Objective
Family planning is recommended as a strategy to prevent adverse birth outcomes. The potential contribution of postpartum contraceptive coverage to reducing rates of preterm birth is unknown. In this study, we examine the impact of contraceptive coverage and use within 18 months of a birth on preventing preterm birth in a Californian cohort.
Study Design
We identified records for second or higher-order births among women from California’s 2011 Birth Statistical Master File and their prior births from earlier Birth Statistical Master Files. To identify women who received contraceptive services from publicly funded programs, we applied a probabilistic linking methodology to match birth files with enrollment records for women with Medi-Cal or Family Planning, Access, Care, and Treatment Program (PACT) claims. The length of contraceptive coverage was determined through applying an algorithm based on the specified method and the quantity dispensed. Preterm birth was defined as a birth occurring <37 weeks’ gestation, and calculated from the medical record. We further examined differences in preterm birth using subcategories defined by the World Health Organization: extremely preterm (<28 weeks); very preterm (28 to <32 weeks); and moderate to late preterm (32 to <37 weeks). We built a multivariable regression model to examine the effect of contraceptive coverage on the odds of a preterm birth and control for key covariates.
Results
The cohort consisted of 111,948 women who were seen at least once by a Medi-Cal or Family PACT provider within 18 months of delivery. Of the cohort, 9.75% had a preterm birth. Contraceptive coverage was found to be protective against preterm birth. For every month of contraceptive coverage, odds of a preterm birth <37 weeks decrease by 1.1% (odds ratio, 0.989; 95% confidence interval, 0.986–0.993).
Conclusion
Improving postpartum contraceptive use has the potential to reduce preterm births.
See related editorial, page 602
Preterm birth (<37 weeks’ gestational age) remains a significant cause of neonatal morbidity and mortality in the United States and globally. Preterm birth is now the second most common cause of death in children age <5 years, with an estimated 1.1 million infants dying annually from complications of preterm birth. It is also an important factor in long-term morbidity, such as cognitive, visual, and learning impairments. Reduction in mortality from preterm birth, and prevention of preterm birth, is the focus of national and international efforts. However, although prevention of preterm birth is a public health priority, international estimates reveal it to be a persistent and substantial problem in a wide range of countries. In the United States, the rate of preterm birth increased by nearly 30% from 1981 through 2006. In 2007, this trend began to reverse, with a decline from a high of 12.8% in 2006 to 11.73 in 2011. Despite this promising trend in preterm birth rates, the overall number of premature infants born in the United States remains higher than in any single year from 1981 through 2006, and significant racial and ethnic disparities exist.
The causes of preterm birth are multifactorial and poorly understood, making prevention challenging. A range of different interventions such as progesterone supplementation, cerclage, smoking cessation, reduction in transfer of multiple embryos with assisted reproductive technology, and eliminating nonmedically indicated inductions have been evaluated for their potential to reduce preterm birth. Even with full implementation of these complex interventions however, the estimated reduction in rates of preterm birth would be small. To effectively and equitably reduce preterm birth rates, there is an urgent need to address the underlying social determinants of poor reproductive health. Individual socioeconomic status, inequalities in income and education, social policies, neighborhood deprivation, and intermediary factors such as health behaviors are examples of social determinants of health.
Influencing health behaviors through preconception care is one strategy that has been adopted to try and improve reproductive outcomes. Attention has focused on the potential for preconception and interconception care to reduce preterm birth by promoting birth spacing and preventing unintended pregnancy, in particular among adolescents. Optimizing interpregnancy interval through birth spacing is one proposed solution for reducing adverse birth outcomes. Evidence suggests that risks of preterm birth, low birthweight, and small-for-gestational-age infants are minimized when interpregnancy intervals are between 18–23 months. However, optimizing interpregnancy intervals alone is not enough to reduce adverse birth outcomes; it has also been proposed that interpregnancy intervals are a proxy for maternal variables that cannot be easily monitored, such as attitudes, lifestyle, and cultural norms. Analyses that compare interpregnancy intervals and birth outcomes from the same woman control for these factors, and can clarify the association between adverse birth outcomes and different variables.
Data from California provide an opportunity to examine the relationship between postpartum contraception and birth outcomes in detail. In 2011, the costs of half of all births in California (50.4%) were paid for by Medicaid (Medi-Cal). Women who have their delivery paid by Medi-Cal continue to be eligible for postpartum and pregnancy-related services, through the end of the second month after the month of delivery, and may then be eligible to receive family planning services through Medi-Cal, or California’s family planning program, Family Planning, Access, Care, and Treatment Program (PACT).
Previous research from California has established the importance of publicly funded family planning programs in establishing optimal interpregnancy intervals through ensuring access to postpartum contraception and promoting use of highly effective methods of contraception. Women with a postpartum visit were 33% less likely to have a short interpregnancy interval than women who did not (relative risk 0.67). The protective effect varies by contraceptive method dispensed. Women receiving the most effective forms of contraception (long-acting, reversible methods, eg, the intrauterine device and implant) had significantly increased odds of achieving an optimal birth interval. Because of the uncertainty of a causal mechanism between interpregnancy intervals and preterm birth, we wanted to directly examine the potential impact of provision of postpartum contraception in reducing preterm birth.
Family planning is recommended as a strategy to prevent adverse birth outcomes. The potential contribution of postpartum contraceptive coverage to reducing rates of preterm birth is unknown. We hypothesized that postpartum contraception would be associated with a significant reduction in preterm birth. In this study, we examine the impact of contraceptive coverage and use within 18 months of a birth on preventing preterm birth in a Californian cohort.
Materials and Methods
The data analysis was approved by the University of California, San Francisco, Committee of Human Subjects Approval and the California State Committee of Human Subjects Protection. We built a cohort of women aged 12–44 years who received publicly funded contraceptive services in the 18 months after birth. We identified records for second or higher-order births among women from California’s 2011 Birth Statistical Master File and their prior births from earlier Birth Statistical Master Files. Outcomes are reported for the 2011 birth. The birth immediately prior to the 2011 birth is referred to as the “index birth.” Women whose index births occurred before Jan. 1, 2005, or outside California were excluded. Other exclusions were: multiple births, unknown last menstrual period, missing birth date data, documented sterilization at time of index birth, birth intervals of <30 days, and missing or improbable maternal age (eg, age <12 years at the time of the birth).
To identify women who received contraceptive services from publicly funded programs, we applied a probabilistic linking methodology to match Birth Statistical Master Files maternal data with enrollment records for women with Medi-Cal or Family PACT claims. The linking algorithm determines whether a pair of records from 2 disparate data files belongs to the same person.
We used the state’s Management Information System/Decision Support System to analyze administrative Medi-Cal and Family PACT clinic and pharmacy claims, and encounter data, to identify the provision of contraceptive methods. Women were defined as receiving contraception if they had at least 1 Medi-Cal or Family PACT claim for receipt of a contraceptive method. Emergency contraception was excluded from the definition of method of contraception received.
Variables
This study explores whether contraceptive coverage and use within 18 months of the index birth were associated with decreased odds of a preterm birth. Contraceptive coverage estimates the amount of contraceptive supply that a woman received. The length of coverage is determined through applying an algorithm based on the specified method and the quantity dispensed (eg, the number of pill packs or condoms distributed) from pharmacy and on-site claims during the study period. For cases of method switching, coverage was calculated on the aggregate of both methods without double counting periods of overlap. In cases of multiple contraceptive method provision, coverage was estimated based on the most effective method. The cutoff and maximum length of coverage was set at 18 months from a woman’s index birth. For long-acting reversible contraceptives (LARC) such as the intrauterine device and implant, unless a removal claim was found, we assigned the maximum length of coverage. The length of coverage was summed across service dates from the first postpartum visit until the 18-month cutoff.
Preterm birth was defined as a birth occurring <37 weeks’ gestation, and calculated from the medical record. Preterm birth was further examined using subcategories defined by the World Health Organization: extremely preterm (<28 weeks); very preterm (28 to <32 weeks); and moderate to late preterm (32 to <37 weeks). Births with ≥37 weeks’ gestational length were defined as term births.
Client demographics for the mother were determined from the information recorded in the 2011 Birth Statistical Master File. Demographic variables consisted of education level (less than high school graduate, high school graduate/some college, or college graduate or more), race/ethnicity (white, black, Hispanic, Asian/Pacific Islander, Native American, or other/unknown), nativity (United States or foreign-born), age at index birth (continuous variable), and parity (2 births or >2 births). Univariate analyses were conducted on the demographic variables to examine the distribution of the cohort.
Variables
This study explores whether contraceptive coverage and use within 18 months of the index birth were associated with decreased odds of a preterm birth. Contraceptive coverage estimates the amount of contraceptive supply that a woman received. The length of coverage is determined through applying an algorithm based on the specified method and the quantity dispensed (eg, the number of pill packs or condoms distributed) from pharmacy and on-site claims during the study period. For cases of method switching, coverage was calculated on the aggregate of both methods without double counting periods of overlap. In cases of multiple contraceptive method provision, coverage was estimated based on the most effective method. The cutoff and maximum length of coverage was set at 18 months from a woman’s index birth. For long-acting reversible contraceptives (LARC) such as the intrauterine device and implant, unless a removal claim was found, we assigned the maximum length of coverage. The length of coverage was summed across service dates from the first postpartum visit until the 18-month cutoff.
Preterm birth was defined as a birth occurring <37 weeks’ gestation, and calculated from the medical record. Preterm birth was further examined using subcategories defined by the World Health Organization: extremely preterm (<28 weeks); very preterm (28 to <32 weeks); and moderate to late preterm (32 to <37 weeks). Births with ≥37 weeks’ gestational length were defined as term births.
Client demographics for the mother were determined from the information recorded in the 2011 Birth Statistical Master File. Demographic variables consisted of education level (less than high school graduate, high school graduate/some college, or college graduate or more), race/ethnicity (white, black, Hispanic, Asian/Pacific Islander, Native American, or other/unknown), nativity (United States or foreign-born), age at index birth (continuous variable), and parity (2 births or >2 births). Univariate analyses were conducted on the demographic variables to examine the distribution of the cohort.
Multivariable Model
We constructed a multivariate logistic model examining the relationship of contraceptive coverage with the outcome of preterm birth. Contraceptive coverage was defined as a continuous variable from 0–18 months. We controlled for demographic variables, including education level, race/ethnicity, nativity, age at index birth, and parity. We used software (SAS, version 9.2, PROC LOGISTIC; SAS Institute, Cary, NC) for all analyses.
Results
Sample characteristics
The cohort consisted of 111,948 women who were seen at least once by a Medi-Cal or Family PACT provider within 18 months of delivery ( Table 1 ). Of these women, 9.8% had a preterm birth. Among women with a preterm birth, 86.9% were moderate to late preterm, 8.6% very preterm, and 4.5% extremely preterm births. In the study cohort, nearly 95% of the births were normal-weight infants.
| Characteristic | n 111,948 | Total percentage |
|---|---|---|
| Preterm birth | ||
| Term | 101,034 | 90.25 |
| Moderate to late preterm (32–37 wk) | 9483 | 8.47 |
| Very preterm (28–32 wk) | 939 | 0.84 |
| Extremely preterm (<28 wk) | 492 | 0.44 |
| Birthweight | ||
| Normal (>2500 g) | 106,136 | 94.81 |
| Low (<2500 g) | 4913 | 4.39 |
| Very low (<1500 g) | 478 | 0.43 |
| Extremely low (<1000 g) | 421 | 0.38 |
| Patient characteristics | ||
| Age at index birth, y | ||
| <20 | 23,930 | 21.38 |
| 20–29 | 71,289 | 63.68 |
| 30–39 | 16,400 | 14.65 |
| >40 | 329 | 0.29 |
| Race/ethnicity | ||
| White | 16,198 | 14.47 |
| Hispanic | 79,087 | 70.65 |
| black | 8432 | 7.53 |
| Asian/Pacific Islander | 5885 | 5.26 |
| Native American | 687 | 0.61 |
| Other/unknown | 1659 | 1.48 |
| Education | ||
| <12th grade | 42,411 | 37.88 |
| High school graduate/some college | 62,352 | 55.7 |
| Bachelor degree or greater | 4209 | 3.76 |
| Data missing | 2976 | 2.66 |
| Country of birth | ||
| United States | 61,294 | 54.75 |
| Foreign born | 50,654 | 45.25 |
| Parity | ||
| 2 births | 51,923 | 46.38 |
| >2 births | 60,016 | 53.61 |
| Data missing | 9 | 0.01 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
