Typical-use failure rates describe the proportion of individuals expected to become
pregnant should they use a contraceptive for a period of time such as one year. This
study produced new estimates of contraceptive failure for more methods than past studies,
and, for the first time, combinations of methods (“dual-use”).
Estimates were based on data from the 2002, 2006–2010 and 2011–2015 National Surveys
of Family Growth and the 2000, 2008 and 2014 Abortion Patient Surveys. We developed
a Bayesian hierarchical hazard model to estimate how the odds of pregnancy vary by
method, year, and duration of method use. Individuals were considered at risk of contraceptive
failure during months in which they used a method and had intercourse.
One-year failure rates ranged from an estimated 2.1% for the intrauterine device (IUD)
to 13.4% for withdrawal. Estimates were 7.1% for the pill, 8.5% for the ring, 9.5%
for condoms, 10.2% for the patch, and 11.8% for calendar methods, when these methods
were used by themselves. Dual use substantially reduced failure rates for users of
the pill, to 5.3–5.4%, and for users of calendar methods, to 6.9–8.4%, depending on
whether they used withdrawal or condoms.
These findings are in line with previous studies in their overall ranking of methods
by effectiveness. Additionally, these new analyses find smaller differences in effectiveness
between hormonal and coital-dependent methods compared with previous studies. In the
post-Roe world, this information may become increasingly important to individuals having sex.