Using a Clustering Technique to Develop a Usable Information System

This study describes the Delphi method and clustering technique used to develop a usable medical information system. To assist in the development of new information architecture, we modified an approach developed at the Rand Corporation for polling qualitative expert opinion and using information gained early in the process to later move towards a quantitative group consensus. Our approach can help discover the mental constructs of groups of typical site users as well as determine where they expect to find particular types of content. Knowing where users expect to find things provides validation for development of intuitive information architecture and supporting design. The particular strength of the insights provided by the clustering technique is that these insights were generated in a systematic way by experts. This systematic process may lead to a more usable information system