**7.3 Mechanism of action**

Barrier methods provide a physical barrier to prevent the migration of sperm from the vagina to the upper reproductive tract where fertilization occurs, thereby preventing pregnancy. Male and female condoms work by keeping sperm out of the vagina and forming a barrier to ejaculate, pre-ejaculate, and vaginocervical secretions [78]. In turn, diaphragms and caps keep sperm from reaching the cervix. However, as only the cervix is covered by the latter two methods, they do not prevent exposure of the vaginal mucosa to semen or exposure of the penis to vaginocervical secretions.

#### **7.4 Effectiveness**

Data from the U.S. suggest that there is a 5% failure rate with the female condom and a 2% failure rate with the male condom with perfect use (i.e., correct and consistent use). With typical use (which includes incorrect and inconsistent use), failure rates are 21% and 18%, respectively [88]. Nevertheless, many factors other than user error may influence the efficacy of condoms in the prevention of pregnancy, including background fertility, coital frequency, or the use of emergency contraception. A study assessing semen exposure following condom failure suggested that even when condoms break or slip, the risk of pregnancy may be reduced in comparison to when using no method of contraception at all [164]. Although condom studies often report clinical breakage and slippage rates, these are not considered valid surrogate endpoints of pregnancy [165].

Data have suggested that, with perfect use, 4.3–8.4% of women using a diaphragm with a spermicidal cream or jelly experience an unintended pregnancy within the first year of use. With typical use, the percentage increases to 12% [88]. In a comparative study, the only contraceptive cap available in the UK was found to be less effective at preventing pregnancy than the diaphragm to which it was compared. The unadjusted typical-use probability of pregnancy at 6 months of use was 13.5% for contraceptive cap users and 7.9% for diaphragms users, with the adjusted risk of pregnancy being 1.96 times higher in the former group [166, 167].
