Materials and Methods

CHOICE is a prospective cohort study of 9526 St. Louis–area women. Through convenience sampling, we recruited women 14-45 years old in university and community clinics that provided contraceptive care and 2 facilities that provided abortion services. CHOICE promoted the use of LARC methods by removing the cost, education, and access barriers that often are experienced in contraceptive services. A detailed explanation of the methods of CHOICE has been published previously. All participants provided written informed consent, and the study was approved by the Washington University School of Medicine Human Subjects Protection Office before participant recruitment.

To be eligible to participate in CHOICE, women had to meet the following inclusion criteria: 14-45 years old, sexually active with a male partner (or intent to be within the next 6 months), willing to try a new contraceptive method or not currently using contraception, not desiring pregnancy in the next 12 months, residing in or seeking clinical services at recruitment sites in the St. Louis area, and consenting in English or Spanish.

To recruit a diverse participant sample that included Hispanic women, we created recruitment and study materials in English and Spanish. Research assistants who were fluent in Spanish recruited Spanish-speaking study participants and conducted follow-up surveys. At enrollment, study participants underwent standardized contraceptive counseling that presented evidence-based and accurate information about each contraceptive method in the order of method effectiveness. Participants chose and were provided their method at no cost for 2-3 years. Staff members administered a baseline survey to collect demographic information and a reproductive history. Participants were followed with telephone surveys at 3 months, 6 months, and every 6 months thereafter for the duration of follow-up period. Participants were paid $10 for each completed survey.

Information regarding method use and satisfaction was collected at each follow-up survey. To be included in this analysis, a participant had to receive and start her baseline chosen method of contraception within 3 months of enrollment. We defined method continuation as reporting method use at 3, 6, and 12 months without a period of nonuse of >1 month. Women were coded as having discontinued their method if they reported discontinuation or removal of their contraceptive method at <12 months or reported a gap in contraceptive coverage longer than 1 month at any follow-up survey. Women who experienced IUD expulsion during the first 12 months were considered continuers if the IUD was replaced within a month; women who did not have it replaced or chose a new method were coded as discontinuers. We measured 12-month method satisfaction as “very satisfied,” “somewhat satisfied,” or “not satisfied” with their current contraceptive method when asked during the 12-month survey. Women who discontinued their method were coded as “not satisfied.” Women who discontinued because of a pregnancy or a desire to conceive were excluded from the satisfaction analysis.

At baseline, we recorded Hispanic ethnicity and race as 2 separate categories. In this analysis, we defined 3 groups. A woman was coded Hispanic when she reported her ethnicity as Hispanic, regardless of race. The other 2 groups represented women who reported themselves as non-Hispanic white and non-Hispanic black. Women who self-identified as other race were not included in this analysis.

To describe the demographic characteristics of the participants who were included in this analysis, we used frequencies, percentages, means, and standard deviation. To compare these variables, we used χ ² tests for categoric variables and Student t tests for continuous normally distributed variables. A histogram was used to assess normality. We estimated continuation rates by constructing Kaplan-Meier survival functions and compared continuation between ethnicities and among different methods using the log-rank test. We used Cox proportional hazard models to estimate the risk of method discontinuation. The interaction between ethnicities and covariates was evaluated, and cofounders were defined as those covariates that altered the hazard ratio of ethnicity and continuation by >10%. We described satisfaction rates using frequencies and percentages. To account for participants who were lost to follow up, we censored participants at their last completed survey date and used contribution time and method use data up to the last contact. Participants who discontinued because of a pregnancy or a desire to conceive were censored at the point of discontinuation. All analyses were completed using STATA software (version 11; StataCorp, College Station, TX). The α-level for all analyses was set at .05.

Results

Among the 9256 women enrolled in CHOICE, 7913 women (85%) met the inclusion criteria for this analysis ( Figure 1 ). Hispanic women represented 5% of the total cohort, which is similar to the percentage of Hispanic persons in the St. Louis area. Forty-five percent of Hispanic participants were born in another country, of which 73% were born in Mexico. The baseline characteristics of the analysis sample are presented by race and ethnicity in Table 1 . Hispanic women in CHOICE were more likely to be receiving public assistance, to have a high school education or less, and to desire ≥3 children. They were less likely to have insurance and to report a history of abortion. Hispanic women were more likely to choose a LARC method, compared with non-Hispanic black and non-Hispanic white women (80%, 73%, and 75% respectively; P < .05).

Exclusion criteria for final analysis sample

Flowchart depicts the exclusion criteria that were used to determine which participants would be included in the final analysis.

CHOICE , Contraceptive CHOICE Project.

Werth. Contraceptive continuation in Hispanic women. Am J Obstet Gynecol 2015 .

Because we detected a significant interaction between contraceptive method and race/ethnicity, Table 2 presents continuation rates that were stratified by contraceptive method. We found no confounders in this analysis. The 12-month continuation rates were higher for LARC methods than combined hormonal methods (OCPs, patch, ring) for all race/ethnicities (Hispanic women, 87% vs 40%; non-Hispanic black women, 85% vs 46%; non-Hispanic white women, 87% vs 56%, respectively). We found no statistically significant difference in LARC discontinuation at 12 months by race/ethnicity ( Figure 2 ; P = .27). Hispanic and non-Hispanic white women were equally likely to continue LARC methods (hazard ratio, 0.96; 95% confidence interval, 0.69–1.34). Hispanic women who used the OCP, patch, or ring were more likely to discontinue combined hormonal methods by 12 months, compared with non-Hispanic white and non-Hispanic black women ( Figure 3 ; P < .01). Hispanic women were twice as likely to discontinue the ring (hazard ratio, 2.2; 95% confidence interval, 1.35–3.54) compared with non-Hispanic white women.

Table 2

Twelve-month continuation rates stratified by race/ethnicity and hazard ratios for risk of discontinuation

Contraceptive method	Race/ethnicity	n	Continuation, %	Hazard ratio (95% confidence interval)
Long-acting reversible contraception	Hispanic	333	87	0.96 (0.69–1.34)
	Non-Hispanic black	3008	85	1.11 (0.96–1.29)
	Non-Hispanic white	2519	87	Reference
Combined hormonal methods (pill, patch, ring)	Hispanic	68	40	1.58 (1.13–2.21)
	Non-Hispanic black	673	46	1.35 (1.15–1.57)
	Non-Hispanic white	727	56	Reference
Levonorgestrel intrauterine system	Hispanic	108	87	1.03 (0.65–1.64)
	Non-Hispanic black	1781	86	1.17 (0.97–1.42)
	Non-Hispanic white	1600	87	Reference
Copper intrauterine device	Hispanic	63	83	1.16 (0.6–2.25)
	Non-Hispanic black	352	86	0.96 (0.66–1.38)
	Non-Hispanic white	506	85	Reference
Implant	Hispanic	102	89	0.67 (0.34–1.31)
	Non-Hispanic black	875	84	1.04 (0.77–1.4)
	Non-Hispanic white	413	84	Reference
Depot medroxyprogesterone acetate	Hispanic	18	69	0.48 (0.19–1.19)
	Non-Hispanic black	448	54	0.83 (0.62–1.12)
	Non-Hispanic white	119	50	Reference
Oral contraceptive pill	Hispanic	38	52	1.25 (0.75–2.07)
	Non-Hispanic black	354	47	1.45 (1.17–1.81)
	Non-Hispanic white	363	58	Reference
Patch	Hispanic	4	25	1.25 (0.38–4.13)
	Non-Hispanic black	92	37	1.17 (0.73–1.88)
	Non-Hispanic white	53	46	Reference
Vaginal ring	Hispanic	26	27	2.19 (1.35–3.54)
	Non-Hispanic black	227	49	1.18 (0.91–1.52)
	Non-Hispanic white	311	56	Reference

Only gold members can continue reading. Log In or Register to continue

Share this:

• Click to share on Twitter (Opens in new window)
• Click to share on Facebook (Opens in new window)

Related

Related posts:

1. Comprehensive maternal hemorrhage protocols reduce the use of blood products and improve patient safety
2. Effects of exogenous progesterone on fetal nuchal translucency: an observational prospective study
3. Influence of interpregnancy interval on neonatal morbidity
4. An incidental finding of endometrial osseous metaplasia during office hysteroscopy

Stay updated, free articles. Join our Telegram channel

Join