Contraceptive failure rates based on the 1988 NSFG

Fam Plann Perspect. Jan-Feb 1992;24(1):12-9.


Analysis of data from the 1988 National Survey of Family Growth--corrected for the underreporting of abortion--reveals that contraceptive failure during the first year of use remains a serious problem in the United States, contributing substantially to unintended pregnancy. The pill continues to be the most effective reversible method for which data were available (8% of users accidentally became pregnant during the first year of use), followed by the condom (15%). Periodic abstinence is the method most likely to fail (26%), but accidental pregnancy is also relatively common among women using spermicides (25%). Failure rates vary more by user characteristics such as age, marital status and poverty status than by method, suggesting the extent to which failure results from improper and irregular use rather than from the inherent limitations of the method.

PIP: Use-effectiveness of contraceptives during the 1st 12 months of use is measured in this study based on respondents' observation period of January 1984-88 in the 1988 National Survey of Family Growth (NSFG)--IV. Best fitting models that estimate the proportions of contraceptive failures (method used and duration) among subgroups of women were determined by log linear hazard analysis. Correction procedures were performed to compensate for underreporting of abortion using the 1987 Alan Guttmacher Institute (AGI) survey data. The effect was substantial and created difficulties. The limitations of the correction procedure are that the time frame is not exactly the same between the AGI and the NSFG, and few adjustments have been made for the changes in contraceptive practice since 1984. Abortions may have been concealed in the reporting to the NSFG, and the failure rates may be biased downward. The results are topically discussed in terms of method-specific failure rates, the effects of characteristics, and the comparison between 1982 and 1988 results of the NSFG. One of the conclusions reached was that US contraceptive failure rates are much higher than necessary. For example, women who used the pill in the 1980s had a failure rate of 80/1000, when it should be 1/1000. Perfect use of condoms yields an estimated failure of 20/1000 vs. the actual rate of 150/1000. The comparison of 1982 and 1988 data that contraceptive failure rates for the pill and periodic abstinence have risen. The changes in user characteristics do not account very much for the differences in the rates. Plausible interpretations are that the 1982 periodic abstinence failure rates may be too low, and that unknown pill user bias accounts for the difference. Failure rates by method show that oral contraceptive users are the least likely to experience an unintended pregnancy, followed by condom and diaphragm users. The highest failure rates were found among users of periodic abstinence and spermicides. In the analysis of characteristics, failure rates in general are more common among disadvantaged women. Those most likely to experience contraceptive failure are younger unmarried women who are poor and from a racial or ethnic minority. The 1988 results differ in that poverty has a greater impact than race or ethnicity, particularly among fundamentalist Protestants. Other non-Catholic poor had the highest rates. Effective use needs reinforcement through education and counseling, particularly among young couples. Individual factors which affect contraceptive failure need to be adequately researched. Contraceptive research and development need to focus more on contraceptive options for differing stages of reproductive lives. Better abortion reporting is needed.

Publication types

  • Research Support, U.S. Gov't, P.H.S.

MeSH terms

  • Abortion, Induced / statistics & numerical data*
  • Adolescent
  • Adult
  • Cross-Sectional Studies
  • Female
  • Health Knowledge, Attitudes, Practice*
  • Humans
  • Incidence
  • Infant, Newborn
  • Pregnancy
  • Pregnancy, Unwanted / statistics & numerical data*
  • United States / epidemiology