Reply to the Letter to the Editor: “Impact of clinical information on CT diagnosis by radiologist and subsequent clinical management by physician in acute abdominal pain”

by Hattori S, Yokota H, Takada T, Horikoshi T, Takishima H, Mikami W, Mukai H, Funakoshi H, Naganuma T, Yano T, Terada K, Inokuchi R, Tsuchiya S, Akutsu A, Saiga A, Kubota Y, Wada T, Takada A, Fujimoto H, Uno T (hjmykt@chiba-u.jp)

Impact of clinical information on CT diagnosis by radiologist and subsequent clinical management by physician in acute abdominal pain

Dear Editor,

We appreciate the letter providing insightful opinions on our published manuscript.

“It would be interesting for your study to perform an intraobserver kappa analysis after a prudential time to avoid memory and recall bias and evaluate if there is a change in diagnosis and lecture times for each radiologist.”

Response:
It is difficult to exclude recall bias completely. We often remember impressive cases even after several years. We prioritized removing recall bias and adopted the study design that the same radiologists and physicians did not evaluate the same patient with or without clinical information.
Various factors (e.g., degree of fatigue, the environment of interpretation) affect interpretation time. Therefore, it is unclear how reliable to compare the interpretation time. We intended to emphasize that the interpretation time was not extended if the radiologist referred to the clinical information.

“It could also be essential to know if having clinical information impacts incidental findings that could be relevant in the clinical context of the patient given satisfaction bias. For example, in the study by Littlefair et al., they demonstrated that having previous expectations before lectures can impact the diagnosis efficiency and assign false positives in regular studies [2].”

Response:
It is not easy to obtain an accurate diagnosis for all acute abdomen. For example, enteritis does not have disease-specific markers and is diagnosed based on medical history and symptoms. Therefore, the board chose the more appropriate diagnostic report.

“Lastly, a subgroup analysis could be performed depending on the type of study, with or without contrast, since some diagnosis is harder to make in a simple CT. Having the appropriate clinical information and suspected diagnosis could lead to more guided studies to better address the clinical inquiry.”

Response:
We did perform subgroup analyses. Please refer to the supplementary information on the website.