A smart algorithm for the prevention and risk management of QTc prolongation based on the optimized RISQ-PATH model

Area of Science:

• Cardiology
• Medical Informatics

Background:

• QTc prolongation is a significant clinical issue with multiple contributing risk factors.
• The existing RISQ-PATH score aids in identifying patients susceptible to QTc prolongation.

Purpose of the Study:

• To refine and validate the RISQ-PATH risk score in an extensive patient cohort.
• To develop an algorithm for generating intelligent QT interval alerts within electronic medical records.

Main Methods:

• A retrospective analysis of electrocardiograms from 17 hospitals in Belgium (2015) was conducted.
• The RISQ-PATH model was optimized using logistic regression on a training dataset.
• Model performance was evaluated on a separate validation dataset (early 2016).

Main Results:

• The study included 60,208 patients; 5.9% exhibited QTc ≥ 450/470 ms and 0.8% had QTc ≥ 500 ms.
• The optimized RISQ-PATH model demonstrated an AUC of 0.772 for predicting QTc ≥ 450/470 ms.
• A cutoff probability of 0.035 yielded a sensitivity of 87.4% and specificity of 46.2%.

Conclusions:

• The validated RISQ-PATH model accurately predicts prolonged QTc intervals (≥ 450/470 ms or ≥500 ms) with high sensitivity (±87%).
• The model's performance was consistent across different QTc thresholds and in the validation cohort.
• The RISQ-PATH model is suitable for integration into clinical decision support systems to generate smart QT alerts.