Link to the University of Pittsburgh Homepage
Link to the University Library System Homepage Link to the Contact Us Form

Deciphering the impact of health disparities and medication use on risk of developing adverse events amongst post-traumatic stress disorder patients using electronic medical records

Oshin, Miranda, (2024) Deciphering the impact of health disparities and medication use on risk of developing adverse events amongst post-traumatic stress disorder patients using electronic medical records. Doctoral Dissertation, University of Pittsburgh. (Unpublished)

[img] PDF
Restricted to University of Pittsburgh users only until 30 April 2025.

Download (5MB) | Request a Copy


The dynamic field of precision medicine relies heavily on integrating artificial intelligence (AI) methods to overcome persistent challenges in biomarker identification, understanding health disparities, and exploring novel drug repurposing. This dissertation employs a novel approach, leveraging electronic medical records (EMRs) with advanced deep learning (DL), natural language processing (NLP), and statistical methodologies for more efficient risk prediction.
Beginning with a focus on predicting suicide-related events (SREs) in post traumatic stress disorder (PTSD) patients using DL models, the research identifies biomarkers, explores drug repurposing, and fosters multi-systemic hypotheses. It unravels the intricate interactions within multi-modal data, shedding light on dynamic factors in PTSD and co-occurring disorders. To address health disparities, a novel NLP approach is introduced, utilizing context-specific dictionaries and a sentence transformer model to identify social determinants of health (SDoH) and transdiagnostic factors. Simultaneously, the DL model is enhanced to predict worse outcomes in PTSD patients, incorporating contribution-based analyses for improved risk prediction accuracy.
A clinical trial emulation study assesses the impact of antidepressants on SRE risk, successfully identifying a beneficial antidepressant associated with decreased risk of SREs. The dissertation introduces DeepBiomarker, a novel deep learning model and its iterations as tools for biomarker identification, accelerating the process by integrating NLP-extracted multi-modal data.
In conclusion, this research effectively combines computational modeling and AI to address key challenges in PTSD research, expanding the horizon for discovering novel risk factors. These methodologies set a new benchmark in pharmaceutical research, emphasizing the potential of precision medicine in the realm of drug discovery and development.


Social Networking:
Share |


Item Type: University of Pittsburgh ETD
Status: Unpublished
CreatorsEmailPitt UsernameORCID
Oshin, Miranda,osm7@pitt.eduosm7
ETD Committee:
TitleMemberEmail AddressPitt UsernameORCID
Thesis AdvisorWang, LiRongliw30@pitt.eduliw30
Committee ChairKirisci, Leventlevent@pitt.edulevent
Committee CoChairRyan, Nealnryan@pitt.edunryan
Committee MemberTarter, Ralphtarter@pitt.edutarter
Committee MemberWang, Junmeijuw79@pitt.edujuw79
Committee MemberWang,
Date: 18 April 2024
Defense Date: 25 March 2024
Approval Date: 30 April 2024
Submission Date: 17 April 2024
Access Restriction: 1 year -- Restrict access to University of Pittsburgh for a period of 1 year.
Number of Pages: 273
Institution: University of Pittsburgh
Schools and Programs: School of Pharmacy > Pharmaceutical Sciences
Degree: PhD - Doctor of Philosophy
Thesis Type: Doctoral Dissertation
Refereed: Yes
Uncontrolled Keywords: Artificial intelligence, Electronic medical records, Post traumatic stress disorder, Health disparities, Personalized medicine
Date Deposited: 30 Apr 2024 13:32
Last Modified: 30 Apr 2024 13:32


Monthly Views for the past 3 years

Plum Analytics

Actions (login required)

View Item View Item