Optimizing radiologist workflow by customizing tools
Radiology workstations can be improved by evaluating individual preferences and work-related stress, and as a result, providing them with customized workstations, according to a presentation by Bruce Reiner, MD, of Baltimore VA Medical Center, at SIIM 2007 last week in Providence, R.I.

Reiner, who worked with colleagues at VA Baltimore on a study of workstation workflow, pointed out that when it comes to improving productivity, it is not “uniform amongst individual radiologists.” The practice of workflow engineering is able to dissect workflow at a microscopic level, including multiple tasks all at the same time. This practice is widely adopted outside of medicine.

Auditing tools can be used that provide tabulated results, structured tables, and evaluation of how a radiologist is spending time. This is important because “radiologist workflow is very heterogeneous,” said Reiner, so technology must be fitted to diverse models, rather than to workflow that is homogeneous. This is problematic because most current workstation tools are “one size fits all.”

Reiner suggested that adaptive technology must be developed, and different usage trends must be understood, while not ignoring the commonalities that exist as well. Key variables include personality, computer experience, education and training, sensory and motor skills, and intelligence.

Any changes that are considered must include the broad areas of interface design, the variety of workstation tools available, and navigation strategy. And even small components such as mice, which many view as insufficient, must come under scrutiny.

The future may very well include efforts to do radiologist workflow profiling to assemble data on each specific user which would be stored in a centralized location, said Reiner. The could be used to create user-oriented tools that would be used dynamically, and applied as needed depending on stress level, to modify tools as they are used. Moreover, this type of data could be collected in real-time and analyzed by the workstation to make adjustments in real-time in response.