Utility of triage early warning scores in the care of critically ill surgical patients at kenyatta national hospital

Abstract

Critical illness is a life-threatening condition involving one or more organ systems resulting in significant morbidity or mortality. Often, it is preceded by a period of physiological deterioration. Such early signs of critical illness are frequently missed, leading to late recognition by clinicians with consequent worsening morbidity and mortality. In such patients, icu mortality has been estimated to be at 8-18%. Tools have been invented to capture such patients and prevent these outcomes by early intervention. Such tools include, modified early warning score, national early warning score; and triage early warning score. The general wards at knh do not have a decision support tool to help identify and manage acutely deteriorating patients. Therefore, the ward care of patients with critical illness is suboptimal. Objectives The broad objective was to establish the appropriateness of care level of critically ill surgical patients using the triage early warning score in knh. Specifically, the intentions were to: identify critically ill surgical patients using the triage early warning score tool in knh a&e, establish a relationship between the triage early warning score and outcome of critically ill surgical patients after 72 hours of follow up; and determine the specificity and sensitivity of the cut off values of the triage early warning scores at knh. Methods This was a prospective observational study involving 168 critically ill surgical patients who were followed up for 72 hours following recruitment. 4 hourly vitals, decisions regarding intervention, level of care and clinical outcomes were recorded. A relationship between the tews and clinical outcomes was established using logistic regression, while the specificity and the sensitivity of the cut-off score for the tews were established using the receivership operating characteristics curve. Results 94 % of the cases presenting to knh were due to trauma, while 6% were non traumatic cases. The most common score was 5 and the highest score recorded was 11. After 72 hours of follow up, 4.23% (7) had unplanned icu admission and their average tews was 7.71, while 6.67% ( 11)

Patients had died and their average tews was 7.55. The odds ratio for bad outcome (death and unplanned icu admission) was 7.708 ( 95% ci 3.48-17.073). The tews was found to have good sensitivity at identifying patients at risk of adverse outcome, with cut off values of 6.5 and 7.5 for prediction of mortality and icu admission respectively. Conclusion Icu strain and the burden of trauma pose a significant challenge at knh. Based on our findings, the tews is a sensitive tool for predicting risk for unplanned icu admission and death. Timely identification and action for patients at risk of deterioration using the tews may reduce adverse events and outcomes. However, since the tews is a modification of the modified early warning score, it may over-triage patients, due to addition of the immobility component.