Approximately half of U.S. women aged 15-24 who currently practice contraception use the pill,1 and more than 80% of all women aged 15-44 have ever used the pill.2 Many women who initiate pill use discontinue within one year. In one study, for example, more than half of women who discontinued pill use in the first six months did so within two months of starting, and the discontinuation rate was highest in the first month of use.3 Estimated discontinuation rates range from 30% to 60% at one year,4 and 20% of unintended pregnancies that occur each year in the United States may be attributable to pill discontinuation.5
Several studies have found that a male partner's influence is associated with contraceptive use among both white and black women and among adolescents.6 Twenty percent of women reported in a national survey that their partner had a lot of influence on their decision to adopt the pill.7 Many variables may be proxies for a partner's influence in contraceptive decision-making, including his education, the duration and stability of the relationship, the level of intimacy in the relationship, the extent of the partner's support, his feelings about pill use and the age difference between partners. For example, extensive research on the influence of the age difference between partners, especially among adolescents, suggests that the prevalence of contraceptive use in general is lower when the man is five or more years older than the woman.8 Other research indicates that women in longer relationships are more likely than those in shorter relationships to stay on the pill.9 Moreover, women are more likely to choose the pill and to use it consistently when they report that they are in a stable, intimate relationship and that their partner is happy with the method than otherwise.10 These measures of partner influence may indicate that a mutual commitment to the relationship is associated with a woman's decision to use and remain on the pill.
Relatively little is known, however, about how partners influence contraceptive use and about overall patterns of use among the rapidly growing population of Hispanics, who now make up 12% of the U.S. population.11 Data from the 1995 National Survey of Family Growth indicate that the proportion of pregnancies that are unintended is the same among Hispanics and non-Hispanics (49%); however, because Hispanics have more pregnancies than non-Hispanics, they have a far higher rate of unintended pregnancies (69 vs. 42 unintended pregnancies per 1,000 women aged 15-44).12 Moreover, 25% of Hispanic women aged 15-44 use the pill for contraceptive purposes.13
The purpose of this exploratory study is to evaluate the male partner's influence on early pill discontinuation in a predominantly Hispanic urban clinic population. We hypothesize that among Hispanic women, a partner's influence on the decision to use the pill is mediated by the level of commitment in the relationship. The variables under study include those identified in previous studies and others that reflect this hypothesis. The study aims to identify the variables that are associated with early pill discontinuation among Hispanic women so clients and providers can better understand the barriers to effective pill use in this population.
METHODS
Clients at a New York City family planning clinic who were first-time pill users or repeat users initiating a new segment were eligible for the prospective study. (Ongoing users who were renewing a pill prescription were ineligible.) Fewer than 10% of eligible women refused to participate; 250 women enrolled between April and September 2000.
Baseline interviews were conducted at the end of the visit in the language that the woman preferred (English or Spanish). Follow-up telephone interviews were conducted six weeks later; 91% of the women enrolled were successfully contacted. The 23 women who were lost to follow-up did not differ significantly from those who completed the study with respect to age, past pill use, language or partner characteristics. Once we excluded 14 women who discontinued use because they became pregnant or desired a pregnancy, the final sample for analysis consisted of 213 women.
In follow-up interviews, women were asked about each day of pill use from the start through day 29. The study outcome was not having started a second pack of pills—i.e., discontinuing oral contraceptive use before pill number 29. The interviews also elicited women's reasons for discontinuing use. (Most discontinuation occurs within two months of initiating use;14 because our study population typically has poor follow-up, we used an even shorter follow-up period in the hopes of limiting the rate of loss to follow-up.) Some women took their first pill at the clinic, while others were instructed to take the first pill later, a decision that was determined clinically by the provider and the client. Data regarding when and where the first pill was taken are presented elsewhere.15
The baseline interviews solicited information on demographic and social characteristics, reproductive and oral contraceptive history, fertility desires and sexual behavior, degree of acculturation and measures of influence of the male partner. The partner's influence was assessed through questions about the partner's age and fertility intention, the duration of the relationship and whether the woman thought it would last, whether the couple were cohabiting, whether the partner was aware that the woman planned to start pill use that day, how certain she was about wanting to use the pill, how strongly her partner approved of pill use and how satisfied the woman was with her partner. Responses to the last three items were scaled, but we dichotomized them for ease of analysis.
Variables that were significantly associated with discontinuation in a univariate analysis were entered into a multivariate analysis. We used logistic regression to determine which variables were independently associated with discontinuation of the pill. We performed survival analysis to identify when in the 29 days of pill use discontinuation was most likely to occur. All data were analyzed with SPSS 8.0 software. We also assessed how the partner and relationship variables were correlated by examining the chi-square statistic.
RESULTS
On average, the women were 22.4 years old (standard deviation, 5.7); they ranged in age from 13 to 46, and 42% were younger than 21 (Table 1, page 258). The vast majority (88%) were Hispanic, and 81% of these women were Dominican (not shown). Slightly more than half of women were unemployed, and a similar proportion had completed at least 12 years of school (or a GED). A minority spoke only English at home (21%) or with friends (38%). Sixty-one percent of the sample had used the pill, 70% had been pregnant and 47% had given birth.
Of the 204 women who said they had a partner, 37% lived with him, 73% said that he knew of their planned pill use and 64% were very satisfied with him. Thirty-six percent of women with a partner said that he was five or more years older than they were, and 27% reported that he wanted them to become pregnant within the next six months.
More than half of women were very certain they wanted to use the pill and intended to do so for more than one year; however, close to half said they would be happy if they became pregnant within the next six months. Twenty-seven percent of clients took their first pill in the clinic. A majority had intercourse at least once a week (64%) and had had only one partner in the past year (74%).
Overall, 48 women—23%—discontinued the pill before starting the second pack. Women stopped the pill within 29 days for user-related reasons (e.g., they forgot to take pills or were confused about how to take them) or because of side effects or the fear of side effects.
Seven women who accepted the method in the clinic nonetheless never initiated use at home. There was no association between not taking any pill and the partner's awareness of the woman's original intentions to use the pill. The rates of discontinuation during the first week of use were similar between women whose partner knew of their planned pill use and those whose partner did not; however, a higher proportion of women whose partner did not know discontinued use by day 29. Moreover, women whose partner was unaware of their intended pill use discontinued at a steady rate throughout the 29 days, whereas those who informed their partner mostly discontinued within the first 14 days or at day 29. Reported discontinuation tended to occur at the end of a week of use (i.e., at days seven, 14, 21, and 28 or 29), but we do not know whether women actually discontinued on these exact days or tended to report discontinuation that way.
At the univariate level, two background and fertility-related variables were associated with significantly reduced odds of discontinuation (Table 2, page 259)—having at least 12 years of schooling (odds ratio, 0.4) and having intercourse at least once a week (0.5). On the other hand, six variables were associated with significantly increased odds of discontinuation—being younger than 21 (not shown), not taking the first pill in the clinic, not having used it before, being less certain about wanting to use the pill, intending to use it for a year or less, and feeling happy about the prospect of a pregnancy within the next six months (2.0-2.6).
Only one of the variables related to the partner or relationship was significantly associated with discontinuation: Women whose partners were unaware of their plan to initiate pill use had elevated odds of discontinuing use prematurely (3.5). When we limited the analysis to women younger than 18 (not shown), there was a marginally significant association between having a partner who was at least five years older and stopping the pill in the first month of use (odds ratio, 2.1; 95% confidence interval, 0.5-9.0).
Once we controlled for all the variables that were significantly associated with pill discontinuation at the univariate level, five variables remained significant ((Table 2). In the multivariate logistic regression model, older age was significantly related to lowered odds of stopping pill use before starting the second pack (odds ratio, 0.9).* The odds of early discontinuation were significantly elevated if women did not take the first pill at the clinic (3.0), intended to use the pill for a year or less (2.3), were happy with the idea of a pregnancy (2.4) or said that their partner did not know of their planned pill use (3.4).
Many of the partner and relationship variables were interrelated (not shown). The partner's approval of pill use, in particular, was associated with his awareness that the woman planned to initiate use, his desire for a pregnancy within six months and the woman's satisfaction with him. For example, partner's awareness of planned pill use was significantly associated with discontinuation among women whose partner desired a pregnancy within six months (odds ratio, 4.4; 95% confidence interval, 1.9-10.1) but not among those whose partner did not desire a pregnancy (odds ratio, 1.9; 95% confidence interval, 0.6-6.1). A partner's awareness of planned pill use was highly correlated with the duration of the relationship, the age difference between partners and whether they were cohabiting. Finally, a woman's happiness about a potential pregnancy was associated with her partner's desire for one in the next six months, but not with her certainty about pill use. Her attitude toward a possible pregnancy was also not related to how long she intended to take the pill.
DISCUSSION
In this exploratory study of partner influence in a mostly Hispanic (Dominican) population, women whose partners were unaware of their plans to initiate pill use had elevated odds of discontinuing prematurely. A male partner's knowledge of planned pill use may reflect that the woman explicitly told him or that he had an implicit understanding of it. Both of these possibilities require a degree of communication within the relationship. Communication in a relationship affects contraceptive use in general and may be a marker for a serious and committed relationship.16 Partners who are in a committed relationship may find it easier to communicate about contraceptive use, which may lead to higher continuation rates.
Conversely, the lack of commitment in a relationship may be associated with poor contraceptive use. For example, women who are in a short-term relationship or who are at least five years younger than their partner are more likely than others to use no method.17 Although in our study the duration of the relationship, the age difference between partners and other relationship commitment variables were not independently associated with discontinuation, they were highly correlated with the partner's being aware of planned pill use. Unfortunately, this study was far too small to disentangle the separate effects of highly correlated variables.
The association between the partner's knowing about planned pill use and discontinuation may also suggest that the woman's decision to tell her partner depends on her perception of his supportiveness. (A study conducted among Indonesian women found that discontinuation of a method was less associated with partner agreement over the method than with whether the woman had received the method she wanted.18) In our study, a partner's knowing about planned pill use and his approving of it were highly correlated, and women might have been more likely to tell their partner if they thought that he would approve.
Poor motivation to practice contraception was also independently associated with pill discontinuation: Although all women in our study requested and received pills from the clinic, nearly half were ambivalent about avoiding pregnancy (i.e., they reported being happy about the prospect of a pregnancy in the next six months). Since a woman's intention to use the pill for one year or less was significantly associated with her odds of discontinuation, planned duration of use may serve as another measure of contraceptive motivation.
Several limitations of this study should be mentioned. The information on the partner's desires, attitudes and awareness of planned pill use was obtained from the woman and not from the partner himself, so some of it might have been inaccurate; our data thus report the woman's perception of her partner's knowledge about and attitude toward pill use. Further, the woman's perception of her partner's knowledge was assessed in the baseline interview only; women who reported that their partner did not know of their planned pill use may have informed him about it later. Another limitation is the sample's small size; a larger sample would allow the associations between partner influence and pill discontinuation to be more precisely characterized. Finally, grouping all Hispanic women together obscures the ethnic, racial and cultural diversity among Hispanics; our findings may not apply to all subgroups.
The dynamic between a woman and her partner that influences the decision to continue pill use is highly complex. The partner's awareness of the decision to initiate use likely reflects aspects of the relationship that are associated with discontinuation, such as levels of communication and commitment in the relationship. Communication about contraceptive use likely benefits the couple in many ways, but women who do not disclose their planned pill use may make that choice because they consider their partner unsupportive or fear adverse consequences. It is therefore important for clinicians to take into account individual women's circumstances and needs before counseling them to inform their partner of their intention to use this method. Studies that incorporate longer follow-up periods are needed among different Hispanic populations to help elucidate the interplay of partner support and communication and its effect on pill discontinuation.