To the Editor In their carefully conducted and well-reported study, Sarvepalli et al¹ conclude that prior reports of lower polyp detection rates among nongastroenterologists were flawed by residual confounding. However, their study may be flawed by the opposite error: overadjusting for factors that operate as part of the causal pathway linking endoscopist characteristics with polyp detection. Specifically, eAppendix 1 in the Supplement lists preparation adequacy (which could be a function of quality of patient education delivered by different clinician types), cecal intubation, and withdrawal time as control variables. Prior studies have noted associations between each of these factors and polyp detection,^2-4 and it is not unreasonable to assume that any or all would be associated with endoscopist specialty. It would be interesting to examine the results with these 3 variables omitted.

致编辑在严谨的报告中，Sarvepalli等人[ ^1]得出结论，先前关于非消化内科医师息肉检出率较低的报道存在残留混杂因素，这是有缺陷的。但是，他们的研究可能因相反的错误而存在缺陷：过度调整作为内窥镜特征与息肉检测联系起来的因果关系的因素。具体来说，补编中的eAppendix 1将准备充分性（可能是不同临床医生类型提供的患者教育质量的函数），盲肠插管和停药时间作为控制变量。先前的研究已经注意到这些因素中的每一个与息肉检测之间的关联，^{2 -4}并假设任何或所有与内镜专家相关联是不合理的。在省略这3个变量的情况下检查结果将很有趣。