Abstract
.png)
.webp)
Authors: Sara Sauer, Timothy Wen, Noam Finkelstein, Mia Charifson, Adesh Kadambi, Chloe Li, Shreyas Kadambi, Kartik Venkatesh, and Isabel Fulcher
Conference: American Heart Association's Hypertension 2025 Scientific Sessions
Key Findings: Machine learning models significantly outperformed the standard USPSTF checklist in predicting hypertensive disorders of pregnancy (HDP). At a fixed specificity level of 0.62, the baseline machine learning model achieved a sensitivity of 0.76 (AUC = 0.75) compared to just 0.58 sensitivity for the USPSTF checklist. Incorporating longitudinal blood pressure data from remote at-home and in-clinic monitoring into a dynamic model further enhanced predictive power, reaching an AUC of 0.88 and a sensitivity of 0.92. Ultimately, dynamic machine learning models leveraging ongoing blood pressure tracking offer superior early risk identification for HDP over static clinical guidelines.
Abstract
Authors: Sara Sauer, Timothy Wen, Noam Finkelstein, Mia Charifson, Adesh Kadambi, Chloe Li, Shreyas Kadambi, Kartik Venkatesh, and Isabel Fulcher
Conference: American Heart Association's Hypertension 2025 Scientific Sessions
Key Findings: Machine learning models significantly outperformed the standard USPSTF checklist in predicting hypertensive disorders of pregnancy (HDP). At a fixed specificity level of 0.62, the baseline machine learning model achieved a sensitivity of 0.76 (AUC = 0.75) compared to just 0.58 sensitivity for the USPSTF checklist. Incorporating longitudinal blood pressure data from remote at-home and in-clinic monitoring into a dynamic model further enhanced predictive power, reaching an AUC of 0.88 and a sensitivity of 0.92. Ultimately, dynamic machine learning models leveraging ongoing blood pressure tracking offer superior early risk identification for HDP over static clinical guidelines.
Abstract
Authors: Sara Sauer, Timothy Wen, Noam Finkelstein, Mia Charifson, Adesh Kadambi, Chloe Li, Shreyas Kadambi, Kartik Venkatesh, and Isabel Fulcher
Conference: American Heart Association's Hypertension 2025 Scientific Sessions
Key Findings: Machine learning models significantly outperformed the standard USPSTF checklist in predicting hypertensive disorders of pregnancy (HDP). At a fixed specificity level of 0.62, the baseline machine learning model achieved a sensitivity of 0.76 (AUC = 0.75) compared to just 0.58 sensitivity for the USPSTF checklist. Incorporating longitudinal blood pressure data from remote at-home and in-clinic monitoring into a dynamic model further enhanced predictive power, reaching an AUC of 0.88 and a sensitivity of 0.92. Ultimately, dynamic machine learning models leveraging ongoing blood pressure tracking offer superior early risk identification for HDP over static clinical guidelines.
Abstract
Authors: Sara Sauer, Timothy Wen, Noam Finkelstein, Mia Charifson, Adesh Kadambi, Chloe Li, Shreyas Kadambi, Kartik Venkatesh, and Isabel Fulcher
Conference: American Heart Association's Hypertension 2025 Scientific Sessions
Key Findings: Machine learning models significantly outperformed the standard USPSTF checklist in predicting hypertensive disorders of pregnancy (HDP). At a fixed specificity level of 0.62, the baseline machine learning model achieved a sensitivity of 0.76 (AUC = 0.75) compared to just 0.58 sensitivity for the USPSTF checklist. Incorporating longitudinal blood pressure data from remote at-home and in-clinic monitoring into a dynamic model further enhanced predictive power, reaching an AUC of 0.88 and a sensitivity of 0.92. Ultimately, dynamic machine learning models leveraging ongoing blood pressure tracking offer superior early risk identification for HDP over static clinical guidelines.
Abstract
Authors: Sara Sauer, Timothy Wen, Noam Finkelstein, Mia Charifson, Adesh Kadambi, Chloe Li, Shreyas Kadambi, Kartik Venkatesh, and Isabel Fulcher
Conference: American Heart Association's Hypertension 2025 Scientific Sessions
Key Findings: Machine learning models significantly outperformed the standard USPSTF checklist in predicting hypertensive disorders of pregnancy (HDP). At a fixed specificity level of 0.62, the baseline machine learning model achieved a sensitivity of 0.76 (AUC = 0.75) compared to just 0.58 sensitivity for the USPSTF checklist. Incorporating longitudinal blood pressure data from remote at-home and in-clinic monitoring into a dynamic model further enhanced predictive power, reaching an AUC of 0.88 and a sensitivity of 0.92. Ultimately, dynamic machine learning models leveraging ongoing blood pressure tracking offer superior early risk identification for HDP over static clinical guidelines.
Abstract
Authors: Sara Sauer, Timothy Wen, Noam Finkelstein, Mia Charifson, Adesh Kadambi, Chloe Li, Shreyas Kadambi, Kartik Venkatesh, and Isabel Fulcher
Conference: American Heart Association's Hypertension 2025 Scientific Sessions
Key Findings: Machine learning models significantly outperformed the standard USPSTF checklist in predicting hypertensive disorders of pregnancy (HDP). At a fixed specificity level of 0.62, the baseline machine learning model achieved a sensitivity of 0.76 (AUC = 0.75) compared to just 0.58 sensitivity for the USPSTF checklist. Incorporating longitudinal blood pressure data from remote at-home and in-clinic monitoring into a dynamic model further enhanced predictive power, reaching an AUC of 0.88 and a sensitivity of 0.92. Ultimately, dynamic machine learning models leveraging ongoing blood pressure tracking offer superior early risk identification for HDP over static clinical guidelines.
Abstract
