The findings are subject to selection bias and confounding, which will have been minimised but not eliminated by propensity score matching.

• 该发现是受选择偏倚和混杂设计的影响，这些可以通过倾向评分匹配最大限度的降低但不能消除。