PhD Candidate · Public Health · Tokyo

I build statistical evidence forpolicy that moves populations.

Physician from Syria, doing a PhD in public health in Tokyo. I work with statistics and health data.

/ posterior samplingiter 0accept 0.0%chains 4target π(θ|y)
01

About

I'm from Homs, Syria. I trained as a physician through the war (MD, Homs University), then came to Tokyo on a MEXT scholarship. Six years in, I'm a PhD candidate in public health at St. Luke's International University.

Most days I work with real-world health data (national registries, claims, federated EHR networks), asking things like whether a policy actually changed anything, how two treatments compare once you follow patients past the trial, or where inequality hits hardest.

Medicine taught me to think in decisions. Statistics taught me to think in populations. The work I want to do lives somewhere between them.

02

Research focus

Tobacco control

Forecasting and evaluating tobacco-control targets

Forecasting smoking-prevalence trajectories for 190 countries. A published paper on regional disparities in Japan's Health Japan 21 progress. Ongoing work on tobacco-advertising exposure and heated-tobacco use in Japan.

Stroke & cerebrovascular

Stroke at national scale

Stroke and cerebrovascular disparities at national scale, mostly on the National Inpatient Sample. Race × sex gaps in thrombectomy access. A decade of poststroke spasticity data that will inform the 2026 AHA Scientific Statement. Also an NIH-funded head-to-head of stroke drugs on a federated real-world network.

COVID-19 & long COVID

COVID-19 mortality and long COVID

COVID-19 mortality gradients across Japanese municipalities, plus long-COVID neurological outcomes at Einstein. Right now I'm also decomposing Japan's pandemic-era excess mortality by ICD-10 cause, to see what actually drove the excess beyond COVID itself.

Healthcare delivery

Stress-testing healthcare access

Stress-testing the U.S. Medicaid workforce on 2.5 billion T-MSIS claims across 51 jurisdictions (2018–2024). Also structural barriers to NCD care in conflict-affected Northwest Syria.

/ interlude · covid-19 excess mortality

Japan lost tens of thousands more lives than the baseline predicted.

Weekly all-cause mortality, 2015–2023. Seasonal Poisson baseline (expected; black) vs observed (gray). The shaded overshoots mark the six pandemic waves. The January 2023 peak (8th wave) is Japan's deadliest week on record.

03

Publications

/ interlude · bayesian disease mapping

Cancer mortality in Japan isn't uniform. It's geography.

47-prefecture tile grid. Each cell shows the share of (municipality × cancer) pairs with elevated (red) or protected (blue) mortality. Aomori has the highest elevated share (55.3%), Oita the highest protected (29.2%). Tohoku and Kyushu sit at opposite ends.

04

Experience & education

Research & clinical

  1. Feb 2026 – Present
    Research Fellow (Remote)
    Rockefeller Neuroscience Institute, West Virginia University
    • Five National Inpatient Sample stroke studies, including race × sex disparities in thrombectomy access and a decade of poststroke spasticity data feeding the 2026 AHA Scientific Statement.
    • NIH-funded (U54 GM104942) comparative-effectiveness work on TriNetX (120M+ patients), comparing tenecteplase vs alteplase and cangrelor vs eptifibatide for acute stroke with 90/180-day outcomes.
    • Stress-test framework for 2.5 billion Medicaid claims (T-MSIS 2018–2024, 51 jurisdictions). First-author manuscript.
  2. Feb 2025 – Present
    Co-Investigator
    National Center for Global Health and Medicine, Tokyo
    • Lead biostatistical framework for the AMED e-ASIA Joint Research Program linking climate change to nontuberculous mycobacterial (NTM) disease risk across Asia.
    • Engineer an Environmental Risk Index from satellite-derived exposures using dimensionality reduction; stratify prefectures into ordinal risk tiers for targeted surveillance.
  3. Feb 2025 – Present
    Research Assistant
    Hitotsubashi Institute for Advanced Study, Tokyo
    • Forecast prefecture-level colorectal cancer screening uptake using Bayesian regression on logit-transformed rates; estimate that <5% of prefectures will meet Japan's 60% target by 2028.
    • Quantify excess stroke deaths during Japan's pandemic period (2020–2022) using log-linear Poisson with parametric bootstrap.
    • Model structural barriers to NCD care in conflict-affected Northwest Syria using modified Poisson regression with district fixed effects.
  4. May 2023 – Present
    Research Fellow
    National Cancer Center, Tokyo
    • Municipality-level COVID-19 mortality gradients via Bayesian hierarchical spatial models, with a publication linking socioeconomic deprivation to excess pandemic deaths.
    • Bayesian projection framework tracking prefectural progress toward Health Japan 21 tobacco-reduction targets; policy briefs to regional decision makers.
    • Northward shift of cancer mortality inequality across Japan (1995 vs 2023) via leave-one-out age-standardised excess rates.
  5. Apr 2024 – Sep 2024
    Research Consultant
    WHO FCTC Secretariat (via St. Luke's Int'l University)
    • Bayesian hierarchical models projecting tobacco-prevalence trajectories for ≈190 WHO Member States under varying policy scenarios; deliverables used for treaty monitoring.
  6. Jun 2023 – Jan 2026
    Research Volunteer (Remote)
    Albert Einstein College of Medicine, New York
    • Patient cohorts in OMOP CDM; SQL / R queries against multi-site EHR warehouses for observational pharmacoepi.
    • Fine–Gray competing risks, Cox PH, GEE for long-term neurological outcomes (Parkinson's progression, post-COVID sequelae).
  7. Jun 2023 – Jun 2024
    Research Assistant, Dept. of Cell Therapy
    National Center for Child Health and Development, Tokyo
    • ML methods (feature engineering, model comparison, cross-validation) to prioritise pathogenic variants in hemophilia-related bioinformatics data.
  8. Mar 2021 – Aug 2021
    Internal Medicine Resident
    Tishreen University Hospital, Latakia, Syria
    • Acute and chronic inpatients across general internal medicine, including clinical assessment, differential diagnosis, and multidisciplinary ward rounds.
05

Get in touch

Happy to hear from you. Collaborations, questions, or just to say hi.

Manuscripts in preparation and references available on request.