The client is a mid-size US-based healthcare firm that offers a full range of primary, secondary, and tertiary healthcare services and is dedicated to providing affordable and quality patient care while ensuring efficient access to specialty and hospital services whenever the need arises. The healthcare client intended to enhance their patient care in a non-urban community. Sparity developed an AI-powered health monitoring solution with the aid of IOT devices.

ClientHealthcareServicesData Science, AI, MLYear 2022

Key Challenges

  • Across the healthcare organization, it is challenging to consistently identify patients at risk, and the client needed a solution to specifically handle rural medical emergencies and support them with high-quality care ​​​​
  • They needed a health monitoring model that collects, processes, and analyzes clinical data from IoT devices


  • Sparity designed the AI-powered preventive diagnosis/risk assessment health monitoring model in combination with the IoT devices that track the health status of patients from the recorded vitals of each patient, such as blood pressure, blood glucose, pulse, respiration, temperature, weight, height, bowel movement, fluid intake & output, and more
  • Employed Amazon EC2 on-demand compute infrastructure for optimized performance ​
  • Adopted AWS IoT Core for connecting Internet of Things (IoT) devices to monitor, manage, and scale their device fleets easily
  • The prediction model analyzes the recorded data to detect patterns or any adverse changes through a supervised learning approach
  • AI-powered prediction model accurately analyzes PQRST waves to detect heart-related emergency incidents (Ex: Angina vs Myocardial Infarction)
  • The implementation of these models leads to intelligent, informed, and smarter patient experiences ​
  • Data from Stethoscope, Blood Glucose, Pulse & Oxygen Saturation (SPO2), Temperature, Blood Pressure, ECG (3/12 Lead) was analyzed to detect patterns in the cardiovascular risk factors through a supervised learning approach
  • PQRST Curves were analyzed through OpenCV libraries to detect abnormalities in ECG ​
  • The model also triggered automated alerts to notify caregivers about the negative changes ​


  • The AI model reduced false positive results by 25%
  • Reduced workload by more than 35% ​
  • Implemented structured reporting tools for gathering near-real-time feedback
  • Improved Preventive diagnosis
  • Effectively handled Risk/Emergency situations
  • Automated processes to reduce dependency on caregivers
  • Rural medical practitioners are empowered with AI, which assists them in deciding the best action to handle any emergency
  • The current system in place is a “Predictive solution” helping thousands of rural patients survive medical conditions