Review Assesses Failure Rates for Contraception Methods Available in the U.S.

April 28, 2011 — Summary of "Contraceptive Failure in the United States," Trussell, Contraception, May 2011.

When evaluating contraceptive methods, couples may find it useful to consider pregnancy rates associated with perfect and typical use, as well as pregnancy rates during imperfect use and how likely a given method is to be used perfectly. Currently, only information on pregnancy rates based on perfect and typical use of a contraceptive method is available.

James Trussell of the Office of Population Research at Princeton University and the University of Hull's Hull York Medical School in England examined contraceptive effectiveness by looking at pregnancy rates for typical and perfect use of a number of contraceptives. Trussell wrote, "Virtually all method failure rates reported in the literature have been calculated incorrectly and are too low," and thus the researchers cannot "empirically justify our estimates except for four fertility awareness-based methods, the diaphragm, the sponge, the male condom, the female condom, spermicides and methods for which there are extensive clinical trials with very low pregnancy rates."

Trussell defined typical use based on how each woman personally defined "use." For example, typical use of a condom might mean that a woman uses a condom only occasionally, or that a woman who reports taking birth control pills has not replenished her supply for several months. A method is used perfectly when it is used consistently in accordance with a specific set of directions. The review's estimates are based on data from women interviewed in the 1995 and 2002 National Survey of Family Growth, as well as from the results of related surveys and clinical trials.

Results

The study estimated that 85 of 100 sexually active couples will experience a pregnancy in the first year if they use no form of contraception. The estimate is based on populations in which contraceptive use is rare and on couples who report having used no contraception with the intention to conceive. It is also the best guess of how many women using a reversible method of contraception would become pregnant after one year if they abandoned their current birth control method but did not employ other contraceptive methods.

Trussell used the NSFG to estimate pregnancy probability for typical use of withdrawal (22%), fertility-awareness based methods (24%), the male condom (18%), the pill (9%) and Depo-Provera (6%). The estimates were corrected for underreporting of abortion, though the correction could have produced estimates that are too high because women in abortion clinics tend to overreport use of a contraceptive method at the time they became pregnant. In addition, NSFG does not distinguish between combined and progestin-only oral contraceptive pills, but because the combined method is more common, the survey for the most part reflects estimates for that method.

Pregnancy rates for perfect use of the sponge and diaphragm were based on data from two clinical trials in which women were randomly assigned to receive either method. Results from those trials indicate a difference in efficacy of the sponge between women who have never given birth and women who had given birth. According to Trussell's review, 9% of women who had never given birth and used the sponge perfectly would become pregnant in a year. Among women who had previously given birth, about 20% of perfect sponge users would become pregnant in a year. Among diaphragm users, there were no differences between pregnancy rates based on whether women had previously given birth. According to the review, 6% of women with perfect use of the diaphragm would become pregnant in a year.

The review based its pregnancy estimate for typical use of the female condom on a six-month clinical trial of the Reality female condom, which is now known as fc. There have been no efficacy trials of the newer version of the female condom -- the fc2. According to the review, about 5% of women with perfect use of the female condom would become pregnant in a year, as would 21% of those with typical use.

Trussell's estimate of pregnancy rates related to the perfect use of the withdrawal method (4%) is based on the assumption that the risk of pregnancy resulting from pre-ejaculation fluid is "modest." One study cited in the review examined pre-ejaculatory fluid samples within two minutes of production and found that 37% of subjects "produced pre-ejaculatory samples that contained motile sperm" and that the concentration and percentage of motile sperm were similar in individual's pre-ejaculatory and ejaculatory specimens. However, the study found that the number of sperm in pre-ejaculatory fluids was low.

Based on a National Institutes of Health trial of five spermicides used over a six-cycle probability, the review estimates the pregnancy rate for perfect use of spermicides to be 18%.

There is an estimated 5% chance of pregnancy among women who perfectly use the fertility awareness-based method known as Standard Days; a 4% chance among women perfectly using the TwoDay method; a 3% chance among women using the ovulation method; and a 0.4% chance among women using symptothermal method. Trussell notes, "Published 'method failure' rates for other variants of natural family planning are incorrect, because exposure includes all use, not just perfect use."

For the pregnancy rate related to perfect use of the male condom (2%), the review relied on three studies, which all reported efficacy during consistent use, but only one reported efficacy for perfect use.

The lowest reported pregnancy rate for typical use of the combined oral contraceptive pill is 0%. However, other studies suggest that pregnancies do occur "rarely" during perfect use, leading Trussell to estimate pregnancy rates among women taking the pill correctly and consistently for one year at 0.3%. The lowest reported pregnancy rate for the progestin-only pill exceeds 1%, but it is likely that the pill is less effective than combination pills during typical use, Trussell writes.

Two trials of Depo-Provera showed no reported pregnancies during perfect use. Because this form of contraception is given through injection, it is difficult to judge the typical use. Pregnancies were reported among women who received one injection but became pregnant before the next injection was scheduled, Trussell wrote. The pregnancy rate after one year of perfectly using Depo-Provera is an estimated 0.2%.

According to the review, "It is possible that the patch and ring will prove to have better efficacy than the [oral contraceptive] pill during typical use because of better adherence with the dosing schedule," though pregnancy rate estimates for typical and perfect use were equal to those for oral contraceptives (9% and 0.3%, respectively). However, current research does not yet illustrate the superior efficacy.

Pregnancy rates for the interuterine contraceptives ParaGard (0.6% for perfect use; 0.8% for typical use) and Mirena (0.2% for both perfect and typical use) were based on studies of the methods.

Estimated pregnancy rates for perfect use of male sterilization based on the weighted averages of results from nine vasectomy studies are "undoubtedly too low" because "clinicians are ... loath to publish articles describing their surgical failures and journals would be reluctant to publish an article documenting poor surgical technique," Trussell wrote. He noted that the "difference between typical-use and perfect-use pregnancy rates for vasectomy would depend on the frequency of unprotected intercourse after the procedure had been performed but before the ejaculate had been certified to be sperm-free." Trussell's research "arbitrarily" sets the chance of pregnancy one year after a vasectomy at 0.15% for typical use and 0.10% for perfect use.

There is no scope of user error for female sterilization based on pooled results from the U.S. Collaborative Review of Sterilization study of 10,685 women who had tubal sterilization. Trussell estimates that after one year of either typical or perfect use, women have a 0.5% chance of becoming pregnant.

The lactational amenorrhea method is a "highly effective," yet temporary, method of preventing pregnancy, the review said. When an infant is exclusively breastfed or given supplemental non-breastmilk or pumped milk and the women has not experienced her first postpartum menses, breastfeeding provides more than 98% protection from pregnancy in the first six months following a birth.

Conclusion

According to the review, "Pregnancy rates during perfect use reflect how effective methods can be in preventing pregnancy when used consistently and correctly according to instructions. Pregnancy rates during typical use reflect how effective methods are for the average person who does not always use methods correctly or consistently. Pregnancy rates during typical use of adherence-dependent methods generally vary widely for different groups using the same method, primarily due to differences in the propensity to use the method perfectly. Additional empirically based estimates of pregnancy rates during perfect use are needed."