By Ben Boulden
Amy Hester almost shed tears of joy when she saw the search results from the UAMS Enterprise Data Warehouse.
“I literally almost cried when I saw the data set,” Hester, associate administrator for patient care at UAMS, said. “I was so pleased. I knew how much time that saved me in my research. It’s a very powerful mechanism for being able to get data that’s highly reliable and valid. There are many aspects of the research process that were improved because I was able to use the data warehouse to do it.”
She and fellow researcher Dees Davis, R.N., developed a model to forecast which patients are likely to have a fall so that health care providers could intervene early and prevent falls from happening.
To see if the predictive model was working, and before the data warehouse was online, Hester spent about 800 hours going through almost 2,000 patient records to extract the information she needed to test the model.
Doing the same thing with the data warehouse was lightning fast by comparison and did not require the use of protected health information from the data warehouse. Hester’s complete statistical analysis and output took 30 hours — less than 4 percent of the time it took her to just do the research before.
The data warehouse’s capability also will allow her to better and more frequently monitor the performance of the model in the future.
Using the fall prediction model has been instrumental in reducing fall accidents by 11 percent and reducing the injury rate by 60 percent.
“The data warehouse greatly reduces the time it takes to get from the workbench to the bedside,” Hester said. “That’s really the goal, and it’s working.”