MaternalShield

Inspiration:

According to the WHO (2023) states that in 185 countries there were 287,000 maternal deaths in 2020. This amounts to about 800 maternal deaths daily, or one death every 2 minutes. In addition, Sub-Saharan Africa alone accounts for 70% of global maternal deaths, and 95% of these deaths are in low and lower middle-income income countries. As someone who is from Sub-Saharan Africa and with a pregnant partner, I was inspired to look into creating an app that mitigates this. My team bought into the idea and fully supports of it.

What it does:

To create a comprehensive and accessible platform that leverages technology for early detection, prevention, and mitigation of maternal health issues, hence reducing global maternal mortality rates. It puts pregnant people into categories such as "Low risk", "Middle risk", or "High risk", by using factors like age, body temperature, heart-rate etc, then determines if they need urgent medical attention.

How we built it:

What sets MaternalShield apart is the integration of user-friendly application capabilities, real-time symptom monitoring, and data-driven decision support. We leverage artificial intelligence (AI), machine learning (ML), and data analysis to empower users with timely and personalised recommendations. We used Python programming language, as it was efficient in finding the best classification models, importing the dataset, and converting the string values, numerical values etc. There will be ongoing fine-tuning to enhance the app.

Challenges we ran into:

It was our first time working on IBM LinuxOne Cloud, hence it was a bit sketchy to set up and took a bit of our time. The datasets we used were hard to source as many of the more comprehensive ones are copyrighted; which we cannot use. Another challenge was that only three out of six members of the team worked on the project, this put a huge strain on our capabilities, but we did the best we could.

Accomplishments that we're proud of:

This is my first Datathon and real experience into using this technology, as University undergraduates, we are proud of participating, and attempting to solve a problem which is personal to us. Our accuracy on execution was a maximum of about 85%, we intend to improve this and get closer to 100%.

What we learned:

We learned the importance of teamwork, how we can think logically, harness the power of technology to make a real change. We also learnt that solutions to many problems are already out there, but in bits and pieces which needs to be incorporated and executed appropriately. In addition, we learnt about the usefulness and capabilities of IBM Z, we will explore it further and believe it will provide more opportunities.

What's next for MaternalShield:

• Partner with existing maternal health organisations to ensure the platform meets the needs of the target audience and is effectively disseminated. Collaboration with these organisations will facilitate feedback gathering on the platform design, content, and features. Additionally, outreach and dissemination strategies will be developed in collaboration with these partners to maximize reach. • Development of mobile applications, a website, and a text messaging service for symptom reporting and analysis. • Data collection, integration, AI model development, and technology implementation. • Develop a plan for evaluating the platform's impact on maternal mortality and other maternal health outcomes. This includes identifying relevant metrics, data collection methods, analysis methods, a timeline for evaluation, and a process for sharing the results with stakeholders. • Incorporate feedback from end users and stakeholders throughout the development and implementation process. Feedback will be gathered through user surveys, focus groups, interviews, and a feedback forum on the platform for users.