Contraceptive failure in the United States: an update

Abstract

This report provides an update of the authors' previous estimates of first-year probabilities of contraceptive failure for all methods of contraception. Estimates are provided of failure during typical use (which includes both incorrect and inconsistent use) and during perfect use (correct use at every act of intercourse). The difference between these two probabilities provides a measure of how forgiving of imperfect use each method is. These revisions are prompted by recent studies that provide the first estimates of failure during perfect use for periodic abstinence and the cervical cap, by more complete evaluations of implants, and by the appearance of the Copper T 380A and disappearance of other IUDs from the US market. Also provided is a more complete explanation of how the previous estimate of the probability of becoming pregnant while relying solely on chance should be interpreted, and this estimate is revised slightly downward.

PIP: Previous estimates of 1st-year probabilities of contraceptive failure for all contraceptive methods are updated. Estimates are given for typical use (both incorrect and inconsistent use) and perfect use (correct use at every intercourse). The difference between these 2 probabilities gives a measure of how "forgiving of imperfect use" each contraceptive method is. The revisions were done because new studies "provide the 1st estimates of failure during perfect use" for the cervical cap and periodic abstinence, more complete implant evaluations are available , and the Copper T 380A (TCu 380A) is now on the US market. Other IUDs have disappeared from the same market. The previous estimate of the probability of getting pregnant while depending only on chance should be revised downward a little. The lowest expected typical, and lowest reported failure rates during the 1st year of use of many contraceptive methods and 1st-year continuation rates are given for the IUD. The estimate for "parous sponge users" has been upwardly revised. The final estimate of spermicide rates is 3%. The efficacy of the ovulation method (OM) of periodic abstinence is quite high, but the method is very unforgiving of less than perfect use. The estimates of the probability of failure of the newer methods, of sexual abstinence, symptothermal and postovulation, have been lowered. Clinical trials of implants have shown higher efficacy. Only 2 IUDs are available in the US--the TCu 380A and Progestasert. These have greater efficacy that the IUDs that were available. 1st-year failure probabilities should be revised downward. The estimate of the proportion of women who would get pregnant during 1 year of unprotected intercourse is questioned. In the National Survey of Family Growth only 43.1% of the married women who said they didn't want to get pregnant but reported no contraceptive use, did get pregnant within the 1st year. The probability estimate could be lowered from 89% to 85%.