Abstract
The potential cost and logistic improvements of picture archive and communication system (PACS) over film-based medical image management awaits the development of viable radiology workstations (RWS) targeted for the primary interpretation task. While the quality of electronically displayed images has been highly investigated, only recently have design and experimental work been devoted to the other critical aspect of workstation design, mainly its computer human interaction, and, in particular, its navigation. By RWS navigation we include its underlying mental model or metaphor, and the commands and hand motions used to access patient folders and to display images. For the last 5 years, the University of North Carolina (UNC) Medical Image Display Research Group has analyzed the primary interpretation task and designed, developed, and evaluated the FilmPlane series of RWS prototypes. This work has helped us understand both RWS requirements and viable design approaches. In this paper, we present our workstation design strategy and our observations and understanding of the issues and problems with RWS navigation. To illustrate the discussion, we describe FilmPlane2, the UNC radiology workstation. We also briefly detail three rapid-evaluation techniques (including two observer experiments) for quickly gaining feedback on a design. These observations and evaluation techniques may aid other RWS designers in producing superior tools for the clinicians.