We read with great interest the article, “Can prevalence expectations drive radiologists’ behavior?” by Reed et al. in the April 2014 issue of the Academic Radiology . We would like to laud Reed et al. for their excellent work, which can be an eye opener for many of us. It apparently seems that the study has the potential to be used in improving the radiologists’ vision for the target.
The authors have very nicely put the subtle difference between the effect of “increased prevalence expectation” and “increased prevalence” on radiologists’ interpretative behavior. As stated, previous studies exploring the correlation between “disease prevalence” and radiologists’ performance have shown conflicting results. On the other hand, the present study, searching for the effect of “increased prevalence expectation” on radiologists’ interpretation of the images, has shown that increased prevalence expectation prolongs image scrutiny time and decreases the radiologists’ confidence in reporting a normal image. Furthermore, the study also shows that increased prevalence expectation increases the number of false positives but by a nonsignificant amount. It should be noted that the present study and a similar previous study have been done with the expert radiologists . It would be very interesting to see what would be the effect of the prevalence phenomenon on learning radiologists’ (ie, radiology residents/postgraduate students) image interpretation behavior . As per our understanding on the subject, we predict that if similar studies are done with the learning radiologists, the effect of prevalence expectation may increase the false positives and false negatives by a significant amount. This aspect becomes much relevant in context of the developing countries such as India where much of the radiology work, especially in government departments, is performed by the residents/postgraduate students. An important scenario can be “reporting chest skiagrams in India.” It is worth mentioning here that most residents/students in India have a big background impression of high prevalence expectation for tuberculosis. Therefore, we would urge Reed et al. and other researchers to generate more knowledge on this subject so that the appropriate measures can be taken to solve this issue.
Conflicts of Interest: The authors declare that they have no conflict of interest.
References
1. Reed W.M., Chow S.L., Chew L.E., et. al.: Can prevalence expectations drive radiologists’ behavior?. Acad Radiol 2014; 21: pp. 450-456.
2. Reed W.M., Ryan J.T., McEntee M.F., et. al.: The effect of abnormality-prevalence expectation on expert observer performance and visual search. Radiology 2011; 258: pp. 938-943.
3. Monnier-Cholley L., Carrat F., Cholley B.P., et. al.: Detection of lung cancer on radiographs: receiver operating characteristic analyses of radiologists’, pulmonologists’, and anesthesiologists’ performance. Radiology 2004; 233: pp. 799-805.