There is strong evidence that racial and ethnic disparities exist across the spectrum of U.S. health services.1 Disparities in reproductive health are particularly stark. African-American women have higher rates of maternal death, infant mortality, and unintended pregnancy than white women2–4 and also have different patterns of contraceptive use.5 According to the 2002 National Survey of Family Growth, the use of any contraceptive method was less for African-American women than for white women, but their rate of contraceptive sterilization procedures was much higher.5 Rates of unintended pregnancy and sterilization for Hispanic women are between those of African-American and white women.5
Although tubal sterilization is an effective method of contraception, it permanently terminates a woman's reproductive ability. Although patient preferences may explain racial or ethnic differences, an interpretation of greater concern involves biased provider counseling or system level factors that restrict the range of contraceptive alternatives for this population of women. Because there has been a history of involuntary sterilization of poor and minority women in the United States,6–9 it is important to understand the factors that underlie persistent racial or ethnic differences in sterilization rates.
Previous studies indicate that tubal sterilization rates among African-American women have exceeded those of white women since its emergence as a popular contraceptive method in the 1970s.5,9–14 These studies, however, did not examine the effect of race or ethnicity after adjusting for many potentially important confounders. These studies also used older data sets compared with the more recent national estimates provided by the 2002 National Survey of Family Growth. To address previous limitations and to further identify key factors that influence observed racial or ethnic variations in tubal sterilization rates, we used the 2002 National Survey of Family Growth database to examine the independent effects of race or ethnicity and insurance status on tubal sterilization rates.
MATERIALS AND METHODS
This study used data collected by Cycle 6 (2002) of the National Surveys of Family Growth, a national cross-sectional survey.15 The National Survey of Family Growth is a periodic study conducted by the National Center for Health Statistics, an agency of the Department of Health and Human Services, to provide national estimates of factors affecting pregnancy and birth outcomes, including sexual activity, contraceptive use, marital status, infertility, and use of medical services for family planning. For the 2002 National Survey of Family Growth, interviews were conducted between March 2002 and March 2003.
The National Survey of Family Growth is a multistage probability sample designed to represent women and men aged 15–44 years in the household population of all 50 States and the District of Columbia. The 2002 National Survey of Family Growth sample included 7,643 women and 4,928 men. Teenaged, African-American, and Hispanic participants were oversampled. The selection of eligible persons for the 2002 National Survey of Family Growth occurred in several steps. First, large areas (cities and counties) of the country were chosen. Within each of these 121 Primary Sampling Units, groups of adjacent blocks were selected at random. Addresses within these blocks were then selected randomly and visited in person. A short “screener” interview was conducted to see if anyone aged 15–44 years lived there. If so, one person was chosen at random for the interview and offered a chance to participate. Interviews were conducted in person by a trained female interviewer in the selected person's home. The overall response rate was approximately 80%. Although the 2002 National Survey of Family Growth data included men, this analysis used only data obtained from women.
The outcome variable in this analysis was whether the respondent had undergone tubal sterilization at any time before interview. Because we were interested in utilization rates of contraceptive sterilization, our analysis did not include women who had been sterilized by hysterectomy, because this procedure is most often performed for medical rather than contraceptive reasons.10
The primary predictors of interest were self-reported race or ethnicity and medical insurance status. Although we were more interested in a woman's medical insurance status at the time of sterilization rather than at the time of interview, this information was not available to us. Instead, we had information regarding method of payment for sterilization for those women who underwent the procedure. Because Medicaid paid for the majority of tubal sterilizations for women who reported either Medicaid, public insurance, or no insurance at the time of interview, these categories were collapsed into one for all our analyses. As such, the public insurance category in this analysis included Medicaid, Medicare, Medi-Gap, Indian health service, Children's Health Insurance Program, state-sponsored, or other government program. Age, education level, income, parity, religion, and marital status have been associated with tubal sterilization in prior literature5,10–14,16 and were, therefore, examined as potential confounders in this study.
We describe variation in baseline characteristics by race or ethnicity using χ2 tests for all categorical variables. We then examined the bivariate association between history of tubal sterilization and each covariate. A multivariable logistic regression model was used to determine the adjusted odds ratios of undergoing tubal sterilization. We assessed for interactions between race or ethnicity and various levels of age, insurance status, marital status, income, education, parity, and religion.
Statistics for this project were produced using Stata 9 software (StataCorp LP, College Station, TX), using appropriate adjustment for the National Survey of Family Growth's complex sample design. All estimates were weighted to adjust for the different sampling and response rates within the survey sample. However, sample sizes are shown in all of the tables to give the reader an indication of the reliability of the estimates. This study was approved by the Institutional Review Board of the University of Pittsburgh.
Table 1 depicts the baseline characteristics of the study sample at the time of the interview. Of the 7,643 women included, 66% were non-Hispanic white, 15% were Hispanic, 14% were non-Hispanic African American, and 6% were non-Hispanic “other,” a category which included Asian, Pacific Islander, Alaskan native, and American Indian women. In the overall sample, 16% of women had undergone tubal sterilization. The majority of women (68%) had private insurance. African-American and Hispanic women were more likely to have undergone sterilization than white women: 21% of African-American women, 20%, of Hispanic women, and 15% of white women had been sterilized. African-American and Hispanic women were also more likely to have public or no insurance compared with white women.
Univariable and multivariable analyses are shown in Table 2. Because the “other” race and “other” religion categories were too heterogeneous and too limited a sample size to produce any meaningful conclusions, these categories were excluded from the multivariable analysis. Women aged 15–19 years were also excluded because they were so unlikely to have undergone sterilization (0.1%). This resulted in a final sample of 5,878 for the final multivariable model.
Increasing age and parity were the strongest predictors of tubal sterilization in univariable and multivariable analyses. After adjusting for insurance status, age, income, education, parity, marital status, and religion, we found that African-American women were significantly more likely to have undergone tubal sterilization than white women (adjusted odds ratio 1.43, 95% confidence interval 1.08–1.88). Hispanic women were more likely to have been sterilized than white women (adjusted odds ratio 1.23, 95% confidence interval 0.87–1.75); however, this association did not reach statistical significance. Public or no insurance, aged older than 30 years, and increasing parity were significantly associated with higher rates of sterilization. Receiving at least some college education, never having been married, and Catholic religion were significantly associated with lower rates of sterilization. Income was not a significant predictor either for or against tubal sterilization.
Significant interactions were found between race and insurance status, poverty level, parity, and education. Additional analyses were performed using stratified levels for each of these variables (Table 3). No significant interactions were identified between race and age, marital status, or religion. Briefly, there were no racial or ethnic variations in tubal sterilization rates among women in the more “disadvantaged” strata (public or no insurance, income below the poverty threshold, less education, and more children). Conversely, within the more “advantaged” strata (private insurance, higher income, higher education, or fewer children) African-American women were significantly more likely to have undergone sterilization compared with white women.
In a sample of 5,878 women, we found that African-American and Hispanic women were more likely to have undergone tubal sterilization compared with white women after adjusting for insurance status, age, poverty level, education, parity, marital status, and religion. We also found that women with public or no insurance were more likely to have undergone tubal sterilization than women with private insurance.
Our stratified analyses based on significant interactions provided insight as to the effect of race or ethnicity across the various levels of education, income, parity, and insurance status. Although African-American and Hispanic women overall were more likely to have undergone sterilization compared with white women, interaction effects indicate that the higher observed rates of sterilization among minority women seems to be largely driven by women in more “advantaged” situations: those with private insurance, higher income, higher education, and fewer children. There were no racial or ethnic differences in the rate of sterilization among the most disadvantaged women: those with public or no insurance, below the poverty line, lower education, and three or more children. The finding that women who presumably have more contraceptive options available to them are more frequently sterilized suggests that the higher rate of sterilization seen among minority women may be due to preference. It remains unclear, however, what factors lead Hispanic and African-American women to choose different contraceptive methods from white women.
Although cultural factors may certainly explain different patterns of contraceptive use, the role of the provider cannot be discounted, especially because provider bias in contraceptive counseling17 and provider influence on contraceptive choices have been observed.18 Furthermore, there is evidence that this bias may be race or class based. Harrison and Cooke19 showed that older age, multiparity, and African-American race were the most significant factors in a gynecologist's willingness to perform sterilization. In our study, women with public or no insurance were more likely to have undergone sterilization compared with women with private insurance, even after controlling for other markers of socioeconomic status (race or ethnicity, income, education, and number of children). Although this study cannot determine the root causes of this finding, it is certainly plausible that, in addition to patient preference, provider-level factors play a role. Prior studies have demonstrated a trend toward permanent or long-acting methods of contraception (sterilization, intrauterine device, levonorgestrel [Norplant, Wyeth Pharmaceuticals, Madison, NJ]) in self-pay or publicly insured women compared with privately insured women.16 The authors of that study proposed that, in addition to discrepancies in contraceptive coverage, health care providers' use of insurance information as a proxy of the patient's ability to support a family might explain their findings.16
An alternative explanation for the observed differences in sterilization rates between privately and publicly or noninsured women is that system-level factors restrict the range of contraceptive alternatives for women with public or no insurance. Although public funds through Medicaid and Title X provide a wide range of contraceptive choices, women must meet strict state eligibility criteria for the former and demonstrate ongoing proof of income eligibility to receive subsidized services for the latter.20 However, many women become eligible for Medicaid coverage at the time of pregnancy because of pregnancy-related Medicaid expansions. This pregnancy-related Medicaid coverage is extended to 60 days postpartum in most states to cover family planning services. This transient insurance coverage may be an incentive for women without private insurance to obtain a permanent method of postpartum contraception (sterilization) rather than rely on a contraceptive method that requires ongoing services or supplies. In fact, MacKay et al11 found that Medicaid was the source of payment for a much higher proportion of postpartum tubal sterilizations than for interval (nonpostpartum) sterilizations (41% compared with 24%).
Our study has several important limitations. First, although the National Survey of Family Growth provides a wealth of information about reproductive health in the United States, it is a cross-sectional survey that obtains information at only one point in time. Information on factors which may be important in the decision to undergo tubal sterilization are obtained at the time of interview rather than at the time the decision was made. However, it is unlikely that this limitation had an effect on the two main primary predictor variables of this study (race or ethnicity and insurance status). Race or ethnicity is a stable variable, whereas insurance status was strongly correlated with payment method at the time of sterilization. Among those women with private insurance at the time of the interview, 86% paid for their sterilization with private insurance. Among those women with Medicaid at the time of interview, 81% paid for their sterilization with Medicaid. Among women with no insurance or public insurance (other than Medicaid), the majority (57% and 62%, respectively), paid for their sterilization with Medicaid. Second, we did not have information regarding geographic data. Regional variation in tubal sterilization rates have been documented, with highest rates in the South and lowest rates in the West10–12,14 Because more African-American people live in the South,21 it is possible that regional differences based on variations of provider attitudes or healthcare delivery account for the higher sterilization rates seen in African-American women. Alternatively, it is also possible that the observed racial variations account for the regional differences. Neither argument, however, explains the higher, although nonsignificant, rate of sterilization observed in Hispanic women who are more concentrated in the West21 (which has the lowest rate of sterilizations). Third, some tubal sterilizations may have been performed for noncontraceptive reasons. However, the vast majority of women (more than 80%) who reported undergoing the procedure in the 2002 National Survey of Family Growth cited nonmedical reasons. Last, because there are significant racial or ethnic variations among vasectomy rates,14 vasectomy may be an important confounder of racial or ethnic differences in tubal sterilization rates. This variable, however, was not included in our final regression model because of a large amount of missing data.
In summary, we found that after controlling for important socioeconomic factors, African-American women and women with no or public insurance were more likely to have undergone tubal sterilization compared with white women and women with private insurance, respectively. Additional research is needed to assess whether these increased rates for African-American women and women with no or public insurance are associated with increased rates for poststerilization regret and to ascertain which external, if any, factors (ie, provider counseling, access, insurance coverage) contributed to the decision-making process. Investigating these observed disparities will help to ensure that women receive high-quality family planning services that allow them to make truly informed choices.
1. Smedley BD, Stith AY, Nelson AR. editors. Unequal treatment: confronting racial and ethnic disparities in health care. Washington (DC): National Academies Press; 2003.
2. National Center for Health Statistics. Health, United States, 2004. Hyattsville (MD): Government Printing Office; 2004.
3. Anachebe NF, Sutton MY. Racial disparities in reproductive health outcomes. Am J Obstet Gynecol 2003;188:S37–42.
4. Henshaw SK. Unintended pregnancy in the United States. Fam Plann Perspect 1998;30:24–9, 46.
5. Mosher WD, Martinez GM, Chandra A, Abma JC, Wilson SJ. Use of contraception and use of family planning services in the Unites States: 1982-2002. Advance data from vital and health statistics; No. 350. Hyattsville (MD): National Center for Health Statistics; 2004.
6. Eliot JW. Fertility control and coercion. Fam Plann Perspect 1973;5:132, 187.
7. Stern AM. Sterilized in the name of public health: race, immigration, and reproductive control in modern California. Am J Public Health 2005;95:1128–38.
8. Roberts D. The pill at 40—a new look at a familiar method. Black women and the pill. Fam Plann Perspect 2000;32:92–3.
9. Shapiro TM, Fisher W, Diana A. Family planning and female sterilization in the United States. Soc Sci Med 1983;17:1847–55.
10. Chandra A. Surgical sterilization in the United States: prevalence and characteristics, 1965-95. Advance data from vital and health statistics; series 23: No. 20. Hyattsville (MD): National Center for Health Statistics; 1998.
11. MacKay AP, Kieke BA Jr, Koonin LM, Beattie K. Tubal sterilization in the United States, 1994–1996. Fam Plann Perspect 2001;33:161–5.
12. Godecker AL, Thomson E, Bumpass LL. Union status, marital history, and female contraceptive sterilization in the United States. Fam Plann Perspect 2001;33:35–41.
13. Piccinino LJ, Mosher WD. Trends in contraceptive use in the United States: 1982-1995. Fam Plann Perspect 1998;30:4–10, 46.
14. Bumpass LL, Thomson E, Godecker AL. Women, men, and contraceptive sterilization. Fertil Steril 2000;73:937–46.
15. National Center for Health Statistics. National Survey of Family Growth, Cycle 6. Available at: http://www.cdc.gov/nchs/nsfg.htm
. Retrieved September 30, 2006.
16. Boardman LA, Weitzen S, Lapane KL. Context of care and contraceptive method use. Womens Health Issues 2004;14:51–9.
17. Baveja R, Buckshee K, Das K, Das SK, Hazra MN, Gopalan S, et al. Evaluating contraceptive choice through the method-mix approach: an Indian Council of Medical Research (ICMR) task force study. Contraception 2000;61:113–9.
18. do Lago, TD, Barbosa RM, Kalckman S, Villela WV, Goihman S. Acceptability of the diaphragm among low income women in Sao Paulo, Brazil. Int Fam Plann Perspect 1995;21:114–8.
19. Harrison DD, Cooke CW. An elucidation of factors influencing physicians' willingness to perform elective female sterilization. Obstet Gynecol 1988;72:565–70.
20. Kaeser L. Title X and the U.S. family planning effort. New York (NY): Alan Guttmacher Institute; 1997.