RWE Scientist | Health Policy | Oncology HEOR | Pharmacoeconomics
I build real-world evidence studies and health economic models in oncology, with a focus on treatment patterns, survival outcomes, and healthcare disparities using SEER and claims data.
A three-paper evidence generation strategy in urothelial carcinoma, modeled on industry HEOR standards.
| Paper | Title | Status | Data |
|---|---|---|---|
| P1 | Stage at diagnosis, treatment utilization, and survival in MIBC before and after ICI availability — a SEER analysis | 🟡 In progress | SEER public |
| P2 | HCRU and economic burden in advanced bladder cancer by treatment line | 🔵 Planned | Claims |
| P3 | Comparative effectiveness of EV+pembro vs. chemotherapy — target trial emulation | 🔵 Planned | Claims |
Reproducible RWE pipeline for Paper 1 — cohort identification, trend analysis, survival analysis, competing risk (Fine-Gray), Cox regression, and logistic regression for no-treatment receipt. SEER public data, Python.
Methods: Interrupted time series · Kaplan-Meier · Fine-Gray competing risk · Multivariable Cox regression · Logistic regression · Joinpoint trend analysis
Cost-effectiveness analysis for cariprazine in schizophrenia. Markov model, PSA, ICER estimation.
Methods: Markov cohort model · Probabilistic sensitivity analysis · Cost-effectiveness plane · CEAC
Languages │ Python · R · SAS (familiar)
RWE Methods │ Survival analysis · Competing risks · ITS · Propensity scoring
HEOR Methods │ Cost-effectiveness · Budget impact · Markov models · PSA
Data Sources │ SEER · CMS SynPUFs · TriNetX (prior experience)
Libraries │ lifelines · statsmodels · pandas · numpy · matplotlib- Retrospective cohort studies using cancer registry and claims data
- Interrupted time series for policy/approval change impact analysis
- Target trial emulation framework for causal RWE
- Cost-effectiveness modeling using Markov state-transition models
Oncology RWE · Urothelial Carcinoma · Immune Checkpoint Inhibitors · Health Disparities · Pharmacoeconomics · Treatment Patterns · HEOR
All code is reproducible and openly available. Data sources are publicly accessible (SEER) or synthetic (CMS SynPUFs). No patient-identifiable data is stored in any repository.