For a deeper understanding of the course of events within a patient record, an analysis of the timing of events is critical. The 2012 i2b2 NLP challenge focused on the extraction of temporal relationships between concepts within hospital discharge summaries. NRC participated in task 2 of this year's challenge, where the goal was to extract correct temporal relations once temporal and medical entities have been identified in text. Our system tied for first place in the competition. It consisted of four specialist relation classifiers, each focused on a different type of link, such as links within a single sentence, or links determining whether an event occurred before or after admission. In this talk, I will describe our system, but I will also discuss some of the interesting algorithmic and evaluation issues that come up when working with temporal graphs.