Advancing Sexual and Reproductive Health and Rights
Family Planning Perspectives
Volume 32, Number 1, January/February 2000

Likelihood That a Condom Will Break or Slip Off Is at Least Partly Related to User's Characteristics

While having a condom break or slip off during intercourse is often the result of design and manufacturing flaws, the user's experience with the method and other characteristics also play a part, according to a large prospective follow-up study.1 Women who are young, black, single or childless and those who engage in risky sex are more likely than others to have condoms break, while women with children have heightened odds of experiencing slippage. Those who have experience with condoms and those who use spermicides as well are less likely than others to have problems with condoms.

Study participants were recruited from among 18-34-year-old nonpregnant women attending a county sexually transmitted diseases clinic in Birmingham, Alabama. Nurses provided them with information on the correct use of barrier methods and recommended use of a condom and spermicide at each act of intercourse or, if that is not possible, use of a condom alone. The participants were given free barrier contraceptive supplies and diaries for recording information on their sexual activity, contraceptive use and any problems that occurred with their method. At six monthly follow-up visits, the women reviewed the diaries with a nurse.

A total of 1,122 women enrolled in the study; the analyses are based on the 892 who reported condom use. Most of these women were young adults (their median age was 24), 90% were black, half had more than a high school education and the majority were poor. On average, they had first had intercourse at age 16, their median number of partners was six and half had used barrier contraceptives. Seventy percent had had a sexually transmitted disease.

Participants' diaries recorded 34,036 instances of vaginal intercourse; latex condoms were used in 64% of the encounters, and in 33% of those cases, women also reported use of a spermicide. In all, 2% of condoms broke during intercourse, and 1% slipped (the study did not differentiate between partial and complete slippage). Other problems with condoms (e.g., they were not used throughout intercourse) occurred in fewer than 1% of acts of intercourse. Rates of breakage and slippage varied according to factors measured both at baseline (e.g., women's race, parity and education) and within the follow-up period (e.g., previous condom failure, type of partner and use of spermicide).

The researchers used logistic regression analysis to compute the odds of breakage and slippage associated with various characteristics. Results of that analysis showed that baseline characteristics had a modest effect on the likelihood of condom breakage. Women younger than 30, black women and participants who had engaged in risky sex (defined as exchanging sex for money or drugs or having sex while under the influence of drugs or alcohol) had an elevated risk of experiencing breakage (odds ratios, 1.6-2.0), while women with children were less likely than childless women to have a condom break (odds ratio, 0.7).

Condom breakage was more dramatically affected by a number of factors measured during follow-up; the strongest predictor of breakage was a woman's experience with condoms. If a woman had had one condom break, the chance that she would have another one break was elevated (odds ratio, 3.6); once she had had two or more condoms break, her odds of experiencing another break were markedly higher (9.3). While slippage was not associated with the likelihood of a condom's breaking, women who had other problems with the method had significantly increased odds of experiencing breakage (2.8).

Participants who had used condoms without problems fewer than five times were substantially more likely to have a condom break than were women who had used condoms more than 30 times without problems (odds ratio, 6.5). Breakage was more common among women who had had unprotected intercourse fewer than five times than among those who had had unprotected coitus 30 or more times (odds ratio, 2.2). The risk of having a condom break tended to increase with a woman's number of partners, though the association was weak and inconsistent. Use of a spermicide had a slightly protective effect on breakage (odds ratio, 0.8).

Condom slippage followed a different pattern. Women with children were somewhat more likely than childless women to experience slippage (odds ratio, 1.4), but no other baseline characteristic had a significant effect. A number of follow-up factors roughly doubled the odds of slippage: having 6-10 partners, having a new or casual partner and having had a condom break (odds ratios, 1.5-1.9). Women who had had a condom slip in the past were 3.0 times as likely as those who had not to report slippage; those who had previously experienced slippage at least twice were at even greater risk (7.4). Again, as the number of times a woman had used condoms without problems decreased, her odds of having a condom slip rose (2.0-10.3). Users of vaginal spermicides had slightly reduced odds of experiencing condom slippage (0.8). Associations between slippage and number of partners and number of unprotected acts of intercourse were inconsistent.

Noting that the rates of slippage and breakage they found are low, the researchers nonetheless urge that women at high risk of contracting HIV from their partner exercise caution, and they encourage health providers to recommend abstinence over condom use for such individuals. Moreover, the researchers emphasize that their results show condom breakage and slippage, which have generally been believed to be a result of product defects, to be at least partly related to users' characteristics. Consequently, they recommend that future research and interventions focus on "the modifiable characteristics of the user" and on improving "user motivation and training."--M.L. O'Connor


1. Macaluso M et al., Mechanical failure of the latex condom in a cohort of women at high STD risk, Sexually Transmitted Diseases, 1999, 26(8):450-458.