Machine Learning Patient Monitoring

Shanshan Jia

864-***-**** *********@*****.*** Burlingame, CA 94010

Interested in the intersection of clinical data analysisand advanced statistical modeling techniques. Proficient in understanding and applying regression analysis, hypothesis testing techniques and machine learning techniques.

Education

Clemson University

Ph.D. in Mathematical and Statistical Sciences (12/2024) M.S. in Mathematical and Statistical Sciences (08/2022) Internship and Work Experience

PRISMA Health Internship-Remote Patient Monitoring Program 03/2023 – Present

• Collected remote patient monitoring (RPM) data from CardioMEMS HF System. Conducted data cleaning and integration to create unified, patient-level datasets.

• Performed exploratory data analysis (EDA) on raw datasets, including statistical summaries, missing data assessment, and visualization.

• Designed and implemented statistical models for risk prediction and clinical decision- making, supporting evidence-based outcomes in healthcare. Clemson University 08/2017-08/2020

• Graduate Lecturer: Instructed undergraduate courses in Business Calculus. Supported instruction in Introductory Business Statistics and Statistical Methods for Process Development and Control.

Research Experience

Clinical Decision Support Research

• Applied Linear and Generalized linear mixed-effect models (LMM/GLMM) to assess associations between remote monitoring diastolic pulmonary artery pressure (PAP) and variables such as heart rate, age, gender, race, smoking status, and alcohol consumption.

• Employed conditional quantile regression to analyze consecutive days of diastolic PAP measurements, uncovering significant correlations between historical and current readings to predict future measurements across different conditional quantiles (10%, 50%, 90%). Identified instances of elevated diastolic PAP (>40 mmHg) as clinical indicators for intervention, typically followed by a subsequent decline in pressure values.

• Performed changepoint detection methods to investigate mean shift patterns in diastolic PAP. Identified high-risk patients requiring aggressive treatment strategies. 5.3% of patients exhibited stable diastolic PAP with minimal interventions and narrow pressure ranges, indicating well-managed conditions. 19.6% of patients experienced a trade-off between higher mean pressures and improved long-term stability after threshold adjustments, reducing intervention needs. 75.3% of patients showed high variability in PAP with frequent fluctuations and interventions, requiring close monitoring and dynamic treatment strategies.

Other research experience

• Employed Unsupervised Learning (K-means Algorithm) to segment RPM data into patient subgroups, enabling targeted interventions and improving care strategies.

• Implemented Supervised Models (Logistic Regression) for patient risk classification.

• Working on conducting Survival Analysis to predict the occurrence of future events based on historical data, informing clinical decision-making and resource allocation.

• Proficient in understanding and applying Regression Analysis and Time series Analysis(ARIMA, SARIMA).

Publication

Jia, S., Gallagher, C., & Ranganathan, S. (2024). Changepoint Analysis for Remote Heart Failure Patient Monitoring Diastolic Pulmonary Artery Pressure. (Under review) Fellowship and Grant

LaMotte Fellowship 2019

Clemson University Doctoral Dissertation Completion Grant 2024 Skills

• Programming: Python, R, SQL, SAS

• Analytics: Regression Analysis, GLMM, Changepoint Detection, Time Series Analysis, Quantile Regression, Survival Analysis

• AI & ML: Generative AI, Logistic Regression, K-means Clustering Certification

• SAS Certified Specialist: Base Programming Using SAS 9.4

• SAS Certified Professional: Advanced Programming Using SAS 9.4

Contact this candidate