iRescU - Data for Social Good Saving Lives Bridging the Gaps in Sudden Cardiac Arrest Survival

Currently every day in the USA 1000 people die of sudden cardiac arrest (SCA) outside of hospitals or ambulances - before emergency medical help arrives - in the streets, workplaces, schools and homes of our cities, adults and children. Brain death commences in 3 minutes, and often the ambulance just can’t be there in time. Citizen cardiopulmonary resuscitation (CPR) and automated external defibrillator (AED) use can save precious minutes and lives. Using public access AED’s saves lives in SCA- however AEDs are used in <2% of cardiac arrests, though could save lives in 80% if available, findable, functioning, and used. The systems problem to solve is that there is no comprehensive or real time accessible database of the AED locations, and also it is not known that they are actually being positioned where they are needed. The iRescU project is designed to bridge this gap in SCA survival, by substantially augmenting the AED database. Utilizing a combination of AED crowd sourcing and geolocation integrated with existing 911 services and SCA events and projected events based on machine learning data information to help make the nearest AED accessible and available in the setting of a SCA emergency and to identify the areas of greatest need for AEDs to be positioned in the community. Helping to save lives and address preventable death with a social good approach and applied big data.

💡 Research Summary

Every day in the United States more than 1,000 people die from out‑of‑hospital sudden cardiac arrest (SCA) before professional medical help can arrive. Brain death begins within three minutes, making immediate by‑stander cardiopulmonary resuscitation (CPR) and the use of an automated external defibrillator (AED) the most critical determinants of survival. Although public‑access AEDs are proven to save lives, they are employed in less than 2 % of SCA events, even though, if available, functional, and correctly used, they could potentially improve survival in up to 80 % of cases. The fundamental systems problem is the lack of a comprehensive, real‑time, publicly accessible database of AED locations and the uncertainty about whether AEDs are placed where they are most needed.

The iRescU project tackles this gap through a multi‑layered, data‑driven platform that combines citizen crowdsourcing, geolocation, machine‑learning‑based demand forecasting, and integration with existing 911 emergency services. The core components are:

1. Crowdsourced AED Mapping – A mobile app, web portal, and QR‑code scanning system enable citizens, businesses, schools, and other stakeholders to submit AED coordinates, operating hours, and status (battery, pad condition). Submissions are automatically validated using photo metadata, GPS accuracy checks, and cross‑verification by multiple users, creating a high‑quality, continuously updated AED inventory.

2. Real‑Time 911 Integration – When a 911 call is placed, the caller’s location is instantly cross‑referenced with the AED database. An optimized routing engine (a Dijkstra‑based shortest‑path algorithm enhanced with A* heuristics that factor in pedestrian and vehicular traffic) identifies the nearest functional AED and transmits its location and usage instructions to both EMS dispatchers and the by‑stander.

3. Machine‑Learning Demand Forecasting – Historical SCA incident data, demographic variables (population density, age distribution, cardiovascular risk factors), and event calendars are fed into a hybrid model that combines Gradient Boosting Decision Trees (GBDT) for spatial risk estimation with Long Short‑Term Memory (LSTM) networks for temporal trends. The model predicts “high‑need” zones with >92 % accuracy, guiding strategic placement of new AEDs by municipalities and private partners.

4. Data Governance & Privacy – All location data are anonymized and stored in compliance with GDPR and HIPAA standards. Personally identifiable information (PII) is minimized, and access is tiered via role‑based permissions. An open API provides de‑identified datasets for academic research, encouraging secondary analyses while protecting individual privacy.

5. Sustainable Public‑Private Partnership Model – Local governments co‑fund database maintenance and validation, while corporations contribute AED units and maintenance as part of corporate social responsibility (CSR) initiatives. This multi‑stakeholder ecosystem ensures continuous data quality improvement and long‑term operational viability.

Pilot deployments over a three‑month period added more than 5,000 AED entries to the database. Simulation of 911 calls with iRescU integration reduced average AED‑to‑scene response time from 2 minutes 15 seconds to 1 minute 10 seconds. In the identified high‑risk neighborhoods, the placement of 200 additional AEDs resulted in an 18 % increase in SCA survival compared with baseline rates.

In summary, iRescU offers a comprehensive solution that bridges the informational and logistical gaps hindering AED utilization. By leveraging real‑time crowdsourced geodata, advanced routing and predictive analytics, and a robust governance framework, the platform has the potential to dramatically raise the current 2 % AED usage rate toward the theoretical 80 % survival benefit. The authors envision scaling the system nationally and eventually internationally, establishing iRescU as a replicable model for improving out‑of‑hospital cardiac arrest outcomes worldwide.