Predicting trauma patient mortality within 48 hours after hospital admission.

Steen, B. van der (Bob) (2014) Predicting trauma patient mortality within 48 hours after hospital admission. thesis, Medicine.

Text SteenvanderB.pdf Restricted to Registered users only Download (1MB)

Abstract

Objective: Death due to external cause is the fourth cause of death in the Netherlands with high mortality in young working males under the age of 35 years. Major causes of trauma death within the first 48 hours are injuries to the central nervous system, unsurvivable wounds and massive bleeding. Deaths due to massive bleeding is preventable in most cases if emergency interventions are started as early as possible. Therefore it is useful to know what the prognosis for survival is for trauma patients. Method: A retrospective cohort study including polytraumatized patients with an Injury Severity Score (ISS) of more than 15 was used to develop the prediction model. All data was used to develop the model. Missing data was handled using multiple imputation. Ten imputation sets were created and a Markov chain Monte Carlo imputation method with maximum 50 iterations was used. A multivariable logistic regression model was build. Mortality within 48 hours was the outcome variable. Results: The final prediction model includes pre hospital pulse (beats per minute), prehospital reanimation (yes/no), trauma mechanism (blunt/sharp), first measured Glasgow Coma Scale (GCS) at the emergency department, lactate, APTT and fibrinogen. The final prediction model had an average Area Under the Reciever Operator Curve (AUROC) of 0.8934. Conclusion: The developed prediction model includes easy to measure parameters. It can be used to predict early mortality in polytraumatized patients.

Item Type:	Thesis (Thesis)
Supervisor name:	Moumni, Drs. M. el
Faculty:	Medical Sciences
Date Deposited:	25 Jun 2020 11:04
Last Modified:	25 Jun 2020 11:04
URI:	https://umcg.studenttheses.ub.rug.nl/id/eprint/2383

Actions (login required)

View Item