JAMIA: Heterogeneous clinical databases can assist clinical association awareness
Researchers from Columbia University determined that, if examined properly, information contained in large, heterogeneous clinical databases can provide detailed illustrations of the temporal patterns of clinical associations and of the types of clinical associations that are made.
The study, the results of which were published Nov. 28 in the Journal of the American Medical Informatics Association, builds off of past research regarding the relationships among clinical variables with a slightly different angle in comparison: the exploitation of time.
Using the New York-Presbyterian Hospitals’ clinical database, which contains 22 years worth of EHR data on 3 million patients, the researchers calculated linear correlation between seven clinical laboratory values and 30 clinical concepts that included diseases, medications and symptoms.
“A relatively simple method, which we use in this study, is to measure linear correlation between co-occurrences of pairs of variables, lagging one variable with respect to the other to assess the change in correlation as variables are shifted in time,” George Hripcsak, MD, of Columbia University, New York City, and his fellow researchers wrote.
The study’s goal was to show that clinical databases can be used to reveal temporal associations, show different types of associations and determine any value associated with the exploitation of time.
The study’s authors believe that large clinical databases will prove to be useful tools in the future because the aggregate data that they provide offer reflections of the healthcare process and the recording process, but that healthcare professionals must develop an understanding of EHRs’ inherent biases when conducting analyses of data.
"We classified associations into three types -- definitional, physiologic, and intentional -- and showed that care must be taken in interpreting the associations because the health record represents the clinical workflow and not just patient physiology,” Hripcsak concluded. “We found that fully exploiting time in the record revealed the most detailed and reliable information.“
The study, the results of which were published Nov. 28 in the Journal of the American Medical Informatics Association, builds off of past research regarding the relationships among clinical variables with a slightly different angle in comparison: the exploitation of time.
Using the New York-Presbyterian Hospitals’ clinical database, which contains 22 years worth of EHR data on 3 million patients, the researchers calculated linear correlation between seven clinical laboratory values and 30 clinical concepts that included diseases, medications and symptoms.
“A relatively simple method, which we use in this study, is to measure linear correlation between co-occurrences of pairs of variables, lagging one variable with respect to the other to assess the change in correlation as variables are shifted in time,” George Hripcsak, MD, of Columbia University, New York City, and his fellow researchers wrote.
The study’s goal was to show that clinical databases can be used to reveal temporal associations, show different types of associations and determine any value associated with the exploitation of time.
The study’s authors believe that large clinical databases will prove to be useful tools in the future because the aggregate data that they provide offer reflections of the healthcare process and the recording process, but that healthcare professionals must develop an understanding of EHRs’ inherent biases when conducting analyses of data.
"We classified associations into three types -- definitional, physiologic, and intentional -- and showed that care must be taken in interpreting the associations because the health record represents the clinical workflow and not just patient physiology,” Hripcsak concluded. “We found that fully exploiting time in the record revealed the most detailed and reliable information.“