Strengths and Limitations
The main limitation of our study is its retrospective observational design in the second part of the study, which does not account for all potential co-founders. Secondly, despite incorporated automatic checkpoints, the computerized database is also prone to human errors. Thirdly, infertile women were heterogenous regarding indication for reproductive surgery which might influence the insignificant difference in preterm delivery in regard to the mode of conception. The main strength of our study is a novel approach to exploring the effect of fertility treatment on preterm delivery as a treatment-related adverse outcome. Using PS matching approach allowed us to balance the treated and untreated group for some of the main co-founders and also for the number of co-existing possible risk factors for preterm delivery per woman making the observational study close to a randomized study. Therefore, the PS matching approach enables higher control over cofounding, reduces bias, and adds to higher prediction estimation20-21. However, the number of outcomes in our study was low, especially of preterm deliveries < 32 weeks, which might influence the precision of estimation. For that reason, the result should be treated with caution.