An Early Warning Tool for Predicting Mortality Risk of COVID-19 Patients Using Machine Learning
Problem
Healthcare systems under pandemic pressure needed fast, early assessment of COVID-19 patient severity to allocate resources and reduce mortality.
Why it matters
Early severity triage during COVID-19 surges helps prioritize scarce care and reduce deaths.
What the AI did
A multi-tree XGBoost model analyzed blood biomarkers from 375 COVID-19 patients to identify key mortality predictors, which fed a nomogram for individual mortality risk.
Results
Identified LDH, neutrophils (%), lymphocytes (%), hs-CRP, and age as key death predictors; AUC value truncated in abstract.
Impact
Offers an early-warning nomogram for COVID-19 mortality; external validation not stated.
Record
| Venue | Cognitive Computation |
|---|---|
| Published | 2021 |
| Work type | article |
| Authors | Muhammad E. H. Chowdhury, Tawsifur Rahman, Amith Abdullah Khandakar, Somaya Al-Madeed, Susu M. Zughaier, Suhail A.R. Doi, Hanadi Hassen, Mohammad Touhidul Islam |
| Citations | 114 (OpenAlex, at import time) |
| References | 46 |
| Model or method | multi-tree XGBoost |
| Organization | not stated in source; data from Tongji Hospital, China |
| Primary source | https://doi.org/10.1007/s12559-020-09812-7 |
| Open access | hybrid |
| Keywords | COVID-19, mortality prediction, XGBoost, blood biomarkers, nomogram |
Importance rating
major — millions of lives, an industry or a field
Abstract states COVID-19 pandemic created extreme pressure on global healthcare services; early severity triage to reduce mortality. Pandemic-scale problem, single-hospital dataset.