Ok, so we just looked at some simple ways to find case materials, in the previous post. But consider the following situation: how will you find a case on, say, non-ischemic chest pain? If you use the simple search tools already, you probably won’t find anything, unless the author happened to use that same phrase in the case description.

Furthermore, what search terms will you use for this? Chest pain covers a variety of different diagnoses and ICD9 codes – you might be looking for chest wall pain, pleurisy, pericarditis, Bornholm’s costochondritis, pneumonia, fractured ribs, or just a case that covers some of these in its differential. The number of synonyms makes things tough… and if the author has used a different term, such as ‘pleurodynia’, to describe their case, you will be out of luck.

We have been working on semantic search tools for OpenLabyrinth. (We are nearly there and hope to have a test release soon). With semantic search, it is more able to understand the concept you are searching for, rather than the exact phrase. A semantic search engine, along with semantic indexing of the case materials, allows you to search for the concept of chest pain, and has access to predefined vocabularies so that it can group these various symptoms along with chest pain. This should pull up a much bigger list of relevant cases.

Because the semantic indexing is applied to all the nodes within the case, you are also not dependent on the case author having inserted all the right metadata about the case in the description or title or keywords.

We will post more about this when our semantic indexing tools are online. In the meantime, if you have comments or suggestions about this, please contact us.