4.1 Main Findings
Our findings highlighted an important possible public health benefits for women that may accrue by moderating the positive and negative aspects of medical abortion: reducing the incidence of it. Firstly, both univariate and multivariate results confirmed a robust reverse causality between MA and EA, AFS, as well as a forward causality between MA and smoking, LNSP. Secondly, the inverse relationship between HI and MA was no longer observed in the MVMR analyses after EA or AFS and LNSP after accounting for EA, or AFS and LNSP, which may be interpreted as a stronger effect of EA or AFS and LNSP on MA than that of HI on MA. Thirdly, AFS and AFB mediated EA to reduce the risk of abortion.