ASHA Sarthi

Inspiration

Our journey into the world of ASHA workers began unexpectedly, stumbling upon a newspaper article detailing their fight for fair pay in Tamil Nadu. Curiosity ignited, and we asked, "Who are these ASHA Workers?" A simple Google search revealed a world of incredible dedication and impact. Even the WHO recognised and commended their efforts in providing primary healthcare in India. Their monumental contributions often hid behind the "Frontline workers" label during the COVID-19 crisis.

These unsung heroes are the pillars of our healthcare system, starting from the grassroots level (rural areas) in primary healthcare centres. Their role in early disease detection and hospital referrals is vital, alleviating pressure on an already strained tertiary healthcare system. Their care extends to expectant mothers, ensuring a smooth journey through pregnancy and post-delivery.

ASHA workers uniquely bridge the gap between resources and the people, performing follow-ups that keep the chain of care unbroken. However, their remarkable contributions didn't match their rewards. This disparity prompted us to delve deeper and question the how and why of it all, sparking the inception of our project.

Our research encompassed various sources – media, literature, government documents, and guidelines from allied workers. We aimed to understand their pain points, daily tasks, and challenges. This journey showed us numerous avenues where technology could lend a helping hand, streamlining their work and improving their lives.

Yet, there lingered a doubt – would they adopt our solution? Would they willingly embrace a tool that might increase their workload? The answer lay in a paradigm shift. We decided that for them to embrace the change, we needed to show them the immediate benefits they could reap.

So, armed with insight and empathy, we embarked on a mission to empower ASHA workers, not just with technology, but with a solution that truly understood and catered to their needs.

What it does

This app is born from understanding their challenges and needs, aimed at enhancing their efficiency and impact.

The app boasts three core features:

AI-Powered Chatbot: A virtual assistant that guides ASHA workers through their daily tasks and inquiries. It's designed for convenience, allowing them to interact in their preferred language by speaking or typing using generative AI and Google Cloud’s WaveNet based lifelike speech synthesis models..

Record Management for Rural Patients: Keeping track of patient information and medical activities becomes effortless. It helps ASHA workers manage prescriptions, referrals, and follow-ups efficiently, enabling them to collaborate seamlessly with doctors.

Assistive AI: A revolutionary tool that adapts to their workflow. By using voice commands, ASHA workers can input patient data effortlessly, eliminating the need for manual typing and transitioning towards a more streamlined, digital approach.

Additional tools back these features:

Incentive Tracker: Empowering ASHA workers to maximise their earnings by ensuring they claim incentives rightfully earned.
Mapped Patient Cases: Geotagged patient cases for easy follow-ups, addressing, and visualisation, utilising the MapBox API.
Notice Boards: Providing real-time information on doctor availability and resource schedules.
Training Guides: Equipping ASHA workers with essential training material, ensuring they always feel supported.

How we built it

Every facet of the app is meticulously aligned with ASHA workers' needs. The design ensures inclusivity, offering language choices and using intuitive icons. It's optimised for rural settings, with a lightweight interface requiring minimal internet connectivity.

Asha Sarthi is a result of carefully considering the Service Blueprint, ensuring seamless integration into their workflow. The app is more than a tool—it's a companion empowering ASHA workers to transform healthcare delivery, one digital step at a time.

For the technical aspect, we created the interface with React, its common helper libraries like React-Router, Python w/ flask as a server-side API b/w the various AI tools and the frontend, OpenAI’s facilities like Whisper AI, and GPT-Turbo, and Google Cloud’s WaveNet based lifelike speech synthesis models.

Challenges we ran into

Roleplaying the AI assistant as an ASHA worker: Having found the ASHA training materials online, we decided it would be best if our AI assistant was specifically trained for it. We finetuned OpenAI’s GPT on the PDF documents of ASHA training material, with the help of their finetuning API. We also initialized conversations with the AI assistant by giving it system prompts to roleplay as an ASHA worker.

Multilingual chat assistant: Assembling an end-to-end multilingual chat assistant that can play the role of an ASHA worker’s assistant. This utlized OpenAI’s whisper AI for multilingual speech -> english transcription.

Creating visuals and illustrations: We designed and digitally painted our own illustrations, characters and animations. This was very important for the application to feel alive, which improves user experience for our target audience.

Design for Accessibility When we generally talk about accessibility, we try to think about the blinds, the ones with some physical impairment, but working on this app was different here. We were dealing with tech for rural area. A population which had access to phones, but had no means to use them. They were able to understand the basics of it, but still required assistance. At the same time, being a ASHA Worker, they would often face temporary accessibility issues, since they worked on ground. Designing of such platform was a challenge itself. Understanding the context was also challenging. Also, Designing could not be super fancy, unlike the apps used in the urban areas, these apps had to be super simple.

Accomplishments that we're proud of

It was a remarkable accomplishment to set out on a 24-hour trek, come up with an idea, and unite our team around it. The ability to make informed decisions paved the road for our success. Finding a solution that appealed to everyone in our diverse workforce was difficult. However, after learning about the world of ASHA Workers, our united resolve to support their cause hardened.

This project carried a very personal meaning for one of our team members, whose entire family is actively involved in Uttarakhand's medical community. It seemed like there was a remarkable purpose alignment.

The seminar's introduction to MapBox sparked our interest in technology in an exciting way. Despite the fact that Open AI and Google offered a wide range of options, we hadn't yet fully utilised their potential as tools for creativity. Even in compared to our involvement in prior hackathons, the sheer scope of what we produced was unmatched. We dared to explore technologies that were novel to both us and the rest of the world, venturing into uncharted waters.

The trip was thrilling, and it gave us more self-assurance to meet any obstacle that comes our way.

In terms of technology as well, we just learnt MapBox in the seminar, and it was super exciting to use something new. Open AI and Google have had a lot of solutions as well, but we never got a chance to try them as a tool to create something. Creating something this big was something new as well, which we had not done before, not even in older hackathons we participated in. This event improved our capacity to work as a team while also broadening our technological horizons. The satisfaction we received from completing this task gave us the confidence to move forward, ready to meet any obstacle life throws at us.

What we learned

This year at RedBrick Hacks, we explored our first AI based project. It was a very new and interesting experience for us. We have expanded our horizons and are now capable of opting for AI based solutions where it fits well. Another aspect that we challenged ourselves to work on was designing AI applications for people in rural areas in India. For this, we leaned towards using multilingual models which are capable of working with languages like Hindi, Gujarati, Marathi, and other Indian languages.

In all, these 24 hours have been an enriching experience for our learning and growth in aspects of design thinking and technical ability.

What's next for ASHA Sarthi

The conclusions drawn from thorough study papers and the complaints aired in news pieces are directly reflected in the current work. It's important to understand that healthcare is a dynamic, state-specific issue that is subject to constant changes. This calls for interaction with those who are actually involved locally. Our approach comprises building relationships with ASHA employees, working closely with them to understand their specific difficulties, and customising the product to meet state regulations. This focused strategy ensures a more effective solution because incentive systems and operational dynamics can differ.

We will soon be working with a select group of ASHA employees to jointly develop a solution that is tailored to their needs. By doing this, we create a smaller but incredibly informative user base that is consistent with the learnings from our insightful fireside talk with Mr. Bikchandani. We plan to widen our audience now that we have a prototype and some early feedback. In order to increase the impact of our project, we will submit our improved solution to well-known businesses operating in the public sphere. We will then make use of their knowledge and resources. As we go from research and development to practical application, addressing the shifting landscape of healthcare with accuracy and empathy, the journey ahead holds enormous potential for transformation.