Data Mining of Blood Incident Databases
Project Award Date: 0000-00-00
The U.S. public and the various health organizations and providers place extreme importance on the quality of the blood supply. While all blood banks, blood handling organizations, and hospitals have very strict quality control procedures, there are still thousands of incidents of incorrect handling of blood and blood products. The goal of the project is to mine very large databases of blood incidents and identify patterns that can lead us to a better understanding of why such incidents occur and how we can minimize them.
We are working together with a group of blood suppliers organized by the University of Texas' Southwestern Medical Center to collect and analyze reports of blood handling incidents. We are using techniques from KDD and data mining-such as clustering and induction-to generate coherent, novel, describable, significant patterns.
Primary Sponsor(s): National Institutes of Health