Research
Clinical AI systems in practice
I build end-to-end clinical AI systems that prioritize measurable outcomes, transparent evaluation, and real-world deployment constraints. My current work includes nephrology, intensive care, cardiology, and infectious diseases research.
Research themes
Clinical AI
Cardiology / Nephrology / ICU deterioration / Physiologic signals
Epidemiology
Outbreak research / Population risk / Surveillance
Systems
Hardware-software integration / Deployment safety / Real-world evaluation
NLP + LLMs
Clinical NLP / BERT pipelines / Retrieval + evaluation
Publications
Prognostic Value of Different Iron Status Definitions in Congestive Heart Failure: A Retrospective MIMIC-IV Analysis of Risk Stratification and Mortality
Retrospective MIMIC-IV analysis of congestive heart failure compares multiple iron-status definitions. It evaluates ferritin and transferrin saturation criteria for 1-year mortality risk stratification. Combined ferritin and TSAT categories improve separation between iron deficiency and hyperferritinemia. The study argues for more precise phenotyping to guide prognostic assessment.
DOI: 10.3390/jcm15010244
SOURCE-WORK-ID: 13c45e33-ff6f-485a-bfa9-403fdbba3f82
ISSN: 2077-0383
Source: Medical University of Warsaw, Poland - WUM.Publikacje
Contributors: Hourani A; Surmeli A; Devarapalli S
Clinical Picture and Outcomes in Patients Diagnosed with Brain Abscess
Retrospective cohort study characterizes the clinical presentation and outcomes of patients with brain abscess. It evaluates factors linked to hospitalization course and overall prognosis. Age and abscess size are highlighted as drivers of longer hospital stay. The work underscores the importance of early recognition and coordinated management for CNS infection.
DOI: 10.3390/jcm14207237
SOURCE-WORK-ID: ad3005e4-6b88-44d8-a67d-dff59f44be86
ISSN: 2077-0383
PMID: 41156107
WOSUID: WOS:001601753900001
EID: 2-s2.0-105020200936
Source: Medical University of Warsaw, Poland - WUM.Publikacje
Contributors: Furman-Dlubala A; Bednarska A; Radkowski M; Paciorek M; Kolodziejska J; Laskus T; Bursa D; Porowski D; Makowiecki M; Hourani A et al.
HALP and mHALP as Effective Tools for 90-Day Mortality Prediction in Heart Failure
Retrospective MIMIC-IV analysis evaluates HALP and modified HALP scores in heart failure. It compares their ability to predict 90-day mortality using risk modeling. Lower scores are associated with higher mortality and mHALP performs better than HALP. The preprint proposes these indices as accessible tools for early risk stratification.
DOI: 10.1101/2025.05.22.25328139
Source: Crossref
Contributors: Abdulla Hourani; Arman David Surmeli; Sai Keertana Devarapalli; Michal Oreziak
The Impact of COVID-19 Non-Pharmaceutical Interventions on Notifiable Infectious Diseases in Poland: A Comprehensive Analysis from 2014-2022
Comprehensive analysis of notifiable infectious disease trends in Poland from 2014 to 2022. It examines patterns across the period that includes COVID-19 non-pharmaceutical interventions. The study frames how population-level measures coincide with shifts in reported incidence across diseases. It aims to inform surveillance and long-term public health planning.
DOI: 10.1101/2025.03.05.25323398
Source: Crossref
Contributors: Abdulla Hourani; Abdelrahman Abdelsalman; Arman David Surmeli
Prognostic Value of Different Iron Status Definitions in Congestive Heart Failure: A Retrospective MIMIC-IV Analysis of Risk Stratification and Mortality
Retrospective MIMIC-IV preprint investigates iron-status definitions in congestive heart failure. It tests ferritin and transferrin saturation thresholds against mortality outcomes. Combined ferritin-TSAT categories provide clearer risk stratification signals. The work presents early findings ahead of journal publication.
DOI: 10.1101/2025.03.05.25323191
Source: Crossref
Contributors: Abdulla Hourani; Arman David Surmeli; Sai Keertana Devarapalli
Tools and repos
Reproducible pipelines, local LLM toolchains, and ML ops tooling.
GitHub links will be added as projects are shared publicly.
Collaboration interests
I partner with clinicians, researchers, and health systems looking to validate AI workflows in real practice.
If you are exploring clinical AI deployment, send a note to hourani03@gmail.com.