IoT & AI Driven Patient Criticality Predictor Program in Hospitals

OBJECTIVES: Arterial blood gas (ABG) analysis can help in predicting mortality risk and managing ventilatory settings for better outcomes

An IoT-based analytical system was developed by EHS to predict the condition of the patient, stratify the risk, and help intervene timely.

METHODS: A predictive model-based program that is implemented in EHS hospital in UAE that accurately identifies patients at risk of mortality using real-time data feeds from a blood gas analyzer. The model leverages the latest technology to continuously monitor patient vitals & laboratory data from the blood gas analyzer, analyze the data using advanced algorithms & machine learning techniques, & provide clinicians with timely alerts & recommendations for appropriate medical interventions.

The model is the first in the region to be implemented in an actual hospital setting, & it led to predictions about patients who require critical care & timely interventions, avoiding unnecessary admissions to the CCU

In this project, a method based on artificial neural networks and gradient boosting algorithm with the aim of estimating and predicting the criticality indicator of the patient. Below are the details of the data set used Nov 2022 to Sep 2023 – Qassimi Hospital

Sample Size of 1,400+ patients

80+ Features – Demographics, Medical history, Vital Signs, comorbidities with rich Clinical data includes Blood Gas Analyzer data

RESULTS:

1. Highly accurate Model built on easily collectible features
2. Improve utilization rates
3. Targeted overbooking improves revenue and patient satisfaction

CONCLUSIONS: The project was presented in Arab Health 2024 and was very well received by healthcare community.