To the Editor, We thank Sun M et al for their commentary on our original study assessing the burden of unwanted e-mails (UEM) in academic surgery1. They made four comments that we will try to address hereunder. First, Sun M et al stipulated that UEM count may not be an appropriate endpoint as it may not be able to capture the impact of UEM, suggesting that UEM exposure should be linked to “objective behavioral indicators and validated measures of occupational well-being to determine whether these messages constitute a clinically relevant and modifiable risk.” (1) The absolute number of UEM received by participants was naturally defined as the primary endpoint since the study primarily aimed to assess the burden of UEM among surgeons2. (2) The reference by Qiao T et al is not precisely applicable here, as it investigated the impact of review on readers’ evaluation and rating of hotels – neither on UEM nor on healthcare3. (3) We agree that UEM count is an imperfect metric, but it is easy to understand, objective, and reproducible2; the idea of using behavioral indicators as well as validated measures of well-being sounds appealing, and future studies using such endpoints – if there are any – would be valuable to fill the gaps. (4) Finally, we are not sure we understand the statement on “whether messages constitute a clinically relevant and modifiable risk”; the stake on UEM is not only about clinical relevance. Second, our study also aimed to identify factors associated with a high burden of UEM. Sun M et al stated that other important variables were not taken into account in our analysis, such as editorial responsibilities, professional society roles, or administrative duties. We definitely agree with them. However, one must be cautious about the high risk of collinearity, which is the reason why we did not integrate these variables into our model. The third comment was in regard to the heterogeneity of UEM, suggesting that analyzing UEM as a single entity would be misleading. “A structured, risk-stratified classification system that distinguishes credibility, financial risk, and potential scientific harm would allow more precise filtering policies, faculty training, and research-integrity safeguards.” This is absolutely correct but overoptimistically feasible. The last comment was related to the length of the study, claiming that 1 month was insufficient. Assessment over a longer period would be an obvious advantage; we agree with this comment and have already addressed this limitation in the discussion. In summary, we thank Sun M et al again for their comments and hope our reply has helped clarify certain aspects. With this study, we took the side of a pragmatic stepwise approach over theoretical principles, aiming to design the perfect study because theory needed to meet reality.
Bernardi et al. (2026) studied this question.