Chapter 7 - Predictive analytics for reducing the risk of QT interval prolongation and torsades de pointes

Abstract

Predictive analytics is the practice of extracting information from existing data sets for the purpose of determining patterns and predicting future outcomes and trends, using techniques such as modeling, machine learning, and data mining. Prolongation of the QT interval and the occurrence of the associated life-threatening arrhythmia known as torsades de pointes (TdP) are largely dependent on the presence of risk factors. Consequently, predictive analytics approaches to QT interval prolongation and TdP may be beneficial for diminishing the likelihood of a potentially catastrophic arrhythmia. Numerous predictive analytics approaches have been used to quantify the risk of QT interval prolongation, including risk scores such as the Mayo Clinic pro-QTc score, the Tisdale risk score in patients hospitalized in cardiac intensive care units, the preliminary and modified RISQ-PATH scores, the modified frailty index, polygenic risk scores, and numerous others. Some of these risk scores perform well with respect to area under the receiver operating characteristics curve, sensitivity, specificity, positive and negative predictive value, while others perform well primarily with respect to sensitivity and negative predictive value, or negative predictive value alone, but have limitations regarding specificity and positive predictive value. Similarly, a variety of predictive analytics models have been developed that incorporate QT interval measurements and numerous other factors to identify individuals and patients at the greatest risk of experiencing TdP or sudden cardiac death (SCD). Many of these risk models utilize easily obtainable patient information and data, including the QT interval, to identify community-based individuals and hospitalized patients at high risk of SCD and so that those with modifiable risk factors can be targeted for intervention and enhanced monitoring strategies can be implemented for those with non-modifiable risk factors. Finally, clinical decision support (CDS) tools incorporating predictive analytics data have been developed to alert clinicians when patients are at moderate or high risk for developing QT interval prolongation or when patients have already developed QT interval prolongation. Implementation of these CDS approaches has been shown to reduce the risk of QT interval prolongation in cardiac ICUs and identify patients at increased risk for mortality so that risk-modifying interventions can be performed.