Title : Detecting Evidence for Guideline Updates Using Topic-centric Method

Presenter Qing Hu
Abstract Ideally, a guideline should be updated immediately after new relevant evidence is published, so that the updated guideline can serve medical practice using latest medical research evidence. However, the update of the guideline is often lagging behind medical scientific publications. Therefore, we propose a topic-centric method to detect evidences for guideline update. We selected two versions of the Dutch Breast Cancer Guidelines (2004 and 2012), and checked if the new evidence items in the 2012 version could be found by using our method. The experiment shows that our method can not only find at least some evidence for 12 out of the 16 guideline statements in our experiment, but it also returns reasonably small numbers of evidence candidates with an acceptable real-time performance.

Title : The impact of advanced preprocessing methods for the prediction of cancer

Presenter Reinier Kop
Abstract I will discuss various methods of preprocessing data and their impact on the performance of standard data mining algorithms. The methods are explored in the context of predicting the occurrence of colorectal cancer using a medical dataset containing 500.000 Dutch citizens. The discussed preprocessing methods include simple aggregation, the analysis of trends, temporal pattern mining, and semantic enrichment.