Unintended pregnancies have significant consequences and occur most frequently in adolescent, low income, and minority women.1 Approximately 78% of pregnancies in adolescents and 72% of pregnancies in black women are unintended.1 Whether these pregnancies end in abortion or unintended births, they pose significant personal, social, and economic burden.2 The prevalence of unintended pregnancy is determined largely by contraceptive practice. Contraceptive practice has two components: method choice and pattern of use. Use of less effective methods, inconsistent or infrequent use, and method discontinuation, have a significant impact on the rate of unintended pregnancies.3
Contraceptive method choice differs by race and age, and both of these demographic factors have been disproportionately associated with contraceptive failure and unintended pregnancy.4–6 Data from the 1995 National Survey of Family Growth (NSFG) revealed that among other demographic factors, age, race, and ethnicity are strong correlates of contraceptive choice.7 The proportion of racial/ethnic minorities using hormonal contraception varied, with blacks and Hispanics choosing the long‐term hormonal methods, such as injections or implants, more frequently over oral contraception.8 Other studies have documented similar differences with blacks and Hispanics being significantly less likely than white women to use oral contraceptives and more likely to use condoms or long‐term hormonal contraception.9
Although there are several epidemiologic studies documenting subgroup differences in contraceptive use, contraceptive failure rates, and unintended pregnancies, our understanding of factors associated with variations in contraceptive choice among adolescents and racial minorities is limited. Race and ethnicity are proximal measures for socially proscribed divisions within populations. The extent to which individuals who belong to a particular race, ethnic group, or culture share certain beliefs, attitudes, and values may determine whether they engage in similar risk behaviors or health practices and experience similar health outcomes. Adolescents from minority ethnic or racial groups may have beliefs or practices that are culture specific and influence contraceptive choice. An alternative explanation is that black and Hispanic women are also more likely to be of lower socioeconomic status, have limited vocational and professional opportunities, and experience situational factors, all of which can contribute to less effective contraceptive use and pregnancy.10
Our hypothesis was that contraceptive method choice and thus unintended pregnancy rates among adolescents and racial minorities are influenced by a combination of demographic, motivational, and cultural factors, and that race/ethnicity may serve as a proxy measure of cultural factors. The purpose of this study was to estimate the association of race/ethnicity with adolescent contraceptive choice and pregnancy after controlling for demographic and motivational factors that are known to influence contraceptive choice. We used data obtained from the retrospective chart review of a culturally diverse population of adolescents making an initial visit to a teen family planning clinic. With increasing condom use due to successful sexually transmitted disease (STD)‐prevention efforts, family planning visits during this time are increasingly seen as an opportunity to initiate more effective hormonal methods.11 Family planning providers need to understand patterns of contraceptive choice in this particularly vulnerable group to be able to help them make contraceptive choices to protect themselves from STDs and unintended pregnancy.
We screened the charts of all clients presenting for an initial visit to a publicly‐funded teen family planning clinic in the Mission District of San Francisco, California, between March 1997 and July 1998 for retrospective review. Female adolescents aged 12–19 were considered eligible for review. There were 1015 clients with initial visits; 56 (5%) clients were male, and 354 (35%) were ages 20–24. The charts of the remaining 605 (60%) female adolescents were reviewed for demographic information including age, race/ethnicity, median household income (determined from zip code using 1990 census block‐group data), pregnancy and reproductive health history, sexual risk behaviors including history of new sexual partners and number of sexual partners in the last 2 months, pregnancy desire, and contraceptive method at presentation (current method) and at the end of the clinic visit (dispensed method). Race/ethnicity was determined from self‐report on two separate forms—funding eligibility and past medical history. Pregnancy desire or intention was assessed by response to the question: “Do you want to be pregnant now?” on a risk‐factor screening form. Pregnancy status was determined using pregnancy test results or last menstrual period charted. Contraceptive method used was based on self‐report on the medical history form or from provider notes. Contraceptive method was categorized based on the most effective method used when more than one method was used. Contraceptive method was collapsed into three categories: none (includes withdrawal and periodic abstinence), barrier (condoms and spermicides), and hormonal methods (oral contraceptive pills, medroxyprogesterone acetate injection, and levonorgestrel implants).
We used χ2 statistics to examine the relationship between pairs of categoric variables and analysis of variance for continuous variables. Multiple logistic regression modeling was used to estimate the association of race/ethnicity with pregnancy status at the initial visit, after controlling for age, median household income, previous pregnancy, new sexual partner, STDs, and current contraceptive method. We used multinomial logistic regression analysis to estimate the association of race/ethnicity with current contraceptive method controlling for confounding variables.12 In this model, the dependent variable—current contraceptive method—consisted of three possible outcomes: use of no method, barrier methods, and hormonal methods. The results are presented as coefficients that represent the log odds of the probability of choosing one method relative to another; the sign of the coefficient indicates the direction of the association. The choice of barrier methods was used as the reference category to which both others, choice of hormonal methods and choice of no methods, were compared. Finally, multiple logistic regression analysis was used to estimate the association of race/ethnicity and other factors with dispensed method. Statistical significance is reported at the .05 level. Data were analyzed using Stata 6.0 (Stata Corporation, College Station, TX).
The study population was young and culturally diverse, ranging from 12 to 19 years, with a mean age of 16.6 (Table 1). Thirty‐three percent of the adolescents were Hispanic, 30% black, 15% Asian/Pacific Islander, 8% white, and 14% multiracial or other. Black adolescents were significantly more likely to be low income compared with adolescents from all other races (P < .01).
The sample was relatively high‐risk for pregnancy and sexually transmitted diseases, with 35% reporting previous pregnancies, 16% reporting teen births, and 16% having a history of STDs (Table 1). White adolescents were significantly less likely to have been previously pregnant compared with blacks, Hispanics, and Asian/ Pacific Islanders (P = .03). Asian/Pacific Islanders and other (multiracial) adolescents were far less likely to have given birth (P = .04). Black adolescents were almost twice as likely as all other adolescents to report a history of STDs (P < .01).
The most common reason for the visit was pregnancy testing, with over half of adolescents (55%) having pregnancy testing and counseling as their main reason for coming to the clinic. There were no significant differences in pregnancy desire by race, with the majority (91%) of clients reporting that they did not want to be pregnant. However, 117 (19%) adolescents were pregnant at presentation and there were significant racial/ ethnic differences in pregnancy status at presentation (P < .01) (Table 1).
At presentation, the large majority (73%) of adolescents reported using barrier methods (94% condoms) as their contraceptive method. Although not significantly different, black adolescents were slightly more likely than adolescents from all other racial/ethnic groups to present relying on barrier methods (81% versus 70%, P = .10). At the end of the visit, there were no significant differences in contraceptive method dispensed by race (P = .63).
In a multiple logistic regression model controlling for the independent variables listed in Table 2, black adolescents were three times as likely as other groups to be pregnant at the initial visit (P = .05). Current contraceptive method and previous pregnancy were the only other significant risk factors for pregnancy. Adolescents using hormonal methods were five times less likely to be pregnant at the first visit than users of barrier methods (P = .01). Adolescents who used no contraceptive method, on the other hand, were almost twice as likely to be pregnant than users of barrier methods (P = .02). Adolescents with a previous pregnancy were almost three times as likely as those without a pregnancy to present to the clinic pregnant (P < .01).
Pregnancy desire was strongly associated with current contraceptive method; adolescents who reported that they did not want to be pregnant were far more likely to be using a contraceptive method than adolescents who reported wanting to be pregnant or were unsure about pregnancy desire (86% versus 50%, P < .01) (Table 3). Ten percent of adolescents who reported that they did not want to be pregnant used hormonal methods of contraception, whereas none of the adolescents who reported being ambivalent or having a desire to be pregnant used hormonal methods.
The results of multinomial logistic regression analysis with current method of contraception as the dependent variable are presented in Table 4. Race/ethnicity was associated with current method, with black adolescents being significantly more likely to present using barrier methods than hormonal methods (P = .04). Adolescents with a previous pregnancy were also significantly more likely to rely on barrier methods than hormonal methods (P < .01). Older adolescents were more likely to rely on no method than barrier methods (P = .02) as well as slightly more likely to rely on hormonal methods than barrier methods (P = .06). Adolescents living in neighborhoods with lower median household incomes were more likely to be using no method than barrier methods (P = .02).
In the multiple logistic regression analysis of factors associated with method dispensed (Table 5), current contraceptive method was a significant predictor, with adolescents using hormonal methods being far more likely to continue with hormonal methods (P < .01). Age was a also a significant predictor of dispensed method, with older adolescents being more likely to receive hormonal methods compared with barrier methods at the end of the visit (P = .04). Neither race/ethnicity nor pregnancy desires were significant independent predictors of method dispensed.
Although the mission of most public family planning clinics is to help young women avoid unintended pregnancy, the reality is that many women are already pregnant and are coming to the clinic for pregnancy testing. Adolescents, low‐income women, and minorities are more likely to use a clinic than a private doctor or health maintenance organization for family planning services.7 The high rate of visits for pregnancy testing in our sample is consistent with data by Zabin and Clark demonstrating that 37% of teenage clinic patients reported the need for a pregnancy test as the precipitating reason for a first clinic visit.13 Adolescents may be more likely to use family planning clinics for pregnancy tests because of the cost of over‐the‐counter tests and for the purpose of avoiding care with a private provider through their parents' insurance. In a convenience sample of about 3000 adolescents from 52 clinics presenting for a pregnancy test, 36% of tests were positive, which was similar to our rate for women presenting for pregnancy testing.14 Providers should be aware that that the majority of clients will have negative pregnancy test results, and this serves as an opportunity to educate and initiate more effective hormonal contraceptive methods.
Black adolescents were more likely than white adolescents to present to the clinic pregnant, which is consistent with data demonstrating disproportionately higher pregnancy rates in black adolescents.15 In the multivariate analysis, race, history of previous pregnancy, and use of less effective contraceptive methods were independent risk factors for pregnancy at presentation. In line with a previous finding that half of teen mothers conceive again within 2 years, the use of less effective contraceptive methods as well as a history of previous pregnancy puts adolescents at even higher risk for a subsequent pregnancy.16 However, the explanation for the independent association between race and pregnancy is less clear.
Black and younger adolescents were significantly less likely to use hormonal methods over barrier methods (Table 4). These findings are consistent with other national data that show that condom and oral contraceptive pill use among adolescents varies by race and age. In the 1997 Youth Risk Behavioral Survey, black and Hispanic female high school students (11.9% and 9.5%, respectively) were significantly less likely to report birth control pill use than whites (20.6%).17 In the same study, black female students (58.9%) were significantly more likely than white and Hispanic female students (49.2% and 40.0%, respectively) to report condom use.
In this study, blacks were more likely to use less effective barrier methods at presentation, but were not more likely to use no method. An important question is whether or not blacks are preferentially using condoms over hormonal methods because of increased risk for STDs and prevalent HIV prevention messages within the black community.18 Blacks were significantly more likely to have had an STD in the past; however, after controlling for previous history of STD and a new sexual partner in the last 2 months, black adolescents were still more likely to choose barrier methods over hormonal methods, indicating that other factors may be more important in the decision to use barrier methods instead of hormones. Research by Overby et al demonstrated that condom use among black adolescents correlated more with current monogamy, lack of intravenous drug use, and implicit trust in their partner's safety rather than history of STDs and number of sexual partners.19 One explanation for the increased preference for condoms is that black adolescents may have a higher “perceived” risk of STDs. Family planning providers should consider real and perceived STD risks and stress dual‐method use as a possible option. An alternate explanation for the increased preference for condoms is that black adolescents may have culture‐specific concerns about the safety and/or adverse effects of hormonal methods. Providers may need to explore what clients consider to be culturally acceptable contraceptive practices and address specific concerns.
Lack of desire to be pregnant should translate into motivation to prevent pregnancy and thus pregnancy desire must also be considered as a factor influencing contraceptive choice and risk of pregnancy.20,21 Pregnancy desire or ambivalence was strongly associated with the use of less effective methods (Table 2), such that no individuals who reported wanting to be pregnant or even being ambivalent about pregnancy used hormonal methods. In a separate multinomial logistic regression analysis of current contraceptive method (not presented), which examined only adolescents who did not want to be pregnant (n = 504), race remained a significant independent risk factor for use of less effective barrier methods.
Age was a significant predictor of method both at presentation to the clinic and at the end of the visit. Older adolescents are more likely to be in more stable partnerships that are conducive to use of more reliable, long‐term hormonal methods.22,23 Neither race nor desire to be pregnant was an independent predictor of dispensed method. However, women who used hormonal methods at presentation to the clinic—who were more likely to be white—were more likely to leave the clinic with hormonal methods.
Previous research in adult women has failed to generate consensus about differences in contraceptive choice by race. A study by Stephen et al using data from the 1976 and 1982 National Surveys of Family Growth, revealed that there were complex racial differences in contraceptive choice that changed over time and with variations in marital status.4 They suggested that socio‐cultural forces underlying differences in family formation between blacks and whites might also affect contraceptive choices. In a similar study by Tanfer et al using data from the 1988 National Survey of Family Growth, after controlling for the effects of various social and demographic factors, white women were found to be more likely to use a method than black women; however, there were no racial differences in methods chosen.5 Although these studies included older women and were done when fewer contraceptive methods were available, they concluded, as we do, that contraceptive choice is complex, and racial differences are not fully explained by known socioeconomic and reproductive health risk factors.
Extensive chart review allowed us to examine factors influencing contraceptive choice in this sample population, but our retrospective approach may have influenced our results. Some of our measures may have been too crude to measure desired risk factors. For example, we used census tract data to measure socioeconomic status and our assessment of pregnancy desire or intention was based on a single question. It is possible that these measures were not sensitive enough, and race is still a residual confounding variable. We only measured birth control method reported or recorded; we did not have information on consistency or actual use. Although associations between contraceptive method choice and race have also been demonstrated in national high school samples, these results cannot be generalized to non‐clinic or low‐risk populations.
1. Henshaw S. Unintended pregnancy in the United States. Fam Plann Perspect 1998:30:24–9.
2. Trussell J, Leveque JA, Koenig JD, London R, Borden S, Henneberry J, et al. The economic value of contraception: A comparison of 15 methods. Am J Public Health 1995; 85:494–503.
3. Ranjit N, Bankole A, Darroch JE, Singh S. Contraceptive failure in the first two years of use: Differences across socioeconomic subgroups. Fam Plann Perspect 2001;33:19–27.
4. Stephen EH, Rindfuss RR, Bean FD. Racial differences in contraceptive choice: complexity and implications. Demography 1988;25;53–70.
5. Tanfer K, Cubbins LA, Brewster KL. Determinants of contraceptive choice among single women in the United States. Fam Plann Perspect 1992;24:155–61.
6. Jones EF, Forrest JD. Contraceptive failure rates based on the 1988 National survey of Family Growth. Fam Plann Perspect 1992;24:12–9.
7. Abma J, Chandra A, Mosher W, Peterson L, Piccinino L. Fertility, family planning, and women's health: New data from the 1995 national survey of Family Growth. National Center for Health Statistics. Vital Health Stat 23 1997;19:49–58.
8. Piccinino L, Mosher W. Trends in contraceptive use in the United States: 1982–1995. Fam Plann Perspect 1998;30:4–10&46.
9. Forrest JD, Frost JF. The family planning attitudes and experiences of low-income women. Fam Plann Perspect 1996;28:246–55.
10. Musick JS. Young, poor, and pregnant. New Haven: Yale University Press, 1993.
11. Finer LB, Zabin LS. Does the timing of the first family planning visit still matter? Fam Plann Perspect 1998;30:30–3.
12. Agresti A. Categorical data analysis. New York: John Wiley & Sons, 1990:315–7.
13. Zabin LS, Clark SD. Why they delay: A study of teenage family planning clinic clients. Fam Plann Perspect 1981;13:205–17.
14. Zabin LS, Emerson MR, Ringers PA, Sedivy V. Adolescents with negative pregnancy test results, an accessible at-risk group. JAMA 1996;275:113–7.
15. Centers for Disease Control and Prevention. National and state-specific pregnancy rates among adolescents United States, 1995–1997. MMWR Morb Mortal Wkly Rep 2000;49:605–11.
16. Rigsby DC, Macones GA, Driscoll DA. Risk factors for rapid repeat pregnancy among adolescent mothers: A review of the literature. J Pediatr Adolesc Gynecol 1998; 11:115–26.
17. Centers for Disease Control and Prevention. CDC Surveillance Summaries, August 14, 1998. MMWR Morb Mortal Wkly Rep 1998;47(No. SS-3).
18. Langer LM, Zimmerman RS, Katz JA. Which is more important to high school students: Preventing pregnancy or preventing AIDS? Fam Plann Perspect 1994;26:154–9.
19. Overby KJ, Kegeles SM. The impact of AIDS on an urban population of high-risk female minority adolescents: Implications for intervention. J Adolesc Health 1994;15:216–7.
20. Dash L. When children want children. New York: Penguin books, 1989.
21. Brown SS, Eisenburg L, eds. The best intentions: Unintended pregnancy and the well-being of children and families. Division of Health Promotion and Disease Prevention, Institute of Medicine. Washington, D.C.: National Academy Press, 1995.
22. Santelli JS, Kouzis AC, Hoover DR, Polacsek M, Burwell LG, Celentano DD. Stage of behavior change for condom use: The influence of partner type, relationship and pregnancy factors. Fam Plann Perspect 1996;28:101–7.
23. Paul EL, White KM. The development of intimate relationships in late adolescence. Adolescence 1990;98:375–400.