Healthcare Analytics
Clinical data analysis, hospital quality, readmissions, public health data, data quality and clinical interpretation.
Medicine × Data × Informatics
Medical Doctor combining clinical reasoning with Python, SQL, healthcare analytics, interoperability standards, and responsible predictive modeling to turn healthcare data into reproducible, decision-ready insights.
01 · About
I am a Medical Doctor building a professional path across healthcare data analytics, clinical analytics, health informatics, and responsible healthcare AI.
My work starts with the healthcare question and unit of analysis, then moves through data quality, reproducible methods, interpretation, and communication. I am particularly interested in hospital quality, clinical data, interoperability, public health data, and decision-support analytics.
02 · Capabilities
A clinical + analytical toolkit for healthcare data work.
Clinical data analysis, hospital quality, readmissions, public health data, data quality and clinical interpretation.
Python, Pandas, NumPy, SQL, exploratory analysis, hypothesis testing, effect sizes and multiple-testing correction.
Power BI, Matplotlib, Seaborn, KPI design, dashboards and decision-oriented communication.
FHIR, HL7 v2, ICD-10-CM, LOINC, UCUM, Scikit-learn, classification and model evaluation.
03 · Selected Work
Projects selected for analytical rigor, clinical context, reproducibility, and transparent limitations.
Hospital Quality · Readmissions · Analytics
CMS HRRP FY2026 analysis covering 18,330 hospital-condition records across 3,055 hospitals and six HRRP conditions, with hospital benchmarking, geographic analysis, data-quality documentation, an interactive dashboard, and a methodologically separate patient-level research workflow.
Cardiovascular Analytics · Machine Learning
Exploratory cardiovascular analysis with leakage-aware preprocessing, Logistic Regression, Random Forest, hold-out evaluation and cross-validation. The published analysis reports a Logistic Regression hold-out ROC-AUC of 0.933.
FHIR · HL7 · Clinical Terminologies
A fully synthetic interoperability portfolio demonstrating Patient, Condition, Observation, MedicationRequest, Encounter and Bundle resources, plus HL7 v2 mapping and terminology use with ICD-10-CM, LOINC and UCUM.
Clinical Data Quality · Statistics
Reproducible data cleaning, EDA, statistical testing, effect-size analysis and multiple-testing correction. Dataset provenance uncertainty is explicitly documented as a data-governance limitation rather than assumed.
Responsible Healthcare AI · Governance
A transparent audit of published model artifacts. Integrity checks are reproducible, while performance claims are intentionally withheld because the original training dataset and complete training workflow are not available in the repository.
04 · Working Principles
Performance numbers are only useful when the evidence behind them is clear.
Define the healthcare question, population, outcome and unit of analysis before choosing methods.
Document missingness, provenance, implausible values, coding assumptions and analytical limitations.
Use documented workflows, version control, automated checks and transparent decision trails.
Predictive models require validation, calibration, subgroup assessment and governance before real-world clinical use.
05 · Career Direction
Focused on the UAE, Gulf region, remote and international healthcare organizations.