A Comparison of the Predictive Value of the Glasgow Coma Scale and the Kampala Trauma Score for Mortality and Length of Hospital Stay in Head Injury Patients at a Tertiary Hospital in Uganda: A Diagnostic Prospective Study

Abstract

Introduction. The prevalence rates of head injury have been shown to be as high as 25% among trauma patients with severe head injury contributing to about 31% of all trauma deaths. Triage utilizes numerical cutoff points along the scores continuum to predict the greatest number of people who would have a poor outcome, “severe” patients, when scoring below the threshold and a good outcome “non severe” patients, when scoring above the cutoff or numerical threshold. This study aimed to compare the predictive value of the Glasgow Coma Scale and the Kampala Trauma Score for mortality and length of hospital stay at a tertiary hospital in Uganda. Methods. A diagnostic prospective study was conducted from January 12, 2018 to March 16, 2018. We recruited patients with head injury admitted to the accidents and emergency department who met the inclusion criteria for the study. Data on patient’s demographic characteristics, mechanisms of injury, category of road use, and classification of injury according to the GCS and KTS at initial contact and at 24 hours were collected. The receiver operating characteristics (ROC) analysis and logistic regression analysis were used for comparison. Results. The GCS predicted mortality and length of hospital stay with the GCS at admission with AUC of 0.9048 and 0.7972, respectively (KTS at admission time, AUC 0.8178 and 0.7243). The GCS predicted mortality and length of hospital stay with the GCS at 24 hours with AUC of 0.9567 and 0.8203, respectively (KTS at 24 hours, AUC 0.8531 and 0.7276). At admission, the GCS at a cutoff of 11 had a sensitivity of 83.23% and specificity of 82.61% while the KTS had 88.02% and 73.91%, respectively, at a cutoff of 13 for predicting mortality. At admission, the GCS at a cutoff of 13 had sensitivity of 70.48% and specificity of 66.67% while the KTS had 68.07% and 62.50%, respectively, at a cutoff of 14 for predicting length of hospital stay. Conclusion. Comparatively, the GCS performed better than the KTS in predicting mortality and length of hospital stay. The GCS was also more accurate at labelling the head injury patients who died as severely injured as opposed to the KTS that categorized most of them as moderately injured. In general, the two scores were sensitive at detection of mortality and length of hospital stay among the study population.