Abstract
There are two potential limitations worth mentioning. The first is that the authors do not discuss how the bladder neck was treated by the five surgeons. Was the bladder neck anatomically spared? Bladder neck reconstruction does not exactly capture this question, because some bladder necks that are not spared anatomically still do not need formal reconstruction. While the literature is still divided, many believe bladder neck sparing to be an important contributor to early continence. To add further confusion, two experienced surgeons may watch the same bladder neck sparing dissection and disagree as to whether the bladder neck was truly spared. Each surgeon seems to conceive this anatomy slightly differently. This question of how the bladder neck was handled simply highlights that we cannot adequately study the effect of unmeasured covariates.
The second and more important issue is a statistical problem. The authors have ignored the key role of propensity scoring to remove bias in observational studies. We generally find comfort in large numbers and multivariate logistic regression. Even if surgical experience was a variable, however, this does not remove the potentially significant bias created by the fact that three surgeons placed suprapubic tubes and two did not. It is possible, for example, that those three are the more experienced or that those three surgeons treat the bladder neck differently, or some other as yet unidentified surgeon factor. In other words, the impact of the unstented anastomosis may have more to do with reasons why the patient got the suprapubic tube (they happened to have a particular surgeon, for instance), rather than the tube itself.
The bottom line is that propensity scoring adjusts for selection bias and confounding. Without it, you have been tempted to conclude that there is a causal effect between an unstented anastomosis and urinary continence, which may or may not be true. In this observational study, patients were not randomized to have a suprapubic tube or not, because this was not the point of the study. One can think of propensity scoring as almost simulating the outcome of randomization between suprapubic tube or conventional Foley catheter, and only with propensity score matching will biases be uncovered.