Welcome to my website

Selected Analytical Projects

Applied work in clinical trial data programming, oncology, population health, and healthcare quality, demonstrating data preparation, statistical modeling, model diagnostics, and interpretation using SAS, R, and SQL.

In progress · Educational project · R

Clinical Trial Data Programming: From SDTM to ADaM

Developing a transparent clinical data programming workflow using public example datasets, connecting standardized source records with participant-level and adverse-event analysis datasets.

  • Linked demographics, treatment exposure, and adverse-event domains to derive preliminary ADSL and ADAE datasets using base R and pharmaversesdtm.
  • Derived safety-population flags, treatment dates, relative study days, event durations, and preliminary treatment-emergent classifications, with explicit handling of partial dates and source traceability.
  • Distinguished event counts from participant counts; planned extensions include admiral-based derivations, metadata preparation, and XPT export with xportr.

Clinical data programming · Analysis dataset derivation · Safety summaries · Traceability

Educational implementation using example data; formal ADaM conformance has not been established.

View R code and project documentation

Completed · SAS

Breast Cancer Survival Analysis

Evaluated the association between hormone receptor status and overall survival among women diagnosed with breast cancer from 2006–2010 using SEER data.

  • Estimated Kaplan–Meier survival curves and compared receptor groups using log-rank tests with Bonferroni adjustment.
  • Fitted multivariable Cox proportional hazards models and assessed model assumptions using Schoenfeld residuals and log-log plots.
  • Identified survival differences across receptor groups and associations with age, stage, tumor size, and lymph node involvement.

Survival analysis · Multivariable modeling · Model diagnostics

View SAS code and analysis

Completed · R & SQL

Hypertension Prevalence and Associated Factors

Analyzed NHANES 2021–2023 data to estimate hypertension prevalence and assess demographic and cardiometabolic associations among U.S. adults.

  • Linked examination and questionnaire datasets and prepared variables for complex survey analysis.
  • Estimated weighted prevalence and confidence intervals using examination weights, strata, and primary sampling units.
  • Fitted adjusted Poisson and logistic regression models; obesity was associated with higher hypertension prevalence (aPR 1.74; 95% CI: 1.57–1.94).

Complex survey methods · Generalized linear models · Results reporting

View code and analytical report

Academic project · SAS

Maternal Nutrition and Birth Weight: A Trial Analysis in Malawi

Compared lipid-based nutrient supplementation with iron and folic acid supplementation in a randomized trial analysis, examining low-birth-weight risk and mean birth weight.

  • Summarized baseline characteristics and applied an intention-to-treat analytical framework.
  • Used log-binomial regression to estimate risk ratios and risk differences, and linear regression to compare mean birth weight.
  • Evaluated effect modification by maternal BMI using interaction terms and stratified analyses.

Randomized trial analysis · Generalized linear models · Effect modification · SAS 9.4

Completed · SAS

Dietary Fiber and Colon Cancer: Case–Control Analysis

Examined the association between dietary fiber intake and colon cancer case status, including effect modification by BMI, education, and age.

  • Cleaned and recoded data, derived BMI and categorical variables, and summarized characteristics by case status.
  • Fitted logistic regression models in PROC GENMOD and evaluated exposure–covariate interactions.
  • Estimated odds-ratio contrasts using ESTIMATE statements.

Data preparation · Logistic regression · Interaction analysis

Course-based analysis; restricted source data are not included.

View SAS code and study overview

Ongoing · Methodological research

Quantile Regression for Longitudinal Biomarkers

Research collaboration between the University of Nevada, Reno and the Department of Health Data Sciences and Biostatistics, Peter O’Donnell Jr. School of Public Health, UT Southwestern.

Developing a conceptual framework to examine whether associations between behavioral patterns and biomarkers differ across the outcome distribution. My interest is in capturing variation that average-based models may overlook, while accounting for repeated measurements and values below laboratory detection limits.

Quantile regression · Longitudinal data · Left-censoring

Methodological reference

In progress · SQL

Nevada Medicaid Quality Benchmarking

Exploring CMS 2023 child and adult healthcare quality measures to evaluate Nevada Medicaid performance relative to national benchmarks.

  • Profiled 5,662 records across seven healthcare quality domains.
  • Examined population labels, reporting programs, and rate definitions to establish appropriate analytical comparisons.
  • Next phase: assess missing, suppressed, and non-numeric rates before comparing Nevada results with national medians and quartiles.

Data profiling · Data quality · Healthcare analytics

View project documentation

Course project · Longitudinal analysis

Longitudinal Modeling of Lipid Outcomes in NAFLD

Analyzed repeated HDL and cholesterol measurements in a nonalcoholic fatty liver disease dataset as part of graduate coursework in biostatistical analysis of cohort studies.

  • Applied linear mixed-effects models and generalized estimating equations to account for within-person correlation.
  • Examined lipid outcomes over time using subject-specific and population-average modeling approaches.

Repeated measures · Mixed-effects models · Generalized estimating equations

Observational analyses describe associations and do not establish causal effects.

Explore GitHub repositories