The field of text mining concerns itself on developing techniques to extract relevant information from textual data. The goal is to distil the text content into structured format that is readily tractable by computers, providing means for fast access to knowledge, and ultimately to automatic inference over the data. Currently, the most successful techniques employ statistical parsers to analyse the text, and machine learning techniques to identify semantic patterns. The full development of this technology can open the way to applications such as the discovery of new medical knowledge from the bulk of published material. In this research we will focus on the area of systematic review (SR) generation for evidence-based medicine (EBM). The goal of these medical reviews is to provide in a summarised way the latest evidence about relevant health related issues. For instance, the effect of applying intubation for trauma injury. Using keyword-based search for these kinds of questions usually returns thousands of publications, and many relevant studies may be missing from the top listed. To tackle this problem, there are different international initiatives that compile the information for relevant medical questions by reviewing the literature periodically. However, due to the pace of medical research, with new trials published currently at a rate of 15,000 a year, this is a formidable task that leaves many relevant questions out. Another problem is the need for updates, which need to be done periodically. These issues make text mining applications needed in this area to make a qualitative gap. Our strategy will be to develop resources and techniques to automatically aid the development of SRs. This effort aligns with the Information Extraction and Semantic Inference tasks that are one of the main goals of the IXA group. We have also foreseen the collaboration with medical experts and the global SR initiatives.