conditions, working with local partner organizations to strengthen approaches built from the bottom up in order to ensure that projects are successful in the long term.

 As part of Welthungerhilfe, the Child Growth Monitor project has access to a huge network of global nutrition experts. The Child Growth Monitor project is built with a team of 15 nutrition, machine learning, data science, software architecture, UI/UX design, hardware engineering, backend engineering, data protection, business development, fundraising, and community management experts. Our leadership team combines 50+years of experience in running and scaling up digital (social) businesses at the intersection of healthcare and software development. 

CGM is still in development mode. Currently, we are active in 5 countries for field tests and data collection purposes. The CGM team has developed a first solution that works well for predicting height & stunting. The testing to prove the app functionality will start in the current month (May 2023) on the ground.

We have collected training data (image data plus manual measurements of highest quality) in the very high 5 figures level. All of the children that have been measured so far have received follow-up treatment if detected as malnourished.

The award prize will be used to continue working on the app enabling it to produce more accurate results and different features (height prediction, weight prediction, immediate feedback to the user, ID-system, user dashboard, privacy and data protection safeguards). It will also support the identification of additional smartphones that are compatible with the CGM app. Finally, it will also support the CGM team to train the Artificial Intelligence for improved height and weight prediction.

The goal is to set up CGM as an open source, non-profit entity that is independent from external funding. To determine the way forward to achieve this goal is part of the project.

CGM will be set up as a software entity, providing products and services to the international nutrition community to conduct high-quality anthropometric measurements, thus, contributing to the goal of “Zero Hunger”. 

Traditional measurement tools (MUAC tape, scale, height board) and methods are expensive, heavy, error-prone and stressful for the child. As measurements are conducted based on tools and expertise available, if precise tool and/or trained enumerators are missing, the quality of measurements drops accordingly.

Comparable tech solutions (BST, SAM Photo diagnostics, Wadhwani newborn app) are either stagnating, delivering sub-par results or targeting a different age group.

The CGM app will help solve these problems. App users take a short video on an off-the-shelf smartphone. The collected data is used to produce a digital model of the child, which is fed into neural networks. The calculated results are compared to harmonized growth charts by WHO, showing the nutrition result of the child. The user receives the following results on their phone: height in millimeters, later weight in grams (and potentially MUAC millimeters), growth charts and z-scores (stunting, wasting, underweight). 

Our strategic goal is to scale up CGM across the Global South after development is completed. CGM has massive potential to reach.

 After our solution is developed, we will create impact in millions. We predict that millions of children will be measured by the end of 2028.

 Once fully developed, CGM can be a game-changer in detecting malnutrition. It has the potential to contribute significantly to the global nutrition community’s vision of Zero Hunger by harnessing the power of AI to measure children easier, more frequently, and more accurately. 

 *We are happy to more detailed numbers in a non-public environment. 

We measure success along the indicators of our theory of change. For our impact, outcome and outputs please see next question.

The Child Growth Monitor project will measure its progress based on the following 3 main indicators:

Product development:

• Indicator 1 – Our app runs on multiple devices
• Indicator 2 – Height and weight prediction work on a good level
• Indicator 3 - Product development is documented in compliance with medical device regulations.

User tests and field integration

• Indicator 1 Product functionality testing have been conducted with users and user organizations. 
• Indicator 2 – A strategy and a plan to integrate CGM into existing workflows and systems are developed.

 Sustainability plan

• Indicator 1 - A strategy for further roll-out and use of the Child Growth Monitor is developed.

_________________________

*We have more detailed information, which we’re happy to share in interviews. We would not like to make this public. 

In order to address malnutrition in children, people working in the nutrition sector need tools they can rely on. Once malnutrition is identified, partners such as health workers, caretakers, iNGOs and governments can provide treatment. But only what is measured correctly can be managed correctly. Correct measurement is the first step in treating children who are affected by insufficient nutrient and energy intake, poor digestion and nutrient absorption, and frequent or prolonged illness. 

For the next year, we follow three major lines of action to achieve our vision that every child can grow up free from malnutrition, reaching their full mental and physical potential. The CGM team is following the logic of the framework bellow:

Impact: Contribute to the SDG2 "Zero Hunger" by providing a mobile app to detect and measure any form of malnutrition via image data and Artificial Intelligence. 

Outcome: The Child Growth Monitor (CGM), a mobile app for anthropometric surveys and growth monitoring, enables the measurement of. Malnutrition is measured accurately, easily and reliably for informed decision-making on the medical treatment of malnourished children aged 6-59 months. The following outputs and activities will be conducted from May 2023 to July 2024.
Output 1: The app is technically able to produce accurate results, has different features (height prediction, weight prediction, immediate feedback to the user, ID-system, user dashboard, privacy and data protection safeguards) and is constantly improved based on user feedback. 

• Activities: Identification of additional smartphones that are compatible with the CGM app. Exploration of new product development acceleration opportunities. Training the Artificial Intelligence for improved height and weight prediction. Further development of the frontend and the backend of the CGM app. Improvement of data input via user feedback.  

Output 2: The Child Growth Monitor app is tested with the users (frontline health workers, nutritionists) and target organizations (NGOs, government bodies, International Organizations, academia, and statistical offices and public-benefit federal enterprises). 

• Activities: Conduct of a needs analysis to define user cases. Conduct of at least two product functionality tests in project target locations.

Output 3: A strategy for further roll-out and use of the Child Growth Monitor is developed.

• Activities: Creation of a strategy document guiding the CGM project to become independent from external outside funding.

Our innovative solution uses image data collected with a smartphone (camera sensor and infrared / LIDAR sensor) and computer vision methods to build 3D models of children's bodies. Image data are used to train deep learning models to accurately predict children's height and weight, which are crucial indicators of malnutrition.

With traditional methods, appr. only 35% of children are measured correctly. I.e. two thirds of children are not measured correctly or not measured at all. With our solution, we can increase the ease-of-use, the speed and accuracy of measurements, providing a tool to the global nutrition community to conduct more and more accurate measurements and following up with treatment. 

Our local partners on the ground help us to ensure that social norms are respected. In addition, following Welthungerhilfe principles, CGM aims at measuring all children, regardless of gender, socio-economic status, religious background, or race. In particular the latter, an important topic discussed in AI, does not appear to pose an issue for CGM. First tests in Namibia have shown that our algorithms, which have been developed 100% based on data from Indian children, work for children from other ethnicities. Our apps will work on children with no physical impairment or disabilities. Together with our users, we define their needs and demands, such as local languages, icons, audio, visual aids for illiterate users, which have already been implemented. When rolling out CGM in different countries, we will adjust our app according to local user needs to make sure the diversity of our stakeholders is reflected.  

CGM will be set up as an open source, non-profit Social Entity, providing products and services to the global nutrition community: Organizations measuring children for malnutrition (governments, International Organizations, NGOs/iNGOs and others.

 Our goal is not to maximize revenue, but to maximize impact. Thus, our business model is to acquire enough money to break even / pay for our expenses. We want to become independent from short-term external funding and set up CGM in a way to become self-sustainable. We will reach sustainability by asking for money for our products and services.

 CGM will be able to replace parts of the traditional – subpar – tools and methods for a fraction of the costs: The app as a measurement tool and services around the app.

 In our business case we will deal with three different stakeholder: Buyers paying for products & services, users using our product and beneficiaries who will gain from our products and services. Buyers are the organizations mentioned in paragraph one. Users are “last mile” health workers and beneficiaries are children between 6 and 59 months.

________________________

  *We are happy to answer more detailed questions in a non-public environment.

The goal is to set up CGM as an open source, non-profit entity that is independent from external funding. To determine the way forward to achieve this goal is part of the project (output 3, development of a roll-out strategy). Afterwards, CGM will be a software entity, providing products and services to the international nutrition community to conduct high-quality anthropometric measurements, thus, contributing to the goal of “Zero Hunger”.

In case, we don’t have a fully developed solution, i.e., CGM as a medical tool, validated in a scientific study and registered as a medical device after the end of the planned project period, the app development will be continued until it achieves the status of a registered medical tool.

There are several examples indicating that we will achieve our goal of self-sustainability:

• First tests with potential paying customers / organizations have shown that our products & services are valuable and are willing to pay.
• We have received significant funding from very reputable organizations helping us to support our vision, esp. the development of a tech product
• We would like to underlie the fact that we have received repeat funding from multiple donors, proving that we are able to reach our goals and match the donors’ expectations
• Here are some of our donors and partners: To maximize impact, we are working together with some of the best partners in the industry such as the World Food Programme, UNICEF, German Corporation for International Cooperation, German Federal Ministry for Economic Cooperation and Development, European Union, Microsoft, GSMA/FCDO, Action Against Hunger, The Boston Consulting Group, Sony, the Indian Government and others to achieve our goals. Academic partners are the university of Ingolstadt, Tilburg university / Zero Hunger lab, university of Namibia.