Predictive Modeling in Health Care

Project Summary

Funding Source:
PCCI | Predictive Modeling and Advanced Clinical Research
Funding Amount:
Principal Investigator:
Dr. Sudha Ram

Exploring the use of Twitter data to predict health outcomes, including utilization and disease detection.

This project seeks to prospectively collect and analyze three years of Twitter data from all over the world. Information from Twitter is timely, location-specific, and can provide information on events and unmet needs that are otherwise not reported.

We propose using advanced analytic techniques, including machine learning, data mining, statistics, and social network analysis to identify risk factors affecting utilization of the certain emergency departments. The analysis will identify indicators for certain environmental, social, and public health factors that may predict disease incidence, as well as build a predictive model for emergency room admissions using a combination of sensor, social media, and electronic medical records.

This will create a novel method for medical organizations to better predict, plan for, and meet the needs of their community.



Twitter analysis graphic courtesy Dr. Sudha Ram.