Expert-Novice Differences in Mammogram Interpretation

This paper examines the results of two initial studies of the problemsolving strategies used by more and less skilled medical professionals during mammogram interpretation. The first study examined the cognitive processing of staff radiologists and radiology residents, while the second looked at surgical residents and medical students‚ as they individually solved a set of breast disease cases. Analyses of 100 verbal protocols from the two studies resulted in the development of a problem-solving model of mammogram interpretation and a characterization of novice expert differences based on performance measures. Results revealed that with increasing levels of expertise there were significant increases in the number of radiological observations and findings, proportion of correct diagnoses, use of data-driven problem solving, and diagnostic planning. The analysis provides a valuable initial characterization of mammogram interpretation across a broad range of expertise levels with implications for the design of computerbased learning environments aimed to train medical professionals to interpret mammograms.