Event Title

Development of a User-Friendly Biomedical Database

Session Number

P04

Advisor(s)

Satyender Goel, Northwestern University
Kathryn Jackson, Northwestern University
Niloufar Safaenelli, Northwestern University

Location

A-133

Start Date

28-4-2016 8:00 AM

End Date

28-4-2016 8:25 AM

Disciplines

Medicine and Health Sciences

Abstract

As an effect of the rising use of health information technology, biomedical databases consist of rising numbers of discrepancies. Ultimately this is affects the efficiency of data analysis in medical research. This study aimed to develop a method of eliminating these discrepancies and developing a biomedical database into a more research-friendly version. Medical lab data was acquired from HealthLNK Data Repository (HDR) which consisted of patient data from seven different medical institutions (Northwestern Medicine, University of Chicago Hospitals and Clinic, Rush University Medical Center, University of Illinois at Chicago Medical Center, Loyola University Medical Center, Cook County Health and Hospital Systems, and the Alliance of Chicago). The medical lab data was clustered based on similarities with the assistance of the Google OpenRefine software. Initially, 429 clusters, were identified by Google OpenRefine meaning the data was highly inconsistent. When concluding the clustering process of this investigation the number of clusters identified had been reduced to 43. As a result of clustering the data, a suggested standard method of labeling medical labs across medical institutions was developed in order to decrease the possibility of discrepancies arising in the future. The results of this investigation have projected a possible solution to making biomedical databases consistently user friendly in medical research


Share

COinS
 
Apr 28th, 8:00 AM Apr 28th, 8:25 AM

Development of a User-Friendly Biomedical Database

A-133

As an effect of the rising use of health information technology, biomedical databases consist of rising numbers of discrepancies. Ultimately this is affects the efficiency of data analysis in medical research. This study aimed to develop a method of eliminating these discrepancies and developing a biomedical database into a more research-friendly version. Medical lab data was acquired from HealthLNK Data Repository (HDR) which consisted of patient data from seven different medical institutions (Northwestern Medicine, University of Chicago Hospitals and Clinic, Rush University Medical Center, University of Illinois at Chicago Medical Center, Loyola University Medical Center, Cook County Health and Hospital Systems, and the Alliance of Chicago). The medical lab data was clustered based on similarities with the assistance of the Google OpenRefine software. Initially, 429 clusters, were identified by Google OpenRefine meaning the data was highly inconsistent. When concluding the clustering process of this investigation the number of clusters identified had been reduced to 43. As a result of clustering the data, a suggested standard method of labeling medical labs across medical institutions was developed in order to decrease the possibility of discrepancies arising in the future. The results of this investigation have projected a possible solution to making biomedical databases consistently user friendly in medical research