Objective
The objective of the study was to evaluate the associations between postpartum contraception and having a recent preterm birth.
Study Design
Population-based data from the Pregnancy Risk Assessment Monitoring System in 9 states were used to estimate the postpartum use of highly or moderately effective contraception (sterilization, intrauterine device, implants, shots, pills, patch, and ring) and user-independent contraception (sterilization, implants, and intrauterine device) among women with recent live births (2009–2011). We assessed the differences in contraception by gestational age (≤27, 28–33, or 34–36 weeks vs term [≥37 weeks]) and modeled the associations using multivariable logistic regression with weighted data.
Results
A higher percentage of women with recent extreme preterm birth (≤27 weeks) reported using no postpartum method (31%) compared with all other women (15–16%). Women delivering extreme preterm infants had a decreased odds of using highly or moderately effective methods (adjusted odds ratio, 0.5; 95% confidence interval, 0.4–0.6) and user-independent methods (adjusted odds ratio, 0.5; 95% confidence interval, 0.4–0.7) compared with women having term births. Wanting to get pregnant was more frequently reported as a reason for contraceptive nonuse by women with an extreme preterm birth overall (45%) compared with all other women (15–18%, P < .0001). Infant death occurred in 41% of extreme preterm births and more than half of these mothers (54%) reported wanting to become pregnant as the reason for contraceptive nonuse.
Conclusion
During contraceptive counseling with women who had recent preterm births, providers should address an optimal pregnancy interval and consider that women with recent extreme preterm birth, particularly those whose infants died, may not use contraception because they want to get pregnant.
In 2012, 12% of all US births were preterm (PTB; <37 weeks’ gestation), and preterm-related deaths are the leading cause of infant mortality. Short interpregnancy intervals (IPI) (ie, conception within 18 months of a previous birth) are associated with approximately 40% increased risk of PTB (<37 weeks’ gestation), low birthweight, and small for gestational age and an increased risk of recurrent PTB.
Short IPI has also been linked to severe maternal complications such as premature membrane rupture, abruption placentae, and placenta previa. Consequently, a Healthy People objective aims to reduce the proportion of pregnancies with short IPI by 10% by 2020 (baseline, 33.1%, 2006–2010).
Use of highly effective contraception postpartum, particularly user-independent methods, is an important strategy for reducing PTB, short IPI, and recurrent PTB. User-independent methods include male and female sterilization for those not desiring another pregnancy and long-acting reversible contraceptives (LARCs) for women who are not ready for child-bearing but want to preserve their fertility.
The American College of Obstetricians and Gynecologists encourages clinicians to offer LARCs as first-line contraception because they are reversible, have very high effectiveness and continuation rates (>99% of women avoid an unintended pregnancy within the first year of use), and are cost effective, even when used short term (12–24 months). Permanent contraceptive methods (sterilization) are also highly effective (>99%), whereas effectiveness rates of other moderately effective, user-dependent methods (ie, pills, patch, ring, and shots) range from 91% to 94% with typical use.
Although several studies have examined contraceptive methods used postpartum, none have focused on women with a recent PTB, a group that is at risk of future PTB and in need of highly effective contraception to prevent short IPI and reduce recurrent PTB. We examined the prevalence of postpartum contraceptive use among women with recent live births and explored the associations between recent PTB and the consequent use of highly and moderately effective methods. We also investigated whether associations vary by insurance type and examined the reasons for contraceptive nonuse.
Materials and Methods
This analysis is based on data from the Pregnancy Risk Assessment Monitoring System (PRAMS), an ongoing population-based survey of women with live births in the past 2–9 months. The PRAMS research design and survey methods have been described elsewhere, and additional details are available from the PRAMS web site ( http://www.cdc.gov/prams ).
Briefly, each participating state draws a stratified random sample from birth certificates and mails up to 3 surveys to each selected participant. Women who do not respond to the mailings are followed up by telephone. The data are weighted to account for sampling frame, noncoverage, and participant nonresponse, thus allowing for population-based inferences. The PRAMS protocol was approved by the Centers for Disease Control and Prevention’s Institutional Review Board, and participating states approved the study analysis plan.
Data
PRAMS surveys comprise core questions that are asked by all participating sites and standard optional questions that sites may choose to add. For this analysis, we analyzed 2009–2011 data from 9 states (Arkansas, Colorado, Michigan, Nebraska, Ohio, Oregon, Rhode Island, Tennessee, and Utah) that asked the optional question about specific contraceptive methods used postpartum and achieved an overall weighted response rate of 65% or greater.
Measures
We estimated gestational age using the clinical estimate reported on the birth certificate and categorized gestational age as term births (≥37 weeks) and PTB (34–36 weeks, 28–33 weeks, and ≤27 weeks [extreme PTB]). The gestational age categories were selected a priori.
To describe postpartum contraceptive use, we examined responses to the following questions: “Are you or your husband or partner doing anything now to keep from getting pregnant?” and “What kind of birth control are you or your husband or partner using now to keep from getting pregnant?” Because respondents could report multiple methods, the most effective method of all responses was selected.
Contraceptive use was categorized according to effectiveness.
- •
Highly and moderately effective methods were those with which fewer than 10% of women have an unintended pregnancy within the first year of use: permanent methods (tubal ligation or vasectomy) and LARCs (intrauterine device or contraceptive implant) and moderately effective user-dependent methods (shots, pill, patch, and ring).
- •
Less effective methods were those with which 10% or more of women have an unintended pregnancy within the first year of use: male and female condoms, diaphragm, cervical cap, sponge, emergency contraception, rhythm, withdrawal, and other.
- •
No contraceptive method (nonuse) was coded when women answered no to current contraceptive use or reported that their only method was abstinence.
All nonusers were asked about reasons for not using contraceptives, specifically, “What are your reasons or your husband’s or partner’s reasons for not doing anything to keep from getting pregnant now?” Multiple close-ended responses were allowed and included the following responses: “I am not having sex,” “I want to get pregnant,” “I don’t want to use contraception,” “My husband or partner doesn’t want to use anything,” “I don’t think I can get pregnant,” “I can’t pay for birth control,” and “other reason.” Respondents also had the option to write in a response.
Analysis
Of 37,089 respondents, 4678 (12.6%) were excluded because of a current pregnancy or hysterectomy (0.6%) or missing information on postpartum contraceptive method (2.8%) or covariates (9.2%). Women who reported abstinence were included in our analysis as nonusers because 90% of postpartum women resume sexual activity by 4 months postpartum and hence are at risk for pregnancy.
Our final analytical sample included 32,411 nonpregnant women with recent live births and data on all covariates. We estimated the prevalence of maternal characteristics and postpartum contraceptive use (highly or moderately effective methods, less effective methods, and no method) stratified by PTB group and used χ 2 tests to assess the statistical differences ( P < .05).
Using a multivariable logistic regression to control for potential confounders, we evaluated the associations between recent PTB and 2 measures of postpartum contraceptive use: (1) any highly or moderately effective contraceptive method (vs less effective methods and no method) and (2) highly effective user-independent methods (vs moderately effective methods, less effective methods, and no method). Potential confounders identified from the literature were age, race/ethnicity, education, income, health insurance, marital status, prenatal care, parity, and smoking.
We conducted sensitivity analyses of the multivariable models among subgroups of women who expressed no concerns about potential infertility (n = 32,309) and multiparous women, additionally controlling for pregnancy intention and previous PTB (n = 14,068). We also examined associations between infant death and contraceptive use among women who had extreme PTB (n = 517).
We assessed effect modification by insurance type at delivery (private, Medicaid, other, none) for the full sample by examining statistical significance of interaction terms between PTB and insurance type for both outcomes ( P < .05). All analyses were conducted using weighted data and STATA 13 (2013; StataCorp LP, College Station, TX) to adjust for the complex survey design, thus allowing for population inferences.
Results
A larger percentage of excluded women had recent PTB (10.4%) compared with the analytical sample (8.7%; P = .005) and reported no postpartum contraceptive method (24.4% vs 8.7%; P < .0001). Of excluded women, recent PTB was even higher among the subset of excluded pregnant women (14.1%). Excluded women were also more likely to be young, minority race/ethnicity, low income, and unmarried and reported late entry into prenatal care and less likely to be college educated or privately insured.
Prevalence of having a recent PTB was 8.7%, of which extreme PTB accounted for less than 1% ( Table 1 ). Compared with women who had recent term births, a higher percentage of women with recent PTBs were non-Hispanic black, low income, unmarried, and current cigarette smokers. Additionally, a higher percentage of women with recent PTBs reported having had 3 or more previous live births, previous PTB, no prenatal care for the most recent live birth, and death of the most recent live-born infant. A smaller percentage of women with recent PTB were college educated or had private insurance (compared with women with recent term births). Infant death varied by gestational age of the recent birth: 41% at 27 weeks or less, 2% at 28–33 weeks, and less than 1% for 34 weeks or longer.
| Maternal characteristics | Recent term birth | Recent preterm birth | P value (χ 2 ) | ||||||
|---|---|---|---|---|---|---|---|---|---|
| ≥37 wks (n = 25,946) | 34–36 wks (n = 3987) | 28–33 wks (n = 1872) | ≤27 wks (n = 606) | ||||||
| Weighted, % | 95% CI | Weighted, % | 95% CI | Weighted, % | 95% CI | Weighted, % | 95% CI | ||
| 91.3 | 90.9–91.6 | 6.46 | 6.1–6.8 | 1.72 | 1.6–1.8 | 0.56 | 0.5–0.6 | ||
| Age, y | .0330 | ||||||||
| ≤19 | 8.4 | 7.9–8.9 | 8.3 | 6.8–10.1 | 10.1 | 8.3–12.4 | 7.9 | 5.5–11.2 | |
| 20–24 | 23.5 | 22.8–24.3 | 23.0 | 20.7–25.5 | 23.6 | 20.8–26.7 | 30.4 | 24.9–36.6 | |
| 25–29 | 31.5 | 30.6–32.3 | 29.8 | 27.2–32.6 | 26.5 | 23.5–29.7 | 24.9 | 20.5–29.9 | |
| 30–34 | 24.3 | 23.5–25.1 | 24.5 | 22.1–27.1 | 23.9 | 20.8–27.4 | 26.1 | 21.1–31.8 | |
| ≥35 | 12.3 | 11.8–12.9 | 14.3 | 12.4–16.3 | 15.8 | 13.7–18.3 | 10.8 | 8.1–14.2 | |
| Race/ethnicity | < .0001 | ||||||||
| White, non-Hispanic | 72.6 | 72.0–73.3 | 71.5 | 69.1–73.7 | 63.8 | 60.3–67.2 | 57.4 | 51.5–63.1 | |
| Black, non-Hispanic | 9.8 | 9.5–10.2 | 14.1 | 12.5–15.9 | 19.6 | 17.1–22.4 | 27.6 | 22.3–33.5 | |
| Other, non-Hispanic | 5.6 | 5.2–6.0 | 5.0 | 4.0–6.2 | 6.9 | 4.5–10.3 | 4.9 | 3.4–7.1 | |
| Hispanic | 12.0 | 11.5–12.5 | 9.5 | 8.2–10.9 | 9.7 | 8.3–11.4 | 10.1 | 7.7–13.2 | |
| Education, highest level | .0001 | ||||||||
| Less than 12th grade | 14.4 | 13.8–15.1 | 16.3 | 14.2–18.7 | 16.7 | 14.4–19.4 | 16.4 | 11.7–22.4 | |
| 12th grade, GED, or high school graduate | 24.6 | 23.8–25.4 | 27.2 | 24.6–29.9 | 29.3 | 26.3–32.5 | 35.7 | 30.1–41.7 | |
| Some college or more | 61.0 | 60.1–61.9 | 56.5 | 53.6–59.3 | 54.0 | 50.5–57.5 | 48.0 | 42.2–53.8 | |
| Household income, dollars | < .0001 | ||||||||
| <10,000 | 20.3 | 19.6–21.1 | 24.8 | 22.3–27.4 | 29.3 | 26.2–32.6 | 29.9 | 24.3–36.1 | |
| 10,000–19,999 | 15.8 | 15.1–16.5 | 16.0 | 13.9–18.3 | 16.2 | 13.3–19.7 | 20.6 | 16.4–25.6 | |
| 20,000–34,999 | 17.3 | 16.7–18.1 | 15.3 | 13.4–17.4 | 16.8 | 14.6–19.4 | 18.9 | 14.7–23.9 | |
| 35,000–49,999 | 11.5 | 10.9–12.1 | 10.0 | 8.6–11.7 | 9.3 | 7.7–11.2 | 8.1 | 5.8–11.0 | |
| ≥50,000 | 35.1 | 34.3–36.0 | 34.0 | 31.3–36.7 | 28.4 | 25.4–31.5 | 22.6 | 18.4–27.5 | |
| Marital status | < .0001 | ||||||||
| Not married | 34.1 | 33.3–35.0 | 38.7 | 35.9–41.6 | 43.6 | 40.1–47.3 | 53.6 | 47.8–59.3 | |
| Married | 65.9 | 65.0–66.7 | 61.3 | 58.4–64.1 | 56.4 | 52.8–59.9 | 46.4 | 40.7–52.2 | |
| Health insurance at delivery | .0001 | ||||||||
| Uninsured | 1.8 | 1.5–2.1 | 2.0 | 1.4–2.9 | 2.0 | 1.2–3.3 | 3.2 | 1.8–5.8 | |
| Medicaid | 42.8 | 41.9–43.7 | 43.5 | 40.6–46.3 | 49.8 | 46.3–53.3 | 53.1 | 47.2–53.8 | |
| Other b | 8.2 | 7.7–8.7 | 10.3 | 8.5–12.4 | 9.0 | 7.3–10.9 | 9.5 | 7.1–12.8 | |
| Private | 47.3 | 46.4–48.2 | 44.3 | 41.4–47.1 | 39.3 | 36.0–42.7 | 34.2 | 29.1–39.7 | |
| Prenatal care | < .0001 | ||||||||
| None | 0.4 | 0.3–0.5 | 1.6 | 1.0–2.5 | 2.8 | 1.9–4.2 | 3.9 | 1.9–7.9 | |
| Late | 17.1 | 16.5–17.9 | 13.4 | 11.8–15.2 | 17.5 | 15.0–20.4 | 17.8 | 13.5–23.1 | |
| Early | 82.4 | 81.7–83.1 | 85.0 | 83.1–86.7 | 79.7 | 76.7–82.4 | 78.3 | 72.7–83.1 | |
| Current smoker | < .0001 | ||||||||
| Yes | 18.9 | 18.1–19.6 | 24.7 | 22.1–27.5 | 23.0 | 20.2–26.1 | 23.6 | 18.9–29.1 | |
| No | 81.2 | 80.4–81.9 | 75.3 | 72.5–77.9 | 77.0 | 73.9–79.8 | 76.4 | 70.9–81.2 | |
| Previous live births, n | .0001 | ||||||||
| 0 | 39.4 | 38.5–40.3 | 38.3 | 35.5–41.1 | 43.4 | 40.0–46.9 | 41.2 | 35.7–46.9 | |
| 1–2 | 49.7 | 48.8–50.6 | 47.4 | 44.5–50.3 | 41.8 | 38.3–45.4 | 43.3 | 37.7–49.2 | |
| ≥3 | 10.9 | 10.3–11.5 | 14.3 | 12.4–16.5 | 14.8 | 12.6–17.3 | 15.5 | 10.8–21.8 | |
| Previous preterm birth c | < .0001 | ||||||||
| Yes | 4.0 | 3.5–4.6 | 14.2 | 11.0–18.1 | 17.8 | 12.5–24.8 | 13.6 | 9.4–19.3 | |
| No | 96.0 | 95.4–96.5 | 85.8 | 81.9–89.0 | 82.2 | 75.2–87.5 | 86.4 | 80.8–90.6 | |
| Recent live-born baby died | < .0001 | ||||||||
| Yes | 0.1 | 0.1–0.2 | 0.3 | 0.1–0.5 | 2.0 | 1.3–4.6 | 40.7 | 34.7–47.1 | |
| No | 99.9 | 99.8–99.9 | 99.7 | 99.5–99.9 | 98.0 | 95.4–98.7 | 59.3 | 52.9–63.4 | |
a Based on Pregnancy Risk Assessment Monitoring System, 9 US states, 2009–2011 (n = 32,411)
b Includes Tri-Care, other military, Indian Health Services, state-specific Children’s Health Insurance Plan, Children’s Health Insurance Plan
Postpartum contraceptive use varied by gestational age of the most recent birth ( Figure , P < .0001). Nearly half of all women with a recent PTB (39%) reported using a less effective method or no method at all. Except for those with a recent extreme preterm birth, most women reported using moderately effective user-dependent methods. Women with extreme PTB most frequently reported no method (31%), at approximately twice the prevalence that was reported by all other groups of women (15–16%).
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
