Arogya Saathi: A Unified Healthcare Ecosystem for Rural India
Description
In rural India, healthcare delivery is often hampered by systemic fragmentation. Patients frequently travel long distances only to find doctors unavailable, medical records are often lost or siloed in physical files, and rural health centers (PHCs) struggle with inventory management of critical medicines. For the government, a lack of real-time data makes it impossible to allocate resources effectively during health crises. Arogya Saathi was conceived as a digital bridge to close these gaps, specifically tailored for the linguistic and infrastructure-challenged environments of states like Punjab.
As the Lead Full-Stack Developer, I designed a multi-tenant, high-concurrency architecture capable of serving millions. The system is built on Next.js 15 for the frontend, utilizing the App Router to optimize performance and SEO—critical for low-bandwidth rural connections. The backend, powered by Node.js and TypeScript, follows a modular service-oriented architecture to ensure maintainability.
One of the core technical challenges was handling "registration bursts"—moments when thousands of citizens might register simultaneously during a local health drive. To solve this, I implemented distributed locking using Redis. This prevented race conditions in the database when multiple workers tried to update shared resources (like hospital bed availability or medicine stock) simultaneously. For the data layer, we used PostgreSQL for its robust relational integrity, which is vital for medical records where data consistency is non-negotiable.
The platform provides a unified experience across three distinct user roles: Patients (can book appointments, view centralized medical history, multi-language support), ASHA/Staff Workers (mobile-optimized portal for field updates), and Doctors (streamlined dashboard with holistic patient history). Arogya Saathi was vetted by SIH judges and government officials for its potential to reduce patient wait times by an estimated 40%.
Technologies Used
How I Did It
I focused on high-concurrency handling using Redis for distributed locking and optimized the frontend for low-bandwidth environments using Next.js 15's latest performance features.
Media
Image/Video Placeholder