Dissertations & Theses (Open Access)
Date of Award
Summer 5-2019
Degree Name
Master of Science (MS)
Advisor(s)
Jose-Miguel Yamal, Phd
Second Advisor
Xianglin L Du, Phd
Third Advisor
David Aguilar, Phd
Abstract
Abstract: Subarachnoid hemorrhage (SAH) is a life-threatening stroke caused by bleeding in the subarachnoid space. Delayed cerebral ischemia, a risk factor for death during non-traumatic SAH treatment, is often prevented by inducing hemodynamic augmentation through the administration of vasopressors. The three common vasopressors administrated to SAH patients are dopamine, phenylephrine, and norepinephrine. A recent study using Electronic Health Records (EHR) of nontraumatic SAH patients identified that the administration of phenylephrine is associated with lower hospital mortality; however, this study did not consider heterogeneity in treatment effects. Therefore, the goal of this paper is to study the heterogeneity in vasopressor treatment effects by exploratory subgroup analyses. The subgroups were identified by studying the vasopressor and pretreatment covariates interaction effects. We employed a machine learning and a generalized linear model approach to study the interaction effects. We also employed propensity scoring and inverse probability weighting to minimize the confounding and selection biases in treatment effect estimates due to inherent pretreatment group differences in the data. Our results showed that dopamine had the highest mortality among those who did not have ondansetron ( a pretreatment medication used to prevent nausea) compared to phenylephrine and norepinephrine; whereas norepinephrine has the highest mortality among those who had ondansetron compared to phenylephrine and dopamine. However, for the subgroup who did not have ondansetron but had fentanyl and lidocaine, there was no significant difference in the mortality rate between vasopressor treatment group. Overall, those who had ondansetron had better survival compared to those who did not have ondansetron with respect to all the three vasopressors.
Recommended Citation
Thomas, Emy Anu, "Machine-Learning And Statistical Subgroup Analyses Of Vasopressor Choice For The Treatment Of Non-Traumatic Subarachnoid Hemorrhage Using Electronic Health Records" (2019). Dissertations & Theses (Open Access). 96.
https://digitalcommons.library.tmc.edu/uthsph_dissertsopen/96