EPICS
SERVICE

Fall 2024 (FSE 104) & Spring 2025 (FSE 404) — EPICS at ASU

Over two semesters in ASU's EPICS program, I worked on the Automated Medical Resource Allocation System (AMRAS), a desktop application designed to efficiently allocate Extracorporeal Membrane Oxygenation (ECMO) machines to patients throughout Arizona. Our community partner was Ms. Sumedha Attani, a medical student at Mayo Clinic Alix School of Medicine, who provided expertise on the healthcare struggles faced by rural and underserved areas.

The problem is significant: there are currently only 27 registered ECMO machines statewide, and the existing system heavily favors urban areas. Patients in rural communities face life-threatening delays in accessing these critical devices. Our team of seven built a comprehensive web application that considers factors such as proximity to hospitals, criticality of conditions, age, transportation needs, and urgency to match patients with available ECMO machines.

The system features a priority queue algorithm that automatically ranks patients based on severity, age, location, and special conditions (pediatric, pregnancy, first-responder status). We built inter-hospital communication features so hospitals can coordinate around outlying conditions, a transport arrangement system, and a real-time map interface displaying all ECMO machines and patient locations across the state. The tech stack includes Next.js, ShadCN UI, Clerk for authentication, Supabase for the database, and Sentry for monitoring, all deployed on Vercel.

During my two semesters, I contributed to the allocation algorithm development, the user interface, and critically, the security enhancements including patient data encryption for HIPAA compliance and EMR integration investigation. Our verification testing confirmed the system achieves sub-3-second load times, sub-5-second map display, and sub-2-second automatic sorting, all passing our design requirements.

TECHNICAL
DETAILS

RELATION
TO MY
THEME

This project connects to my GCSP theme of Security through two dimensions. First, data security: handling patient health information required HIPAA compliance, encrypted storage, and secure authentication.

Second, health security: the existing ECMO allocation system disproportionately serves urban areas, placing rural Arizona patients at a significant disadvantage. Our algorithm addresses this inequity by incorporating distance, severity, and special conditions (pediatric, pregnancy, first-responder status) rather than defaulting to proximity to major hospitals. The PESTLE analysis conducted by our team clarified the legal and ethical implications: when an algorithm determines priority for life-saving devices, the consequences of design decisions are immediate and consequential.

EPICS distinguished itself from other GCSP experiences through the immediacy of its impact. The system serves patients who require ECMO machines for survival, and our community partner at Mayo Clinic depended on a reliable, well-tested product. This accountability elevated the rigor of our development process: more thorough testing, more careful consideration of edge cases, and greater attention to failure modes.

The project also integrated competencies developed across my GCSP experience. The PESTLE analytical framework from FSE 150 structured our stakeholder analysis. The data security principles aligned with my AI safety research through FURI. EPICS reinforced that social consciousness is not supplementary to engineering — it is the foundation that determines whether engineering serves its intended purpose.

VALUE

DESIGN
REVIEW

Below is the slide deck from our EPICS design review, presenting the AMRAS system architecture, stakeholder analysis, and verification testing results.

contact@aryankeluskar.com
aryankeluskar.com