Online health information portals provide valuable content to casual consumers. However, the page-oriented nature of these resources makes it difficult for users to understand the overall information space and navigate the complex relationships between various diseases. We have developed a visual analytic system named DiseaseAtlas that helps users navigate a large set of disease-related documents and understand multi-dimensional relationships for key semantic concepts such as symptoms and treatments. This paper describes several unique aspects of DiseaseAtlas and demonstrates its capabilities through a case study.