Symptom Analysis to Predict Heart Failure

The Client

The client – JROP is a research institute of echocardiology based in New Delhi, India. internationally renowned echo-cardiology expert based out of Delhi. In addition to research work, the institute trains about 2,000 students every year in echocardiography. 

The Problem

The client wanted to study the correlation between common heart conditions in patients with diastolic heart failure.  

Key Challenges

  1. The data had to be cleaned and standardized for analysis 
  2. Domain expertise was required to understand medical terminology and the insights obtained from the analysis. 

How ElementalAI Helped 

The team designed the pipeline to create clusters that would best describe the behavior of the body in terms of hormonal imbalance or prevailing heart conditions in those clusters. The model improves with incoming data i.e. reinforced learning.  

Currently the model creates the clusters based on demographics and suggests the most common ailments that patient suffers when they belong to those clusters.  

This can potentially save the medical practitioners time to diagnose the patient for DHF and start early treatment. 

Technologies Used 

The analysis was primarily created and published in Excel