Related Experiment Video
Updated: Mar 31, 2026
08:27
A New Hybrid Quantitative Evaluation Model for Axillary Junctional Hemorrhage in Swine
Published on: December 6, 2024
735
A joint latent class model for classifying severely hemorrhaging trauma patients
Mohammad H Rahbar1,2, Jing Ning3, Sangbum Choi4
1Division of Clinical and Translational Sciences, Department of Internal Medicine, The University of Texas Medical School at Houston, The University of Texas Health Science Center at Houston, Fannin St, Houston, TX, USA. Mohammad.H.Rahbar@uth.tmc.edu.
BMC Research Notes
|October 27, 2015
Summary
The traditional massive transfusion (MT) definition is flawed for classifying trauma bleeding severity. A new latent class model better identifies severely hemorrhaging patients, improving accuracy and clinical prediction.
Area of Science:
- Trauma and Emergency Medicine
- Biostatistics
- Clinical Research
Background:
- Massive transfusion (MT), defined as ≥10 units of red blood cells (RBCs) within 24 hours, is a standard for trauma bleeding severity.
- MT is limited by survivor bias due to early in-hospital mortality, making it an unreliable classification criterion.
Purpose of the Study:
- To develop and validate a more accurate method for classifying severely hemorrhaging (SH) trauma patients.
- To address the limitations of the traditional MT definition in quantifying bleeding severity.
Main Methods:
- A latent-class (LC) mixture model was applied to retrospective trauma transfusion data.
- An expectation-maximization (EM) algorithm was used to classify patients based on RBC units and 24-hour survival.
- LC-based classification was compared to the MT rule, and predictive performance was evaluated using logistic regression.
Main Results:
- The LC model identified 27% of patients as SH, compared to 45% classified as MT.
- The agreement between LC and MT classification was 73%.
- The SH classification captured 91% of patients who died within 24 hours, including those not meeting MT criteria.
Conclusions:
- The traditional MT definition inadequately represents transfusion practices and outcomes in early trauma resuscitation.
- Joint latent class modeling offers a bias-corrected approach for classifying severely bleeding patients.
- Future prospective studies with time-to-event data can further refine this classification method.
