Contraceptive failure rates: new estimates from the 1995 National Survey of Family Growth

Abstract

Context: Unintended pregnancy remains a major public health concern in the United States. Information on pregnancy rates among contraceptive users is needed to guide medical professionals' recommendations and individuals' choices of contraceptive methods.

Methods: Data were taken from the 1995 National Survey of Family Growth (NSFG) and the 1994-1995 Abortion Patient Survey (APS). Hazards models were used to estimate method-specific contraceptive failure rates during the first six months and during the first year of contraceptive use for all U.S. women. In addition, rates were corrected to take into account the underreporting of induced abortion in the NSFG. Corrected 12-month failure rates were also estimated for subgroups of women by age, union status, poverty level, race or ethnicity, and religion.

Results: When contraceptive methods are ranked by effectiveness over the first 12 months of use (corrected for abortion underreporting), the implant and injectables have the lowest failure rates (2-3%), followed by the pill (8%), the diaphragm and the cervical cap (12%), the male condom (14%), periodic abstinence (21%), withdrawal (24%) and spermicides (26%). In general, failure rates are highest among cohabiting and other unmarried women, among those with an annual family income below 200% of the federal poverty level, among black and Hispanic women, among adolescents and among women in their 20s. For example, adolescent women who are not married but are cohabiting experience a failure rate of about 31% in the first year of contraceptive use, while the 12-month failure rate among married women aged 30 and older is only 7%. Black women have a contraceptive failure rate of about 19%, and this rate does not vary by family income; in contrast, overall 12-month rates are lower among Hispanic women (15%) and white women (10%), but vary by income, with poorer women having substantially greater failure rates than more affluent women.

Conclusions: Levels of contraceptive failure vary widely by method, as well as by personal and background characteristics. Income's strong influence on contraceptive failure suggests that access barriers and the general disadvantage associated with poverty seriously impede effective contraceptive practice in the United States.

PIP: This study estimated method-specific contraceptive failure rates in the US. Estimates were adjusted for underreporting of induced abortion in the main survey. The correction made a sizeable impact, as 25% of the 2,157,473 conceptions due to contraceptive failure were aborted. Data were obtained from the 1995 National Survey of Family Growth and the 1994-95 Abortion Patient Survey. Analysis was based on hazard models for failure in the first 6 and 12 months. Data include 7276 contraceptive use segments. The mean duration was 9.6 months. The pill and condom had the largest shares of use segments. The lowest failure rates were for implants and injectables (2-3%). Failure rates were as follows: oral pills (8%), diaphragm and cervical cap (12%), male condom (14%), periodic abstinence (21%), withdrawal (24%), and spermicides (26%). Failure rates were highest among cohabiting and other unmarried women; women with an annual family income below 200% of the federal poverty level; among Black and Hispanic women; and among adolescents and women in their 20s. The failure rate among low income women declined during 1988-95. Women above the 200% of poverty level had stable rates. Poverty continued to have a negative impact on effective contraceptive use. Four models were used to examine the effects of socioeconomic factors on contraceptive failure.