Georgia Institute of Technology
Predicting Gestational Uterine Artery Perfusion with Blood Pressure Pulse Wave Analysis in Ethiopia
Abstract
dc:description.abstractMaternal mortality is a significant contributor to the premature death of women in Ethiopia. As of 2020, Ethiopia had the 4th highest number of maternal deaths worldwide. While some of the conditions precipitating these deaths are unavoidable or already being addressed with novel interventions, others, like preeclampsia (PE), require solutions for improved prevention. PE is a gestational vascular disorder and the 2nd leading cause of maternal death globally. Low-dose (and low-cost) aspirin administered early in pregnancy can reduce the risk of developing PE, but current risk screening methods require high-resource technologies like ultrasound that are widely inaccessible in Ethiopia and other low- to middle-income countries (LMICs). Therefore, there is a need to develop an alternative low-cost screening method for the early risk assessment of developing PE in Ethiopia and other LMICs. To address this challenge, we planned and executed a preeclampsia research study in Ethiopia, gathered a wide array of hemodynamic and biochemical data, and are now developing an alternative method to screen for preeclampsia risk in Ethiopia. In this work, we demonstrate that there are statistically significant differences between femoral and brachial diastolic pressure waveforms in participants with varying uterine artery pulsatility indices (UA PI). UA PI is a metric traditionally measured using ultrasound technology and is commonly used by clinicians in developed countries to screen for risk of developing preeclampsia. Following this promising finding, statistical shape analysis of brachial and femoral pulse pressure waveforms is performed using principal component analysis (PCA) to better characterize 2-D shape characteristics of the diastolic waveform (relative to traditionally used shape parameters). We then use these PCA results to demonstrate again the statistically significant difference between the shapes of pressure waveforms between low and high UA PI groups. In conclusion, we present preliminary classification model results (with the best model having a TPR of 0.7 at FPR 0.2 and AUC = 0.81) to show the predictive potential of this method and the primary direction of future work. Upon further development of these models into a screening metric, this work will produce a low-cost and low-expertise alternative to screen for risk of developing preeclampsia and will help inform and improve antenatal care in Ethiopia and other LMICs.
Degree
thesis:*- Level thesis:degree_level
- Masters
- Department dc:contributor.department
- Mechanical Engineering
- Grantor dc:publisher
- Georgia Institute of Technology
- Year dc:date.issued
- 2024
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Woodall, Julia Lins Arrighi
- Advisor dc:contributor.advisor
-
- Gleason, Rudolph L.
- Committee members dc:contributor.committeemember
-
- Wood, Levi
- Lam, Wilbur
Subjects
dc:subject × 8Rights
- Language dc:language.iso
- en_US
Identifiers
dc:identifier.*- Handle dc:identifier.uri
- https://hdl.handle.net/1853/82005