Today malnutrition is emerging as one of the single biggest contributor of child mortality in India. World Bank data indicates that 44% of Indian children under the age of 5 years are severely malnourished. Malnutrition can increase children’s susceptibility to illness and reduce their ability to learn and be productive later in life. Therefore, identifying these children at the right time is critical for providing the appropriate treatment. mNutrition creates a readily-available platform, which can be utilized by health workers to provide a quick and easy identification of children, who are in dire need of treatment.
How does mNutrition work?
mNutrition is a mobile application which gathers the child’s relevant details along with parameters like age, sex, height, weight, etc. and using the “WHO 2006 weight-for-height/length and MUAC tables”, which are the diagnostic indicators of malnutrition, calculates the data and gives a ready indicator if the child is malnourished. The app can be used with ease by local community workers most of who are not well-literate and poorly trained with limited skills. mNutrition can also be used to store the child’s medical information, and initiate treatment though antibiotics and food supplements and further monitor the child’s response to treatment. All this information can be used by doctors during real-time decision making to aid in the treatment protocol. Thus, mNutrition is a preliminary step in building a technological solution to ease the implementation programs for the health workers and ensure an accurate timely diagnosis.
After a thorough research of existing diagnostic methods, I realized that there is an urgent need of a reliable screening tool that can quickly diagnose malnutrition with maximum accuracy. The solution should be easily accessible, especially in rural areas where the percentage of malnourished children is relatively higher. The real question was ‘How could I create a solution that can help identify the extent of malnutrition in a child while at the same time, be easily accessible, cost-efficient and provide maximum precision?’ Keeping these points in mind, I came up with a possible hypothesis which included a mobile application that could act as a screening tool for health workers, using preloaded data to help evaluate a child’s risk profile. The application would then also store the child’s records online so that they can be easily accessed when required.
Malnutrition & its diagnostic features.
According to the WHO, malnutrition refers to deficiencies, excesses, or imbalances in a person’s intake of energy and/or nutrients.
Malnutrition in children under 5 years is diagnosed by categorizing it into different levels. The highest level of risk is termed as SAM (Severe Acute Malnutrition). SAM can be identified if -
· Weight-for-length/height ratios are less than -3Std Deviation in WHO Charts
· MUAC (Mid-upper arm circumference) is less than 115 mm
· Oedema present in feet.
Children with SAM are at the highest risk of death, nine times more than a child with mild malnutrition
Initiatives taken by Government of India (GOI) to tackle child malnutrition.
The GOI has set up various nationwide and statewide initiatives to tackle malnutrition such as the ICDS (Integrated Child Development Scheme), Balwadi feeding program and the Mid-Day meal program.
The ICDS is the largest nutritional program and is currently operational in 1.34 million anganwadis across India. Their objective is to reduce malnutrition in kids below the age of 5 and provide proper healthcare education to mothers.(3) The health workers are the primary point of contact for parents and they conduct growth monitoring activities, counsel the parents and motivate them to participate in malnutrition management activities. Most of these health workers are local women who are not well-literate and poorly trained with limited skills. Each worker typically cover 150-180 children in their anganwadi cluster.
Existing methods of diagnosing SAM and keeping records.
Weight for height Z-score(WHZ) and MUAC are two independent anthropometric indicators used in the diagnosis of children with SAM. However as they are both measured using different methods, they do not always necessarily identify the same population of children as suffering from SAM. (4)
WHZ has been the preferred indicator for diagnosing malnutrition in primary health care centers. However, it has technical and practical limitations when on site. On the other hand, MUAC is a simple and low-cost method that can be practiced by a single person with minimum training. However, there could be a discrepancy in identification between these two methods. A child identified by one parameter need not necessarily be identified by the other and hence depending on which method amongst the two is used, may not be eligible for the necessary treatment.(5)
Currently, health workers adhere to tracking and site-monitoring systems which are paper-based. However this system is slow, tiresome and subject to human errors and misreporting. Errors can also occur in rounding off the values to the nearest corresponding measure. Even with the correct values for conducting the look-up, a frontline health worker might still look at the wrong column for grade assessment. Also, paper based records are easily prone to damage. In the case of a natural disaster, valuable records could be lost forever.
Therefore there is a need for a screening and record management tool for health workers that is easily implementable, requires minimal prior training and provides accurate results within a short span of time.
Approach- I started by collecting the latest WHO weight-to-height charts(4) and analysed them to check whether there was a pattern that I could use to develop a mathematical equation in order to determine the risk factor. As there was no fixed mathematical formula to determine the values using a given variable, I chose to develop a program that compares the weight-to-height ratios directly to the charts themselves. For this program, I used Eclipse Oxygen IDE (Integrated Development Architecture) by Oracle and the programming language used was Java. I mapped the entire WHO charts in a set of Integer arrays, to which I added a series of nested if-else ladders to sequentially determine the final value. The program currently assesses the child on account of his weight-to-height ratio as well as MUAC. If any one of them shows a sign of a more severe grade of malnutrition, then the risk factor is altered accordingly.
Proof Of Concept - For the Android App development, I used Android Studio. For the database, currently I have used SQLite to create a client-side local database in the beta version. However in the next application update/version rollout, I will import the local database to the cloud by using Google Firebase. Google Firebase uses NoSQL and is also cross-platform, meaning that it can be utilized by android, IOS and web apps. This is to make the solution scalable and accessible to a wider audience, in line with my nationwide rollout strategy.
Solution Design & Components - I started with the overall solution design with the Login Page, where the user will enter his/her username and password to gain access to the application. Upon gaining access, the user will be directed to a Home Page, where he can search the names of children who already have existing records in the database. On clicking any name, a dialog box appears containing details about the child - name, age, DOB, area/ ward where he stays etc. A floating Action Button will also prompt the user to add a new entry into the database. Upon selecting this button, the user will be guided to a series of Information pages where he can enter data such as the Child’s name, location, date of birth, gender, name of parents as well as weight, height, MUAC.
Upon submitting these details, the application sends this data to the program mentioned above. After comparing the values with the table, the program will then display the risk factor on the next page, which is the Results page. The Results page displays the risk factor as an integer as well as a word statement describing the risk factor i.e. “Normal” to “Severely Malnourished”.
I also chose to use colours such as red to denote “Severely Malnourished” as red denotes danger and using it will help the user understand the gravity of the situation.
We did a test rerun to check if mNutrition uses both the weight for height criteria as well as MUAC for detecting malnutrition. In all cases, height was kept at a constant of 60cm. The results of the test were as follows –
The results confirmed the logic built in the app that mNutrition uses both weight for height as well as MUAC to calculate the risk factor to ensure that no child escapes being diagnosed for SAM. Even if the readings on any one of these parameters exceed the normal values, the risk sign is being displayed. Users also have an option to modify/ edit the values if required, or just save the entry and create a historical record of the child. More research is needed on exactly which diagnostic method (Weight-to-height or MUAC) is more reliable and will produce better results so as of now, using both the criteria for detecting malnutrition seems like a more reliable option.
The WHO weight-for-height and MUAC tables are globally used anthropometric indicators to estimate SAM. So any kind of revolution of this particular method which makes the health workers job easier, is consequential. While the mobile application does not eradicate all possible types of human error, it reduces many of the pivotal human errors leading to incorrect diagnosis and thus, provides a more reliable diagnosis for malnutrition. In the long run, the success of the programs led by the government highly depends on successful implementation by the health workers. A technological innovation such as mNutrition being used as a mobile calculator, helps the children being screened correctly and reliably in a short period of time.
KEY CHALLENGES & FUTURE STEPS
The key challenge here is to develop a fully functional software solution with a holistic approach, addressing accurate diagnosis through a complex tabulation as well as guidance regarding the correct treatment protocol, right from the antibiotic intervention to the precise nutritional therapy through RUTFs (Ready to Use Therapeutic Foods) and other means. Since mNutrition further aims to be a health diagnostic and treatment prognostic application, the probability of inaccuracy has to be minimalistic. On the other hand, getting government bodies to implement this program nationwide in a bureaucratic country like ours in a given span of time would be another key challenge.
These challenges can be addressed by having a phased wise approach. Currently I have started with the first phase of developing a diagnostic solution, by digitizing the WHO 2006 weight-for-height/ length and the MUAC tables. After a pilot study I plan to progress further towards future phases encompassing a prognostic, educative and treatment module. I plan to develop a response triggered decision tree algorithm for the treatment protocol which will be accompanied by text, voice and pictures, for ease of the health workers. Counseling messages as well as information regarding the ready to use therapeutic foods sachets and routine medications can also be displayed on the app. This entire data recording the child’s information and follow up will be uploaded on the ‘cloud’ providing live and accurate data for decision making and management.
Continuous innovations should be made to make the implementation program easy and uncomplicated for the health workers and mNutrition is one such effort of doing the same.
My name is Ayush Gharat and I am a grade 9 student studying in Head Start Educational Academy in Bangalore, India. I am a certified Associate Android Developer, and have learned Java, XML as well as a fair amount of python.
My love for technology started when I first came across the movie ‘Ironman’. J.A.R.V.I.S, the AI that Tony Stark uses caught my attention and became an obsession of mine. Since then, I have developed a deep interest for technology, from phones and laptops to Artificial Intelligence and Quantum Computers. I also have a strong passion to eradicate the problems that our world faces today, so that the future is brighter and better than ever.
My inspiration is Elon Musk, the man who chooses to dream big. His life stories, right from leaving his country to pursue education, to funding SpaceX and Tesla even when everyone thought they were failed companies, really inspired me to work hard no matter the cost.
I am currently planning to seek further education in either USA, Canada or UK. I am looking to choose a computer-stream program, possibly at Machine Learning or Artificial Intelligence. My dream career would be as an entrepreneur for a tech-based startup.
If I won the Google Science Fair, it would mean the world to me as it would give me a platform to promote this idea and turn it into a successful venture/product that can change people’s lives for the better.
References and Bibliography: