关键词: 放射科医生 专家 新手 线索提示 决策 眼动

Abstract: With the increasingly important role of modern medical photography in medical diagnosis, the significance of radiologists’ decision making is also growing. However, studies have revealed a high rate of false negative in image interpretation. About 40 percent of radiological examinations are operated on chest with 26% false-negative errors. False negatives do harm to the health of the patients, lead to medical disputes and even causes significant medical negligence. Applied prospectively, double reading is used for quality assurance of current radiology reports. With this method the results of two separate observers are combined to yield one result. It is considered negative if both results are negative and positive if at least one of the individual results is positive. Many studies have found that double reading can reduce errors and increase sensitivity though it may be at the expense of specificity. In China, double reading is also applied in radiology departments in major hospitals, but conducted differently. The specific process begins with the interpretation of a patient’s image by a doctor, who makes a diagnosis and completes the writing of a diagnostic report. The second doctor then signs the image report. The problem of this conduct is that the first doctor’s diagnosis can be a clue to the second doctor. There were arguments about the impact of clue on the radiologists’ decision making. Some studies have shown that the availability of clue increased diagnostic accuracy, while others found no significant increase in accuracy but distraction. The expectation that clue (other’s impact) can increase accuracy has not been confirmed so far. The present study aimed to investigate the impact of clue on different radiologists’ decision-making. Which factor affected the decision-making by radiologists mostly, their expertise or other’s impact on them? We tested the hypotheses in a single experiment by using eye-movement technique. Experts and novices were selected as experimental objects. Their expertise was set as non-variables, while clue words were adopted to control others’ impact. The X-ray was adopted as experimental material with the task of detecting pulmonary nodule. The SMI250 Eye Tracker was selected as our experimental instrument.15 experts and 19 novices participated in the experiment to explore the decision making between experts and novices in different cue categories (yes or no), different size (big or small) and location (concealed or non-concealed) of nodules. The major findings of the present study were: (1) For experts, the false negatives rate of small nodule in concealed position with clues was lower than without clues, while, for novices, the false negatives rate of all sizes of nodule with clues was lower than that without clues. (2) It appears that experts, with more experience, were less affected by clues than novices. (3) With the help of clues, both times of fixation and total fixation time increased in experts and novices. (4) The times of re-fixation indicated that novices spare more cognitive resources on processing of clue words. On the basis of these findings, we concluded that clue made an important positive influence on radiologists’ decision-making. On the other hand, experience and other’s impact played a joint role in radiologists’ decision-makings, while other’s impacts played a bigger role in the novices’ decision-making process.Findings of the present study provided the first evidence that other’s impact may cause the second doctor to have preconceived notions about the results, biasing the overall interpretation. We proposed that radiologists should review radiographs without knowledge of the clininal scenario ing to avoid focusing on the expected findings.

Key words: radiologist, expert, novice, clue, decision-making, eye movement