The main purpose of this paper is to present a new approach for High-Stake Decision Support Systems. We introduce the challenge of the exploitation of heterogeneous community-based knowledge for decision-aid in critical situations. Knowledge shared in e-communities can be very rich, but is also inherently chaotic and questionable. Such uncontrolled knowledge is not usually considered in High-Stake decision support systems, which traditionally rely on expert data. In this paper, we present the new concept of High-Stake Community-Contributed Knowledge Base (HSCCKB) and propose an associated knowledge model. This model constitutes to the theoretical background for the design of HSCCKB-based decision-support systems, by defining the specific knowledge characteristics that should be considered in the reasoning process.