MIT Solve

Solution overview

Our Solution

Child Growth Monitor

Tagline

A game-changing app to diagnose malnutrition

Pitch us on your solution

PROBLEM

200 million children globally are malnourished. 5.6 million child deaths are caused by malnutrition each year. All of those lives could have been saved. In the fight against malnutrition, detection is the most important step. But detection and diagnosis is not easy. Traditional methods of measuring malnutrition are not only slow and expensive but often produce poor data and wrong assessments of the situation, wich leads to wrong or too late action.

SOLUTION

he Child Growth Monitor - uses a mobile phone to record images (3D model, pointcloud) of childrens’ bodies in combination with Artificial Intelligence (AI) to calculate height, weight and middle upper-arm circumference (MUAC). Diagnosing the nutritional status of a child will be cheaper, quicker, better quality and will deliver results in real-time.

IMPACT

Once fully developed we are planning for 13 million to 66 million children to be measured with our app.

Embed your elevator pitch here:

What is the problem you are solving?

200M children globally are malnourished. 5.6M child deaths are caused by malnutrition each year. All of those lives could have been saved. Hunger is the biggest solvable problem in the world.

In the fight against malnutrition, detection is the most important step. But detection is not easy. Traditional methods of measuring malnutrition require:

- costly hardware (scales and measuring boards cost over $500 per set)

- expert knowledge

- and time.

Traditional methods of measuring children frequently result in poor data and wrong assessments of the situation. Leaving a lot of caretakers unable to properly assess the nutrition status of a child. Thus, preventive measures are often taken too late or not at all.

“Malnutrition in children is especially harmful. The damage to physical and cognitive development during the first two years of a child’s life is largely irreversible. Malnutrition also leads to poor school performance, which can result in future income reduction. Adults who were undernourished as children are at risk of developing diseases such as obesity, diabetes and cardiovascular issues.” Unicef: https://www.unicef.ca/en/malnu...

Ending malnutrition could be a watershed moment in global health, helping an entire generation of children escape hunger and poverty and reach their full potential.

Who are you serving?

We have to differentiate between the users and the beneficiaries of our solution - and between the primary and secondary user & beneficiaries groups (according to our scaling plan).

A) Primary users & beneficiaries groups

Users: frontline healthcare workers in 52 countries (countries with “serious”, “alarming” and “extremely alarming” hunger; according to the Global Hunger Index https://www.globalhungerindex....)

Beneficiaries: children from 0-5 years of age in those 52 countries

B) Secondary users & beneficiaries groups

After our solution has proven to work in the primary user group, we will roll it out to further user groups:

Users: Healthcare workers in more than the above 52 countries; parents; teachers; pharmacists, General Practitioners

Beneficiaries: children from 0-5 years of age in all developing countries or even globally

Our solution is already out in the field in the hands the users. Although, at the moment we're using our app to collect data, not yet to give a diagnosis. Our solution is based on Machine Learning. In order to make our existing algorithm more accurate we need more data. We have collected scans from 10k children in India and are about to collect 75k more scans.

What is your solution?

The Child Growth Monitor uses a mobile phone to record images (3D model, pointcloud) of childrens’ bodies in combination with Artificial Intelligence (AI) to calculate height, weight and middle upper-arm circumference (MUAC). It will enable anyone with a smartphone to diagnose the grade of malnutrition of a child in real-time. The Child Growth Monitor will be cheaper, quicker with better quality and deliver results in real-time.

We are collecting the following metrics:

Height (Gold Standard)
Weight (Gold Standard)
MUAC (Gold Standard)
Gender
Age

With our data we will be able to classify our diagnoses into: Moderate Acute Malnutrition (MAM), Severe Acute Malnutrition (SAM), Low Birth Weight (LBW) and stunting. The goal is to produce gold-standard anthropometric measurements for both weight and height. We will deliver results in a “traffic light system” and also as growth charts time-series.

The Child Growth Monitor decouples data-collection and diagnosis from producing measurement results. This will facilitate hardware independence in the future.

Our infrastructure is based on open-source components and can run in any cloud-environment or data-center. Development and operation is agile and highly automated through state-of-the-art DevOps tools and processes.

We aim to deploy our models on dedicated servers (cloud-enabled) and on selected mobile devices (offline-first).

Deploying on mobile devices will enable us to do an automated diagnosis, which will be available shortly after scanning a child even when in an offline region. Only if our confidence values indicate that a scan (and one or two retries) might not produce results accurate enough for a diagnosis, data is uploaded to the cloud for analysis and if possible to be included into the scan sample database including a gold standard manual measurement.

The Child Growth Monitor measurement API (both online and offline) will use multiple approaches to produce the results needed for the nutritional assessment of a child:

Approach 1: Building a 3D model

Approach 2: Building a spatio-temporal model

Approach 2: Other Future Approaches

Select only the most relevant.

Reduce barriers to healthy physical, mental, and emotional development for vulnerable populations
Decrease inequalities, stereotypes, and discrimination, from birth

Where is your solution team headquartered?

Bonn, Deutschland

Our solution's stage of development:

Pilot