With two years to go, prospects for meeting a key reproductive health goal of the federal government’s Healthy People 2010 initiative seem dim. An analysis of data from the last two cycles of the National Survey of Family Growth (NSFG) reveals that women’s probability of experiencing a contraceptive failure within the first year of use of a reversible method was essentially unchanged between 1995 and 2002; the 2002 figure—12%—is considerably above the Healthy People target of 7%. Contraceptive failure varies by method and by women’s characteristics, but the socioeconomic characteristics associated with relatively high probabilities of failure are not the same for all methods.1

Two types of analysis were used to assess contraceptive failure. The researchers first estimated life-table probabilities of failure at different durations of use for each of the five most commonly used contraceptives in the United States (injectables, pills, male condoms, withdrawal and fertility awareness) and for women with various characteristics. In multivariate analyses, they then examined relative risks of pill, condom and withdrawal failure for women of different subgroups. Data were corrected for underreporting of abortion in the NSFG, which is known to be substantial and likely would lead to underestimates of contraceptive failure.

The life-table calculations reveal that 4% of women who begin using any contraceptive experience a failure within three months; the proportion rises to 7% at six months and to 12% at one year. The injectable and the pill have the lowest one-year failure rates (7% and 9%, respectively), and methods based on fertility awareness the highest (25%); male condoms and withdrawal have intermediate probabilities of failure (17% and 18%, respectively). Comparison of these 2002 data with data from the 1995 survey reveals no significant changes in either overall or method-specific failure rates.

Women in their 20s are significantly more likely than those aged 30 or older to experience a contraceptive failure during their first year of use (14–15% vs. 8%), blacks are more likely than whites to do so (21% vs. 10%) and cohabiting women have a greater probability than married women of experiencing a failure (22% vs. 10%). The probability of failure is significantly higher among women who intend to have more children or are not sure, those who have already given birth and those living below 200% of the poverty line (14–20%) than among their counterparts who intend no more births, are nulliparous and are better-off (6–9%).

Different predictors of failure emerge for the three methods studied in the multivariate analysis. The risk of pill failure is significantly elevated for women younger than 30, those who intend to have more children or are unsure, those who are cohabiting or not in a union (relative risks, 1.7–2.5) and, especially, those who have given birth (7.9). A heightened risk of condom failure is predicted by being younger than 30, intending more births or being unsure, cohabiting, being parous, being black and living at less than 200% of the poverty level (1.5–2.5). The risk of experiencing a failure during use of withdrawal is increased among women who intend more children or are not sure, have never married, have given birth or live below 200% of the poverty line (1.9–4.2).

Noting that “the effectiveness of a method greatly depends on vigilance and effective use by the woman and her partner,” the analysts infer that many women, particularly those in some socially and economically disadvantaged subgroups, are unable to overcome difficulties using methods. As a result, they conclude that reaching the Healthy People goal of reducing the contraceptive failure rate to 7% and lowering failure rates among disadvantaged groups by 2010 “will be very difficult in the absence of major policy and programmatic interventions.” Their recommendations include simultaneous efforts to increase contraceptive education and access, and to improve couples’ communication about contraceptive use.


1. Kost K et al., Estimates of contraceptive failure from the 2002 National Survey of Family Growth, Contraception, 2008, 77(1):10–21.