Abstract
The rapid progression of the dengue fever (DF) epidemic in tropical areas along with the expansion of the expatriate community in Indonesia has led to an increase of medical evacuation related to dengue fever.
With the view of designing a diagnostic tool that contributes to make evidence based decisions to medically evacuate patients with complicated dengue fever, we conducted a retrospective survey of a cohort of 46 cases of patients evacuated from Jakarta over the two previous years. Using multivariate logistic regression analysis, we looked for criteria predictive of the fact that the patient would later require a level of care not immediately available locally.
Within dengue fever patients becoming seriously thrombopenic, three parameters - presence of abdominal pain, liver function disturbance and fever - are strongly associated with the occurrence of heavy cares during the course of the hospitalisation that followed the evacuation. The selected multivariate model, provides a scoring classifier; its excellent accuracy was characterised by ROC analysis: AUC= 0.899.
The survey allowed reconsidering the value of clinical features (abdominal pain and fever) besides the ackwnowlegment of the severity of the thrombopenia. The availibility of a prognosis tool implies an assessment regarding the conditions of its use. Other data (older of from other asian regions) will worth being included.
