Patients' disease risk predictive modeling using MIMIC data

Dhanjeet Singh, Vishal Kumar, Robin G. Qiu

Research output: Contribution to journalConference articlepeer-review

Abstract

The latest MIMC III (Medical Information Mart for Intensive Care III) database has rich information on over 58k patient's medical histories for over 11 years. Based on MIMC III database, this paper presents a study of those patient's primary disease risk prediction modeling, which focuses on assessing future disease risks for an individual who is ready for discharge. We explore a framework that combines regression modeling and deep learning techniques to substantially improve the performance of developed models. Firstly, a regression model will be applied to predicting the length of stay for a patient's future ICU visit. Secondly, deep learning approach will be adopted to assess individual's future visit in terms of time and the primary disease. If the modeling gets adopted in a hospital, the predicted results can be promisingly utilized as a reference for medical professionals and experts to offer effective health care guidance for patients. The proposed framework can also be utilized for developing an innovative tool that helps individuals create and maintain a better healthcare plan over time.

Original languageEnglish (US)
Pages (from-to)112-117
Number of pages6
JournalProcedia Computer Science
Volume168
DOIs
StatePublished - 2020
Event2020 Complex Adaptive Systems Conference, CAS 2019 - Malvern, United States
Duration: Nov 13 2019Nov 15 2019

All Science Journal Classification (ASJC) codes

  • Computer Science(all)

Fingerprint Dive into the research topics of 'Patients' disease risk predictive modeling using MIMIC data'. Together they form a unique fingerprint.

Cite this