Explainable Early Patient Deterioration

See it before it becomes an emergency.

EEPD is a wearable ward monitor with edge AI. It watches every patient's vitals around the clock and tells nurses, in plain words, why a patient's risk is rising.

Hospital sign in → See how it works

24 × 7continuous vitals On-site AIworks without internet Every alertcomes with a reason

Scroll

The problem

On general wards, deterioration is noticed too late.

Late intervention means more ICU admissions, higher mortality and a heavier load on hospitals. Manual checks every few hours cannot keep pace with today's wards.

Illustrative example · one patient, one morning

2 h 20 min sooner

SpO₂ Heart rate

08:00
Nurse round 09:40
EEPD alert 12:00
Next round

Between rounds, SpO₂ slides from 98% to 90% while heart rate climbs. EEPD flags the trend at 09:40. On manual checks alone, the next look is at noon.

Nurses can't watch every bed

One nurse covers many patients. Between rounds, a slow decline goes unseen.

ICU monitors stop at the ICU

Existing systems are expensive and need heavy infrastructure, so they stay in critical care.

Early signs are quiet

The first changes are small and spread across several vitals. Each looks harmless on its own.

Falling SpO₂ Abnormal ECG Rising temperature Reduced movement

How it works

From the patient's wrist to the nurse's screen in seconds.

EEPD watch-style wearable on a patient's wrist, wired to the bedside dock

STEP 1

Wrist band senses

A light band on the hand reads ECG, SpO₂, temperature and movement. No battery to charge.

MAX86150 ECG/SpO₂ · MAX30205 temp · BMI270 motion

Bedside dock with a glass display showing live vitals

STEP 2

Bedside dock powers and shows

A cable links the band to the dock, which powers it, shows live vitals on a 5-inch screen and sounds the alarm, even if the network drops.

ESP32-S3 · 5" IPS display · buzzer

STEP 3

Sent over ward Wi-Fi

An ESP32 inside the dock streams readings over the hospital's own network.

Wi-Fi · 250 Hz ECG · vitals every reading

STEP 4

Edge AI scores risk

A Jetson unit on each floor runs the model. Data never has to leave the hospital.

NVIDIA Jetson Orin Nano · works offline

Nurse checking the EEPD ward view on a tablet

STEP 5

Care team gets the why

Nurses and doctors see an alert that names the vitals that changed, so they know where to look first.

Any phone or tablet · notes · acknowledge & review

The model

An alert you can read, not a black box.

EEPD looks at how each patient's vitals move over time, not single readings. It estimates the chance of deterioration in the next 12 hours, then lists the changes behind that number.

12 hprediction window

0.88ROC-AUC on held-out test patients

8 in 10deteriorations flagged at the alert threshold

294kobservations used for training

heart rateSpO₂temperature3-hour change4-hour average4-hour variability

Research prototype. Trained and tested on a public synthetic hospital dataset, not yet clinically validated. Clinical trials come next.

Bed 3P003Escalate

HR108

SpO₂91

TEMP38.4

Deterioration risk91%

SpO₂ fell 5% over 3 hours
Heart rate rose 22 bpm over 3 hours
Temperature 38.4 °C, fever

Where it helps

Healthcare beyond ICU walls.

ICU-grade attention for the beds that have never had it.

Small & medium hospitalsPrimary health centresRural health centres Old age homesAssisted livingAmbulancesHome care

Roadmap

Where we are, and what's next.

Nov 2026 · now
Working prototype
A wrist band and dock reading ECG, SpO₂ and temperature, a live ward portal, and the first risk model.
Dec 2026
Clinical-grade hardware
Medical sensors, a 3D-printed wrist band and dock, and fabricated boards.
2027
Five-bed edge system
Five wrist bands and bedside docks, connected to a Jetson Orin Nano.
Next
Hospital pilots
Pilot wards, clinical validation, and one EEPD portal per hospital.

The team

Built by FrameMind AI Studio.

Chinmay R M

CEO · CTO · Founder

Vision, product strategy, AI architecture and the technology stack.

Poorvi I H

CMO · Co-founder

Brand, go-to-market, and bringing hospitals on board.

Vaishnavi B

COO · Co-founder

Operations, pilot hospital deployment and scaling.

Dr. Bharati M. Reshmi

Mentor · AI

Guides the model's accuracy, robustness and scalability.

Working with

Backed and built with partners.

The EEPD team with their mentor at the NAIN 2.0 kickoff, IIIT Dharwad — **NAIN 2.0 kickoff · 2026–27**The team with our mentor at IIIT Dharwad.

Funding

Funded by NAIN 2.0

A ₹4.5 lakh grant from Karnataka's New Age Innovation Network (2026–27), with K-tech and Startup Karnataka.

Fabrication

Hardware partner, Hyderabad

Board fabrication, enclosures and assembly of the wearable and dock.

Pilots

Your hospital?

We are looking for wards to pilot EEPD. Talk to us about a trial.

Already an EEPD hospital?

Doctors, nurses and administrators sign in to their ward from any device on the hospital network.