UNICEF Innovation Fund Graduate: Avyantra

Machine learning-powered prediction platform to facilitate the early diagnosis of neonatal sepsis in India

Hyma Goparaju, Co-Founder, Avyantra Health Technologies
Machine learning-powered prediction platform to facilitate the early diagnosis of neonatal sepsis in India
30 March 2020

The UNICEF Innovation Fund is proud to see portfolio member, Avyantra, graduate. They’ve come a long way – from numerous product iterations to deep diving into understanding their ecosystem better, strengthening their business model, and gearing up to take their solution to market. They’re now ready to collaborate at a larger scale – as they find new pathways to work with partners, investors, and the open source community. 

Over the past 12 months, Avyantra has successfully moved from the proof of concept stage to a fully accomplished product with support from the UNICEF Innovation fund. Our team has developed PreSco, a cloud-based application to facilitate early diagnosis of neonatal sepsis through machine learning methods.

The first 28 days of a child's life are the most vulnerable. In India, neonatal deaths are comparably higher with a neonatal mortality rate of 23 deaths per 1000 live births. Neonatal infections account for nearly 33 percent of newborn deaths in India, and while preventing deaths such as these is possible through provision of quality healthcare, many mothers and children in rural areas often fall victim due to the lack of  adequate health infrastructure. Our platform aims to address this gap by aiding early diagnosis of neonatal sepsis through artificial intelligence. 


Avyantra initiated the prototyping process after conducting primary research on issues of diagnosis of neonatal sepsis. During this process, we identified data to be collected in a specific format from hospitals. Non-invasive and observable data points of interest, like mother parameters – blood pressure, GBS Infections; baby parameters - baby appearance, distension, skin color, urine output; invasive parameters - blood CRP, WBC Count, immature to total neutrophil ratio etc., were among the many other parameters were reviewed with neonatal specialists, and the format was developed based on their feedback along with key findings from our literature review.

When we began testing the application in public health hospitals, we found that only a few variables/parameters are used due to various reasons such as budgetary and resource constraints, non-availability of equipment for blood tests and culture test equipment. 

Snapshot of PreSco’s Risk Score Screen
Snapshot of PreSco’s Risk Score Screen


Due to the above constraints, doctors resort to treating infants through the administration of antibiotics, regardless of the severity of the symptoms. We have therefore developed a new model with few non-invasive parameters and critical tests (CRP test, WBC test, etc.) which are available at primary hospital level so that the application will be useful for semi-urban and rural areas. The noninvasive parameters for babies include data points such as birth weight, frequency of stools and gestational age; for mothers, these parameters include mother BMI and blood pressure. The score generated by the platform provides a risk score through which a doctor would be able to make a risk assessment of the baby’s condition (whether the baby is moving towards sepsis or not).

During the time the application was being developed, we collected data in excel files and later data collection was done through the open source application. We then commenced the development of the predictive algorithm by analyzing the data collected. Through inputting different neonate data points, the platform generates a predicate score that doctors can use in their diagnosis of neonatal sepsis. The platform provides three levels of risks – Low, Medium and High — through which a doctor can assess a baby’s probability for an onset of sepsis.

We iterated upon the predictive model and fine-tuned the algorithm to generate the predictive score. We then compared the predictive scores generated via the PresCo platform with the blood culture results obtained from laboratories. With culture tests being the gold standard, we compared our score with the culture tests results that came in after about 2-3 days. We achieved an accuracy ranging from 85% to 90%.

The application was then tested by multiple health care professionals for accessibility, user-friendliness and accuracy of the score. Uuser requirements and methods used for identifying sepsis are quite different between urban and rural hospitals, which was indeed a huge learning during our prototype development.

Using open source has enabled us to build our prototype faster through easily available resources, all while keeping the cost of development low.

Open Source

We are hopeful that the open source model will lead us to ease integration and interoperability during the phase of commercialization, while meeting the prime objective of keeping the platform’s cost at an affordable range.


Since our platform is novel, the main challenge for Avyantra lies in getting the platform extensively tested and validated among healthcare professionals. Establishing consistency and achieving high performance metrics are some of the challenges we foresee. For this, we need to extensively and exclusively work with a clear plan for testing and validating with a sample of healthcare volunteers and perform real-time testing of the cloud-based platform. We believe that thorough validation will help us gain the confidence of the healthcare professionals and of the healthcare funding community, leading us to take the platform seamlessly to a commercialization phase.

Place: Shrawasti District Hospital; District: Shrawasti; State: Uttar Pradesh; Country: India

What's Next?

We are looking forward to extensively testing and validating our platform with rural health volunteers in India (ASHAs) as well as collaborating with relevant government agencies. Once we expect the government to be the largest consumer of our platform, collaboration is essential for us to reach rural pockets that are in need of better primary healthcare infrastructure. The next big goal would be to accomplish high performance metrics and move the platform to the commercialization stage.

Working with the UNICEF Innovation Fund

The UNICEF Innovation Fund has played an extremely crucial role in our product development. Since our solution is based on machine learning and artificial intelligence technologies that have not been tried and tested for commercial application of neonatal sepsis elsewhere, it was quite hard to obtain funding for prototype development. The UNICEF Innovation Fund helped us get over the hurdle. In addition to the funding support, the Innovation Fund’s business mentoring sessions, both online as well as during the workshop, have helped us tremendously in fine tuning our business model. 

Along this journey, our work has gotten national recognition as we were among the 25 social start-ups in the Action for India Initiative in November 2019.

In February 2020, we received a grant of about USD 20,000 by ‘Parivarthan’, an initiative to promote start-ups by HDFC, a leading private bank in India. We were among the 20 companies selected for funding.

Overall, the team at Unicef Innovation Fund has been extremely supporting, encouraging and understanding. Learnings during these 12 months have been immense and we hope to take them forward and implement them during all stages of our company’s growth.


About the UNICEF Innovation Fund:

UNICEF’s Innovation Fund invests up to $100k in early stage, open-source, emerging technology digital public goods with the potential to impact children on a global scale. It also provides product and technology assistance, support with business growth, access to a network of experts and partners to allow for scale and growth. The investments can go either to UNICEF Country Offices or to private sector companies in UNICEF programme countries.