Bruce Jackson's Blog

One of the challenges we face in preparing medical reports is the maintenance of a data dictionary or vocabulary.  Compounding this problem is the fact that the same term can have different meanings in different contexts and to different areas in a hospital.  For example, something as simple as the term: "patient" may have a different meaning to an attending physician than to a manager in the finance department.

Even in those places and cases where the definition of a term is consistent, it may not be widely known or understood.  For example, "Length of Stay" may have a standard definition such as "the number of days the patient is counted in the midnight census."  However, this might result in a patient who is in the hospital for two days with a length of stay of only one day;  i.e., only present for one midnight census.

I mentioned to one of my associates how great the New York Times web site was because I could simply double-click any word while reading and a definition would pop up.  He suggested that something similar might help us provide context sensitive definitions to our reports.  And so began my efforts to develop a tool to bring a data dictionary to our reports.

First, I had to list the constraints and objectives for developing the tool:

1. Most of our internal web sites do not have external web access.  Therefore, the tool must be able to function without Internet access.
2. The vocabulary must be very easy to create and edit.
3. It would be nice if we could have a custom vocabulary for different sites and purposes.
4. It must run on any web page on any host web server.
5. It can't cause any latency or delays for our end users.
6. It must be able to grow without requiring much, if any, maintenance as it grows.  In other words, if we change or add a term to the vocabulary, then it must immediately show up in all of our web consumer pages without requiring page edits or other modifications.