Objective
MD Anderson Cancer Center is a university hospital and global leader in
cancer care, located in Houston, Texas. MD Anderson’s electronic health
record (EHR), Epic, currently utilizes DDI warnings provided by First
Databank (FDB), one of the major providers of drug databases in the
nation. Based on internal assessments, DDI warnings accounted for 47%
of all FDB medication warnings from January 2019 through June 2019. Due
to limitations in functionality, these DDI warnings do not take
patient-specific factors, such as age, cardiovascular history, or
electrolytes, into consideration, and alerts generated for
QTc-prolonging medications are overridden more than 90% of the time.
FDB provides all possible DDI warnings, leaving each institution to
determine which warnings to make visible and which ones to filter or
keep silent. Filtered warnings determined to be of low priority are
turned off by the institution to avoid unnecessary alerts. Replacing the
DDI warnings with a custom alert triggered by an accurate
QTc-prolongation risk score would be beneficial in reducing overall
alert burden and fatigue. The aim of our study is to decrease false
positive DDI warnings for QTc-prolonging medications by 20% through
triggering a custom alert based on a patient-specific QTc-prolongation
risk score at MD Anderson Cancer Center. The accuracy of the
QTc-prolongation risk score and its ability to predict the risk for QTc
prolongation will be determined based on sensitivity and specificity
calculations.