Original research••

Prevalence and associated factors with contraceptive use 1 year after childbirth: a cross-sectional analysis of the pregnancy risk assessment monitoring system

•,,,,

...

Abstract

Background Contraceptive use after childbirth can reduce risk of unplanned pregnancy. This study estimated the prevalence of contraception use in the 1 year after childbirth. It further examined factors associated with use of the most and moderately effective forms of reversible contraception in the 1-year period after childbirth.

Methods This cross-sectional study used Phase 8 data from the Pregnancy Risk Assessment Monitoring System (PRAMS) (2016–2020) in the USA (N=194 032). Prevalence, overall and by type, was estimated nationally, by state and by age. Multinomial logistic regression models estimated associations with the most and moderately effective forms of reversible contraception (vs non-use). The most effective reversible contraception consisted of long-acting reversible contraception (LARC); moderately effective reversible contraception was a combined category of the injection, pill, patch, vaginal ring and diaphragm.

Results Moderately effective reversible contraception was the most common method used (24.8%) followed by no contraception (23.9%) and then use of the most effective reversible contraception (17.0%). Variation by age and state in LARC use was demonstrated. Having had a postpartum visit was the factor most strongly associated with LARC use in the 1 year after childbirth (adjusted OR (aOR): 4.96, 95% CI: 4.91 to 5.01). In contrast, being married at the time of the survey was the factor most strongly associated with use of moderately effective reversible contraception (aOR: 1.55, 95% CI: 1.54 to 1.56).

Conclusion Higher prevalence of no contraception or less effective forms of birth control presents an opportunity for greater LARC adoption, if preferred, especially at the postpartum visit.

What is already known on this topic

Contraceptive counselling and use of effective contraceptives in the first year after childbirth can reduce unintended pregnancies.

What this study adds

Moderately effective contraception followed by no contraceptive use was more frequently reported than the most effective reversible form of contraception. Variation by age and state was evident, which has implications for clinical practice and policy surrounding family planning services. The postpartum visit presents an opportunity for shared decision-making, education and increased access to the most effective forms of reversible contraception.

How this study might affect research, practice or policy

The higher prevalence of no contraception and higher use of less effective contraceptives compared with the most effective types presents an opportunity for greater long-acting reversible contraception education and adoption, if preferred, at the postpartum visit.

Introduction

In the USA, access to and use of affordable or no-cost contraception in the first year after childbirth is a critical component of post-delivery care.1 In the USA, where disparities in contraceptive access persist, healthcare providers play a key role in reducing unplanned pregnancies by educating patients about their postpartum contraceptive options and by empowering them to manage their fertility.1 2

Long-acting reversible contraceptives (LARCs), such as intrauterine devices and the subdermal etonogestrel implant, are the most effective reversible contraceptive options available, with <1 pregnancy per 100 women in a year with typical use.3 Moderately effective contraceptives include short-acting reversible contraceptives such as injectables, pills, patches, rings and diaphragms4; these methods are associated with 6–12 pregnancies per 100 women in a year with typical use. Failure rates for typical use of less (ie, condoms, sponge, cervical cap) and least effective (ie, spermicide, natural family planning, withdrawal) methods are between 12% and 28%.1 5

The purpose of this study was to estimate the prevalence of contraception use in the USA up to 1-year post-childbirth; and to examine factors associated with use of the most and moderately effective forms of reversible contraception among individuals up to 1-year post-delivery.

Methods

Design

This cross-sectional study used secondary data from the Centers for Disease Control and Prevention’s (CDC) Pregnancy Risk Assessment Monitoring System (PRAMS). PRAMS is a state-level, population-based surveillance survey on maternal behaviours, attitudes and experiences before, during and shortly after pregnancy. Only individuals with a recorded live birth are eligible and contacted for inclusion. Participating local health departments across the USA collect data via a mixed-mode mail/telephone survey (self-report) and information gleaned from birth certificates for submission to the CDC. The CDC cleans, cumulates and weights the data to derive a nationally representative sample of all individuals who delivered a live born infant in the USA (‘the incident pregnancy’). For a complete description of the PRAMS methodology and access to the data, see Shulman et al.4 Patient and public involvement was not appropriate for this study.

For the current study, PRAMS Phase 8 data were used. These data cover live births from January 2016 through December 2020 and include 44 states as well as the District of Columbia, New York City and Puerto Rico. Inclusion criteria for this study were (1) aged between 18 and 44 years at the time of the incident pregnancy, (2) self-reported that they were not pregnant at the time of the survey and (3) a response to survey item 45: ‘What kinds of birth control are you or your husband or partner using now to keep from getting pregnant?’ or reported that they ‘had my tubes tied’ or were ‘not having sex’ in response to item 44: ‘What are your reasons or your husband’s or partner’s reasons for not doing anything to keep from getting pregnant now?’. A total of 194 032 participants met these criteria.

Key variables

Current contraceptive method

The primary outcome variable was type of contraceptive method—focused on use versus non-use—as reported at PRAMS survey completion which extends 3–12 months after giving birth. Six types of contraceptive groups were derived from responses to item 45: ‘What kind of birth control are you or your husband or partner using now to keep from getting pregnant?’ These contraceptive types included: (1) most effective reversible (ie, LARCs), (2) most effective permanent (ie, sterilisation via vasectomy or tubal ligation, including ‘tubes tied’ responses from item 44), (3) moderately effective (ie, pills, ring, patch, injectables, diaphragm), (4) less effective (ie, condoms, sponge or cervical cap use extracted from qualitative responses listed in the ‘other’ option of item 44), (5) least effective (ie, withdrawal, natural family planning) and (6) no contraceptive use (ie, ‘no’ response to all other contraceptive types listed above or ‘I am not having sex’ response from item 44 or item 45). Respondents who reported use of multiple forms of contraception were categorised into the most effective method reported.

Demographic characteristics

including mother’s age at delivery, marital status, education, race and ethnicity, mode of delivery (C-section or vaginal), and if the mother received support from women, infants and children were determined by information from the infant’s birth certificate. Household income, ever experiencing intimate partner violence (IPV), if the incident pregnancy was intentional, unintentional or mistimed, and if respondents used birth control before the incident pregnancy were gathered from the PRAMS questionnaire. Regional access to contraception was assessed using urbanicity and state of birth, as reported in the PRAMS analytic file. All demographic variables were categorical.

Clinical characteristics

Included maternal comorbidities or delivery complications reported in the PRAMS questionnaire or from birth certificate data. Specifically, this study examined (1) body mass index (BMI) (ie, underweight, normal, overweight, obese); (2) type 1 or type 2 diabetes before the incident pregnancy and/or gestational diabetes diagnosed during the incident pregnancy, (3) high blood pressure or hypertension in the 3 months before the incident pregnancy, and/or pre-eclampsia or eclampsia during the incident pregnancy, (4) postpartum depression in the incident pregnancy and (5) smoking cigarettes or (6) drinking alcohol in the last 3 months of the incident pregnancy. Responses were binary (0=no, 1=yes).

Healthcare resource utilisation (HCRU), used as a proxy for point-of-access to contraception and counselling, included respondents’ health insurance status at the time of survey (‘now’), which was recoded into the following mutually exclusive categories: (1) employer-sponsored insurance, (2) Medicaid, (3) no insurance, (4) multiple and/or other insurance types. Additional indicators included whether the respondent received prenatal care in the first trimester, whether they attended a postpartum check-up defined as ‘a regular check-up a woman has about 4–6 weeks after she gives birth’ and whether they delivered in a state offering Medicaid reimbursement for immediate postpartum LARCs at the time of birth.6 Responses were binary (0=no, 1=yes).

Analysis

Analyses were conducted in SAS Viya (SAS Institute, Cary, North Carolina, USA) to account for the survey design and to adjust for sampling bias using the weighting variable included in the PRAMS dataset. All reported results use weighted analyses. Descriptive statistics described the sample and χ² tests of significance with Rao-Scott correction (two-tailed, α=0.05) assessed differences in sample characteristics by contraceptive users and non-users.

Estimation of prevalence was determined as follows: the number of respondents using contraception divided by the number of respondents in the analytical sample; prevalence was assessed for the overall national sample and by state and age group (18–29, 30–39, 40–44 years).

Multinomial logistic regression was used to estimate associations with contraceptive type, focused on the most and moderately effective forms of reversible contraception. Adjusted models included socio-demographic, clinical characteristics and HCRU. Interactions were not examined. Only respondents with complete data were included in the regression analyses (unweighted N=147 103; weighted N=7 751 731).

Results

Sample description

The weighted sample included 9 787 137 respondents. Most respondents were married, between 25 and 34 years of age, had greater than high school education, lived in an urban setting and were non-Hispanic white (table 1). Comparisons between contraceptive users and non-users show that demographic and socioeconomic factors were associated with contraceptive use. White, higher-income individuals and those unexposed to IPV were more likely to use contraception after the incident pregnancy. Previous use of birth control, an unintended or mistimed incident pregnancy and delivery by C-section was also associated with a higher likelihood of contraceptive use. Rural residency and previous live births were associated with increased contraceptive use.

Table 1

•

Socio-demographic characteristics reported at the time of survey completion, overall and by use versus non-use of contraception in the 1-year post-childbirth period, weighted

Prevalence of contraceptive use

Weighted prevalence estimates indicate that the most common contraceptive method used were moderately effective types (24.8%; 95% CI: 24.49% to 25.11%), followed by no contraceptives (23.9%; 95% CI: 23.63% to 24.23%). Most effective reversible contraceptives had the next largest prevalence (17.0%; 95% CI: 16.72% to 17.26%), followed by less effective contraceptives (16.8%; 95% CI: 16.48% to 17.02%) and the most effective permanent contraceptive (11.1%; 95% CI: 10.85% to 11.30%). The least effective contraceptives were also the least commonly used type (6.5%; 95% CI: 6.28% to 6.63%). See online supplemental data table 1 for estimates stratified by age group.

Contraceptive use varied by age (figure 1). No contraception use remained the most common category across all age groups, except for 18–29 years old where a greater proportion used moderately effective contraceptives (29.1%). Respondents 18–29 years of age chose the most effective reversible contraceptives more often than their older age counterparts. Sterilisation use increased with age. Less effective contraceptives were most common among respondents 30–39 years (17.8%), followed by 18–29 years old (16.0%) and 40–44 years old (13.2%). The least effective contraceptives remained the least prevalent across all age groups.

Figure 1

Request permissions

Prevalence of contraception use in the 1-year period after childbirth, overall and by age.

Contraceptive prevalence by state reveals significant regional variation (figure 2). Prevalence of the most effective reversible contraceptives was highest in Oregon (28.3%) and Utah (28.1%), while New Jersey (9.6%) and Puerto Rico (8.2%) had the lowest prevalence. Puerto Rico exhibited a very high prevalence of the most effective permanent contraceptive (28.9%) while New York City had one of the lowest (5.3%). West Virginia (34.5%) and Mississippi (35.8%) had the highest use of moderately effective contraceptives, compared with lower prevalence in Alaska (19.4%) and Puerto Rico (14.9%). Finally, New York City had the highest proportion of contraceptive non-users (34.0%) while Utah had one of the lowest (15.1%).

Figure 2

Request permissions

Prevalence of contraceptive methods by state 1-year period post-delivery weighted (N=9 787 137).

Multivariable associations with contraceptive types

Compared with contraceptive non-users, having had a postpartum visit was the variable most strongly associated with use of most effective reversible contraceptives (adjusted OR (aOR): 4.96, 95% CI: 4.91 to 5.01) and being married was the variable most strongly associated with use of moderately effective reversible contraception (aOR: 1.55, 95% CI: 1.54 to 1.56) (table 2). Respondents with one or more previous live births were more likely to use the most effective reversible contraceptives (aOR=1.57, 95% CI: 1.56 to 1.58) and moderately effective methods (aOR=1.25, 95% CI: 1.24 to 1.26) compared with women with no previous live births. Among black/African American versus white respondents and those experiencing IPV versus no IPV, there was a decrease in the likelihood of using the most effective reversible methods (aOR=0.68, 95% CI: 0.67 to 0.68; aOR=0.68=3, 95% CI: 0.63 to 0.64; respectively) and moderately effective contraceptives (aOR=0.66, 95% CI: 0.66 to 0.67; aOR=0.50, 95% CI: 0.49 to 0.41, respectively). Asian respondents were less likely to use the most effective reversible contraceptives (aOR=0.43, 95% CI: 0.43 to 0.44) and slightly more likely to use moderately effective methods (aOR=1.01, 95% CI: 1.00 to 1.02) compared with white respondents. Hispanic respondents were more likely to use the most effective reversible methods (aOR=1.58; 95% CI: 1.56 to 1.59) than non-Hispanic respondents. Respondents with a BMI of 30 or above (‘obese’) had an increased likelihood of using the most effective reversible methods (aOR=1.29, 95% CI: 1.28 to 1.30) compared with individuals with ‘normal’ BMI. Similarly, hypertensive respondents were more likely to use most effective reversible contraceptives (aOR=1.13, 95% CI: 1.12 to 1.14) and slightly less likely to use moderately effective methods (aOR=0.98, 95% CI: 0.97 to 0.99). Respondents who reported drinking alcohol in the last 3 months of pregnancy had an increased likelihood of using the most effective reversible contraceptives (aOR=1.21, 95% CI: 1.2 to 1.21), whereas smoking was associated with a decreased likelihood (aOR=0.92, 95% CI: 0.91 to 0.93).

Table 2

•

Multinomial logistic regression results examining contraceptive types ordered by effectiveness compared with no contraceptive use, weighted sample (N=7 751 731)

Discussion

This cross-sectional study estimated the prevalence of contraception use in the first year after childbirth among individuals who gave birth to a live infant in the USA between 2016 and 2020, a period marked by changes in federal and state policies affecting contraception access (eg, Affordable Health Care Act, state-level Medicaid expansion and policy changes) and adoption of professional and quality guidelines (eg, National Quality Foundation 2904).7 It further estimated associations with use of the most and moderately effective forms of reversible contraception in the 1-year period after delivery.

The findings regarding the higher prevalence of no contraception or less effective contraceptives present an opportunity for greater LARC adoption overall, if preferred, and particularly among individuals over age 29 years who were less likely to use these forms of contraceptives. While the choice of contraceptive is a multifactorial decision based on patient and physician perceptions, experiences and personal preferences,6 8 9 this study found that attending a postpartum visit was most strongly associated with LARC use, elevating the importance of patient education and shared decision-making during the postpartum visit. A systematic review found that patient preferences for contraceptive counselling include comprehensive and personalised information, respect for patient autonomy, positive provider-patient relationships and diverse preferences for the context in which counselling occurs (eg, setting and prenatal and postpartum timing).10 Ensuring that individuals can attend their postpartum visit, via automated scheduling and tailored messages, may be critically important as well.11

Variation in the use of most effective reversible and moderately effective contraceptives by state highlights the need for tailored policy and practice changes to address disparities in contraceptive access and counselling. Despite the higher up-front acquisition costs, several studies have reported that LARCs are the most cost-effective and cost-saving form of contraception available.12–16 Access to and availability of postpartum contraceptives play a crucial role in their utilisation.10 17 Study findings regarding state-level variation in contraceptive use may reflect differences in funding and programmes for family planning services. Research shows that LARC device unavailability and the cost or reimbursement of devices contribute to access barriers affecting both patients and providers.17–22 Other studies have found that policies that increase insurance coverage improve access to and lead to greater use of contraceptive services, particularly for LARC methods.23 24 Policy and practice changes that address barriers beyond the immediate postpartum period, like facilitators and clinical protocols enabling same-day provision of methods,7 25 can improve access for all.

Strengths and limitations

This study has several limitations. First, the cross-sectional design means that the reported associations are exploratory and correlational; however, a strength of the study was the large and nationally representative sample which increases generalisability, statistical power and provides a more accurate estimate of associations across diverse subpopulations, reducing the likelihood that the observed relationships are due to chance or sampling bias. Second, key confounding factors, particularly those influencing medical professionals’ decisions to recommend certain contraceptives, were not measured or assessed. Relatedly, the survey methodology did not use a standard timeframe for assessing contraceptive use; respondents were asked if they were ‘doing anything now to keep from getting pregnant’, which could have been 3–12 months post partum.4 Furthermore, PRAMS did not collect data on the nature of contraceptive counselling received, the methods offered but declined and preferences, making it difficult to discern whether the variation in contraceptive use reflects personal preferences, medical norms or access issues. This study also found a high proportion (89.4%) of respondents attended a postpartum visit, which contrasts with studies reporting that as much as 13–45% of individuals miss their postpartum visit,26 27 thus potentially reflecting a sample in the current study more prone to healthcare-seeking behaviour. Inherent limitations of survey research, such as self-reporting and recall bias, data entry errors and non-response bias, must be acknowledged with these data. The CDC takes steps to mitigate these issues through weighted sampling and quality control measures, including data entry verification, supervisor monitoring of telephone surveys and maintaining low non-response rates (1–2% overall, 6% for income-related questions).4 Despite these limitations, the study findings have important implications for strategies to increase the use of the most effective contraceptives among individuals in the 1-year period post-delivery. Future studies that replicate this study with a post-COVID-19 sample are warranted as research suggests the pandemic decreased postpartum visit attendance despite an increase in telehealth visit options.28 29

Conclusions

This study highlights the opportunity that postpartum contraceptive counselling presents for increasing use of the most effective forms of reversible contraception. Education and shared decision-making between patients and providers about contraceptive options available during the postpartum visit is an important strategy for reducing unplanned pregnancies in the first year after delivery.