Sepsis is a transversal pathology and one of the main causes of death in the Intensive Care Unit (ICU). It has in fact become the tenth most common cause of death in western societies. Its mortality rates can reach up to 60% for Septic Shock, its most acute manifestation. For these reasons, the prediction of the mortality caused by Sepsis is an open and relevant medical research challenge. This problem requires prediction methods that are robust and accurate, but also readily interpretable. This is paramount if they are to be used in the demanding context of real-time decision making at the ICU. In this brief contribution, three different methods are presented. One is based on a variant of the well-known support vector machine (SVM) model and provides and automated ranking of relevance of the mortality predictors while the other two are based on logistic-regression and logistic regression over latent Factors. The reported results show that the methods presented outperform in terms of accuracy alternative techniques currently in use in clinical settings, while simultaneously assessing the relative impact of individual pathology indicators.
TopBackground
The incidence of Sepsis and its associated complications: Septic Shock and MODS are difficult to establish and their causes are multiple and varied: longevity, associated pathologies (diabetes mellitus, hepatic cyrrhosis, neoplasias, chronic renal insufficiency, and so on), the inreased use of invasive techniques, corticoid administration, chemotherapy and immunosupressants, organ transplants, and so on (Luce J., 1987).
Studies carried out on the eighties probably underestimated the real incidence of Sepsis. In 1990 the US Center for Disease Control (CDC) calculated that between 1979 and 1987 there were 450,000 cases of Sepsis causing the death of 100,000 people. The incidence of Sepsis increased from 73.6 cases/100,000 people patients in 1979 to 175.9 cases/100.000 people in 1989 (CDC, 1990).
Now it is commonly accepted that the incidence of Sepsis is much higher. Angus et al. (Angus DC., 2001) describes an incidence of 3 cases / 1000 people, which implies that in the US there may appear 750,000 cases of Sepsis per year out which 51.1% will require ICU admission, with a hospital mortality rate of 28.7% (resulting in more than 215,000 deaths/year). These figures are similar to those of secondary deaths of acute myocardial infarction. It is expected that this incidence will increase 1.5% each year due to increased longevity, more aggressive treatment and the increased number of patients taking immunosuppressants. Therefore, the expected incidence for the years 2010-2020 is 934,000-1,110,000 cases. Besides this study, Martin et al. (Martin GS., 2000) studied the incidence of Sepsis during 1979-2000 in the US and found out that the number of Septic patients augmented from 164,072 in 1979 to 659,935 in 2000 (i.e. an increase of 13.7% each year). The reported incidence of Sepsis increased from 82.7 cases / 100,000 people to 240.4 cases / 100,000 people (annual increase of 8.7%).