The article of "The association of gestational age at term and adverse perinatal outcomes: a retrospective study based on 43502 Chinese pregnant women" by Liu Xiaoyang and co-authors published in this issue, was based on the clinical data of 13,123 singleton full-term and natural delivery mothers from West China Second University Hospital of Sichuan University. It systematically analyzed the impact of different delivery timings within 37⁺⁰ to 41⁺⁶ weeks on neonatal outcomes, and pointed out that 39⁺⁰ to 40⁺⁶ weeks is a "window period" with a relatively low risk of adverse outcomes. This study had a clear research idea and advanced statistical methods, especially in handling the non-linear relationship between gestational age and outcomes, it adopted the restricted cubic spline model, which had high clinical reference value. However, as a retrospective study, it also had certain limitations in causal inference, confounding control, and generalizability. This article, in combination with the review of this research report, makes a summary analysis of the value and limitations of retrospective studies.