State reproductive rights policies and unintended pregnancy

Abstract

Objective

States’ reproductive health policies may have consequences for the rate of unintended pregnancy. This study analyzed the association of states’ reproductive rights policies with the likelihood of respondents reporting unintended (wanted later and unwanted pregnancies) and unwanted pregnancy.

Study design

This was a cross‐sectional analysis of the Pregnancy Risk Assessment Monitoring System using 2016–2019 data from 43 states and DC. Institute for Women's Policy Research state reproductive rights rankings were used to categorize states as having restricted, some protection, or protected reproductive rights. Poisson regressions were used to estimate the associations of state reproductive rights rankings with unintended and unwanted pregnancies, controlling for individual‐level sociodemographic characteristics. A sensitivity analysis restricted to Medicaid‐covered births was performed.

Results

A total of 154,062 respondents represented a weighted population of 7,826,006 of whom 25.6% reported an unintended and 6.6% an unwanted pregnancy. When compared to states with protected reproductive rights, states with restricted reproductive rights had a 6% higher likelihood of unintended pregnancy (IRR 1.07, 95% CI 1.04, 1.10). This association was amplified when limited to a Medicaid population (13% higher likelihood, IRR 1.13, 95% CI 1.08, 1.17).

Conclusion

State reproductive rights rankings were associated with unintended and unwanted pregnancies. Further restricting reproductive rights at the state level may increase this already major health and economic burden.

1. INTRODUCTION

States’ health policies are intricately linked with reproductive health outcomes. With the Supreme Court decision in Dobbs v Jackson Women's Health Organization, states have independent jurisdiction over the legality of abortion. Even prior to the Dobbs decision, differences in state‐level policies regarding reproductive rights, access to pregnancy and postpartum care, Medicaid eligibility, and social welfare programs have resulted in different health outcomes for women and children across states [1, 2, 3, 4]. For example, more restrictive state‐level reproductive rights, including abortion restrictions, have been associated with higher rates of maternal mortality [5, 6], fetal and infant mortality [6], low birthweight, and larger Black–White disparities in preterm birth and low birthweight [7, 8, 9].

Differences in state legislative reproductive rights environments may also create an environment where people become pregnant when they do not intend to conceive. Nearly half of pregnancies in the United States are unintended, and rates of unintended pregnancy differ by state [10, 11]. The Centers for Disease Control and Prevention (CDC) define an unintended pregnancy as one that is either unwanted (pregnancy occurred when no more children were desired) or wanted later (pregnancy occurred earlier than desired, also known as mistimed) [12]. Unintended pregnancies are associated with serious adverse maternal and child health conditions across the life course, including intimate partner violence, preterm birth, low birthweight, and perinatal depression [13]. Rates of unintended pregnancy differ between various sociodemographic groups, and evidence suggests unintended pregnancy is not solely driven by individual‐level determinants but also influenced by broader societal factors such as lapses in access to quality medical care, structural determinants of health, and systemic racism [14].

Although unintended pregnancy differs by state [10], there is little evidence about the extent to which this variation reflects a state's legislative reproductive rights environment rather than state population differences. This study evaluates this question by determining whether states with more restrictive reproductive rights have higher rates of unintended pregnancy after controlling for individual‐level sociodemographic characteristics. Understanding the association between state‐level reproductive rights and unintended pregnancy in the pre‐Dobbs era is crucial to understanding ongoing legislative trends and how they may be associated with health outcomes.

2. METHODS

2.1. Data source

The study uses population‐weighted data from the Centers for Disease Control and Prevention's Phase Eight Pregnancy Risk Assessment Monitoring System (PRAMS) [15]. Survey data were collected via mail and telephone surveys and linked to birth certificate data between January 2016 and December 2019. Respondents were surveyed within 2–6 months after delivery. A total of 43 states and the District of Columbia (DC) are represented; Arizona, California, Idaho, Nevada, Ohio, South Carolina, and Texas were not included. This study includes individuals with live births, excluding respondents with neonatal or maternal deaths. PRAMS data are deidentified and publicly available data, and thus this study was exempt from review by Institutional Review Board.

2.2. PRAMS pregnancy intention and sociodemographic characteristics

The primary outcome, unintended pregnancy, was categorized based on responses to a PRAMS question about pregnancy intent. Respondents were asked whether they “wanted to be pregnant sooner,” “wanted to be pregnant then,” “wanted to be pregnant later,” “didn't want to be pregnant then or at any time in the future,” and “[weren't] sure what [they] wanted.” [15] Unintended pregnancy was classified as a pregnancy that occurred when the respondent “didn't want to be pregnant then or at any time in the future” (“unwanted pregnancy”) or “wanted to be pregnant later” (“wanted later pregnancy”). Unwanted pregnancy was classified as a pregnancy that occurred when the respondent “didn't want to be pregnant then or at any time in the future.” Unwanted pregnancies were modeled separately from those that were unintended to evaluate for different effects of reproductive rights at the extremes of pregnancy intention. We acknowledge that the use of the terminology “pregnancy intention” imperfectly characterizes pregnancy desire, which is what PRAMS surveys measure. We use “unintended” and “unwanted” as they are concise and describe their derivation above and discuss this further in our limitations.

PRAMS data were used to control individual‐level sociodemographic characteristics that varied widely between states. Household income was categorized as $20,000 or less, $20,001 to $40,000, $40,001 to $60,000, $60,001 to $85,000, or $85,001 or more. Maternal age was categorized as 18–19, 20–24, 25–29, 30–34, 35–39, and 40 or more years. Race and ethnicity were categorized as Non‐Hispanic White, Non‐Hispanic Black, Non‐Hispanic Asian, Hispanic, or other/unknown (non‐Hispanic). Respondents were also asked if they were married, if the PRAMS survey was completed in Spanish, and if they lacked health insurance coverage preconception or had Medicaid coverage (vs. any other insurance coverage) at delivery. Maternal educational attainment was categorized as less than 12 years, 12 years, 13–15 years, or 16 or more years. Births were characterized as multifetal gestation versus singleton, and parity at the time of survey was categorized as nulliparous 1, 2, or 3 or more births. Year of delivery (2016–2019) was included to control analyses for potential secular trends in pregnancy intention. Income was regression‐imputed for 4.45% respondents with missing data using respondent age, race and ethnicity, marital status, and education. Medicaid coverage at the time of delivery was regression‐imputed for 3.2% respondents with missing data, using the same regression imputation equation as used for household income. The final analysis sample excluded respondents with any missing data for any other covariates or pregnancy intention.

2.3. State reproductive rights rankings

States were categorized using the Institute for Women's Policy Research (IWPR) state reproductive policy scores [16, 17]. These composite rankings were based on a weighted composite score for state legislation on the legal right to abortion, legislative opposition to reproductive rights, parental consent or notification and waiting period laws, Medicaid family planning funding and overall coverage, percent of women living in counties with at least one abortion provider, infertility treatment coverage, and sex education laws. Because 2016–2019 study data spanned two state election cycles (although only one presidential election cycle), we used the average of state ranking scores originally published by IWPR in 2015 and republished in 2022 as a proxy for the overall state ranking scores during the study period (Appendix I) [16, 17]. There were few state rank order changes between published 2015 and 2022 rankings (r = 0.87). IWPR does not publish reproductive rights data during our study period. By averaging data from these periods, we aimed to capture state‐level trends in reproductive rights policies.

We first plotted continuous average state IWPR reproductive ranking scores against each respondent's mean state unintended birth percentage, including the subgroup of unintended births that were unwanted. We then defined three groups of states using their mean 2015–2022 IWPR scores, each accounting for an approximate one‐third of the unweighted PRAMS population. States were categorized as restricted (n = 15), some protection (n = 14), or protected (n = 15). Appendix I provides each state's 2015 and 2022 and average IWPR reproductive rights scores by category.

2.4. Statistical analysis

The significance of bivariate associations between rates of unintended and, as a subgroup analysis, unwanted pregnancy, and state reproductive rights category was tested using chi‐square tests. The referent group when evaluating unintended pregnancy included people who were unsure about pregnancy, wanted pregnancy now, and wanted pregnancy sooner. The referent group when evaluating unwanted pregnancy included people with a pregnancy that was wanted later, were unsure about pregnancy, wanted pregnancy now, and wanted pregnancy sooner. Incidence rate ratios, which provide better estimates of relative risk than odds ratios for the rate of unintended pregnancy and the smaller respondent subgroup of those reporting unwanted pregnancy, were estimated from multivariable Poisson regression with robust variance [18, 19]. These models tested the significance of differences between the three IWPR state score categories after controlling for individual respondents’ age, race and ethnicity (social construct potentially representing exposure to racism), marriage status, household income, education attainment, parity, multifetal gestation, and pre‐pregnancy and delivery insurance status.

Because the percentage of unintended pregnancy was expected to be higher among low‐income respondents whose deliveries were covered by Medicaid, we performed sensitivity analyses restricted to respondents with Medicaid delivery coverage. All analyses were population‐weighted using the complex survey module of Stata Version 17 (College Station, TX).

3. RESULTS

After excluding 5480 (3.4%) of respondents with missing data for pregnancy intention or covariates, there were 154,062 PRAMS respondents with live births between 2016 and 2019. The study population is representative of an estimated 7,826,006 individuals with live births in 43 states and DC. Of these respondents, 25.9% lived in 15 states with restrictive reproductive rights, 36.5% lived in 14 states with some protection for reproductive rights, and 38.5% lived in 15 states with protected reproductive rights (Appendix I). Pregnancy was reported as unintended by 24.6% of respondents and unwanted by 6.6%. State mean unwanted and unintended pregnancy rates were highly correlated, r = 0.92 p < 0.001.

Figure 1 demonstrates the correlation between 2015 and 2022 mean IWPR state reproductive rights rankings and states’ mean percent unintended pregnancy (Panel A) and for the sub‐group reporting unwanted pregnancy (Panel B). There was a significant inverse correlation between state reproductive rights ranking and state percent unintended pregnancy (r = −0.53, p < 0.001) and unwanted pregnancy (r = ‐0.56, p < 0.001). States with the highest reproductive rights rankings and lowest unintended pregnancy rates included New Jersey, Oregon, Connecticut, Maryland, and Vermont.

FIGURE 1.

States’ reproductive rights and unintended pregnancy. Panel A demonstrates the prevalence of unintended pregnancies and panel B demonstrates the prevalence of unwanted pregnancies. Reproductive rights score based on Institute for Women's Policy Research reproductive rights rankings. Percent unintended and unwanted pregnancy from Pregnancy Risk Assessment Monitoring System (PRAMS) 2016–2019. Average state reproductive rights scores from the Institute for Women's Policy Research were used in the analysis. Higher rankings represent state environments more protective of reproductive rights. Unintended pregnancy refers to a pregnancy that was wanted later or unwanted. Unwanted pregnancy refers to a pregnancy not desired now or any time in the future. PRAMS data include 43 states and the District of Columbia. Arizona, California, Idaho, Nevada, Ohio, South Carolina, and Texas are not included in PRAMS dataset.

Unintended and unwanted pregnancies were highest in the most restrictive states and decreased as state reproductive rights increased (Table 1). Lower educational attainment and household income were associated with higher rates of unintended pregnancy. Unintended pregnancy was more frequent among individuals at both extremes of age. Higher rates of unintended pregnancy were reported by non‐Hispanic Black (39.2%) and Hispanic (30.7%) respondents, as well as those who were uninsured pre‐conception (32.4%) or insured by Medicaid at the time of delivery (41.7%). Unwanted pregnancies were highest among people 40 and older (12.8%), people with yearly household incomes less than $20,000 (9.8%) people identifying as non‐Hispanic Black (11.7%), people who did not graduate high school (8.5%), and people insured by Medicaid at the time of delivery (9.4%). Sociodemographic characteristics for all categories of pregnancy intent can be seen in Appendix II.

TABLE 1.

Pregnancy intent by respondents’ sociodemographic and clinical characteristics.

	All respondents, percent	Unintended pregnancy a , b (wanted later or unwanted), percent	Unwanted pregnancy a , c , percent
Population‐weighted N	7,826,006	1,984,536	488,815
All Respondents	154,062	39,771	10,286
State Reproductive rights category d
Restricted	24.94	28.42	7.41
Some protection	36.54	25.88	6.32
Protected	38.52	22.88	5.43
Survey year
2016	20.34	24.43	5.78
2017	24.07	25.39	5.96
2018	27.10	25.68	6.45
2019	28.49	25.66	6.54
Married	62.09	17.13	4.18
Age group
18–19	4.17	49.18	6.71
20–24	18.46	38.42	6.08
25–29	29.18	25.83	5.61
30–34	29.68	18.55	5.31
35–39	15.16	17.16	7.93
≥ 40	3.35	17.04	12.80
Race and ethnicity
Non‐Hispanic White	58.59	20.70	4.82
Non‐Hispanic Black	15.26	39.24	11.65
Hispanic	16.32	30.68	6.71
Asian	5.41	18.57	3.96
Other	4.41	27.86	7.53
Household income
≤ $20,000	24.19	38.56	9.79
$20,001–$40,000	21.63	31.62	7.87
$40,001–$60,000	13.64	24.14	5.64
$60,001–$85,000	11.25	18.65	4.57
≥ $85,000	29.29	12.98	3.05
Educational attainment
Did not graduate high school	11.73	31.47	8.48
Graduated high school	24.44	32.89	8.35
Some college	26.97	29.89	7.93
Graduated college	36.86	15.10	2.91
Parity
0	38.62	24.72	3.09
1	33.19	21.78	4.12
2	16.44	28.74	10.52
≥ 3	11.75	32.98	16.63
Multiple gestation	1.75	22.18	7.27
Uninsured pre‐conception	12.80	32.42	8.07
Insured by Medicaid for delivery	41.71	25.73	9.41
Survey completed in Spanish	7.37	27.71	6.07

Note: All comparisons significant at p < 0.01.

^a Percent unintended and unwanted pregnancy and sociodemographic factors from Pregnancy Risk Assessment Monitoring System 2016–2019 data. Respondents include 154,062 respondents with live, singleton births living in 43 states and the District of Columbia. Total weighted populations = 7,826,006. PRAMS data include 43 states and the District of Columbia. Arizona, California, Idaho, Nevada, Ohio, South Carolina, and Texas are not included in PRAMS dataset.

^b Unintended pregnancy refers to a pregnancy that was wanted later or unwanted.

^c Unwanted pregnancy refers to a pregnancy not desired now or any time in the future.

^d Reproductive rights score based on Institute for Women's Policy Research reproductive rights rankings.

Table 2 displays regression results for the likelihood of unintended and unwanted pregnancies by state reproductive rights category. As compared to respondents living in states with protected reproductive rights, the likelihood of unintended pregnancy was equally higher in states with some protection (aIRR 1.06, 95% CI 1.03–1.10) and in states with restricted rights (aIRR 1.07, 95% CI 1.04–1.10). Differences were very similar when restricted to unwanted pregnancy. Specifically, compared with states with protected reproductive rights, the likelihood of unwanted pregnancy was higher in states with some reproductive rights protection (aIRR 1.10, 95% CI 1.02–1.19) and in states with restricted reproductive rights (aIRR 1.15, CI 1.07–1.23).

TABLE 2.

Poisson regression ^a results for unintended and unwanted births.

	IRR (95% CI) unintended b , c	IRR (95% CI) unwanted b , d
	All respondents
State reproductive rights category e
Protected	Reference	Reference
Some protection	1.06 (1.03–1.10)	1.10 (1.02–1.19)
Restricted	1.07 (1.04–1.10)	1.15 (1.07–1.24)
	Respondents covered by Medicaid
Protected	Reference	Reference
Some protection	1.09 (1.04–1.14)	1.20 (1.10–1.32)
Restricted	1.13 (1.08–1.17)	1.15 (1.04–1.27)

^a Regression covariates included year of PRAMS survey, age group, race and ethnicity, Spanish as preferred language, marital status, household income, educational attainment, parity, multiple gestation, uninsured pre‐conception, insured by Medicaid at the time of delivery.

^b Percent unintended and unwanted pregnancy and sociodemographic factors from Pregnancy Risk Assessment Monitoring System 2016–2019 data. All respondents include 154,062 respondents with live, singleton births living in 43 states and the District of Columbia. Total weighted populations = 7,826,006. Medicaid‐only population included 69,515 respondents representing a total weighted population = 3,264,191. PRAMS data include 43 states and the District of Columbia. Arizona, California, Idaho, Nevada, Ohio, South Carolina, and Texas are not included in PRAMS dataset.

^c Unintended pregnancy refers to a pregnancy that was wanted later or unwanted.

^d Unwanted pregnancy refers to a pregnancy not desired now or any time in the future.

^e Reproductive rights score based on Institute for Women's Policy Research reproductive rights rankings.

State differences were larger among respondents with deliveries covered by Medicaid (Table 2). In comparison to states with protected reproductive rights, the likelihood of unintended pregnancy was marginally greater in states with more restriction of reproductive rights; the IRR for unintended pregnancy in states with some protection was 1.09 (95% CI 1.04–1.14) and for states with restricted reproductive rights was 1.12 (95% CI 1.08–1.17). For unwanted pregnancy, the likelihood in states with restricted rights was greater than those with some protected rights (aIRR 1.20, 95% CI 1.10–1.32) and (aIRR 1.15, 95% CI 1.04–1.27, respectively). The full regression model can be found in Appendix III.

4. DISCUSSION

This study found that residents of states with protected reproductive rights legislative environments had a lower risk adjusted likelihood of unintended or unwanted pregnancy when compared to residents of states with some protection or those with more stringent restrictions. When analysis was limited to people who were insured by Medicaid at the time of delivery, state differences were larger.

Previous research on state‐level policies related to reproductive rights and perinatal health has found that imposing restrictive reproductive health legislation, such as targeted regulation of abortion providers and abortion waiting periods, was (as intended) associated with an increase in barriers and delays for abortion care [20, 21]. Other studies of composite reproductive rights laws have found that states with more restrictive environments had higher rates of pregnancy‐related mortality and preterm birth [8, 22]. Few studies have examined the extent to which state legislative environment is associated with unwanted and unintended pregnancy after controlling for individual‐level sociodemographic factors. Existing research often focuses on the individual determinants of unintended pregnancy, such as younger age, less education, minoritized race and ethnicity, and low income [11, 23, 24, 25, 26]. Our work supports the significance of state‐level policies as possible structural determinants.

We found a stronger association between state reproductive rights and unintended pregnancy when evaluating lower income respondents with Medicaid as the payor for birth. It has been shown that health policies that restrict contraception, abortion, and access to healthcare disproportionally affect people with low income [27]. Conversely, state health policies that have expanded access to contraception or Medicaid have been associated with fewer unintended pregnancies [28]. Our results are consistent with the idea that restrictive reproductive rights policies reduce access to sex education, birth control, and abortion in ways that disproportionally affect populations with lower incomes.

The rate of unintended pregnancy among respondents with live births that we report (< 25%) is lower than the often‐cited CDC national average of 45% [11]. Notably, our data only included unintended births, whereas pregnancies that ended in miscarriage or abortion were not included. The most recent unintended pregnancy data from the CDC are from 2011 [11]. Over the course of a decade, the rate of unintended pregnancy could have further declined to a rate closer to that we have reported with ongoing decreases in the teen birth rate and increasing use of long‐acting reversible contraception [29, 30].

Reducing the incidence of unintended pregnancy is important because of its association with adverse maternal and neonatal outcomes [13]. Unintended pregnancies are concentrated among a low socioeconomic status, minoritized population which already experiences disproportionate risk of adverse pregnancy outcomes and maternal morbidity and mortality [11, 31, 32]. Better understanding of the structural factors associated with unintended pregnancy is critical to improvement in health, social and economic outcomes, and health equity.

We find an association between composite state reproductive rights scores and the likelihood of unintended and unwanted pregnancies. In a recent study by Krieger et al. of state health outcomes from 2012 to 2024, state legislative voting records, trifecta state governments, a state policy liberalism index, and voter political lean were significantly associated with a variety of life course health outcomes including infant mortality, premature mortality, maternity care deserts, and food insecurity [33]. Thus, composite state reproductive rights rankings may reflect the larger ideological and cultural environment of the states. PRAMS data provide the additional benefit of being able to test the ecological effects of state residence after controlling for detailed individual risk factors associated with unplanned pregnancy.  All components of the reproductive rights composite score represent reasonable targets for political change and for future study about the role of specific policies on health outcomes. Medicaid expansion after implementation of the Affordable Care Act has been associated with an increase in use of both short‐acting and long‐acting reversible contraceptives [34]. Only 10 states have not adopted Medicaid expansion, and states in this study that have not expanded Medicaid are grouped as having either restrictive (WY, KS, TN, MS, and AL) or somewhat restrictive (WI, GA, FL) reproductive rights [35]. Medicaid Family Planning Waivers, which provide access to family planning services to people who do not qualify for Medicaid, have been associated with use of family planning and contraceptive services and a decrease in unintended pregnancy [36]. However, only 26 states currently have expanded Medicaid eligibility for family planning services despite its potential to reduce unintended pregnancy [37]. States could additionally work to expand sexual education programs. Increased federal funding for comprehensive sex education has been associated with a decrease in teen birth rates, which are disproportionately unintended [38].

This cross‐sectional study cannot provide a strong causal inference. There may be unmeasured state and regional population differences that are driving our results. While this PRAMS dataset is strengthened by a population that included respondents from 43 states, Texas and California, which have large populations and polarized political contexts, did not participate in PRAMS, which may also affect the ability to generalize these results. PRAMS also does not include people who gave birth but were less than 18 years old, a group with a high unintended pregnancy rate [11]. In this study, we included the substantial proportion (15.5%, or an estimated 1.2 million residents of the PRAMS states) of respondents who answered “unsure” as to the wantedness of their pregnancy in the referent group, rather than define them as having an unintended pregnancy. Therefore, our results may represent a conservative estimate of unintended and unwanted pregnancy rates. Our study uses retrospective data from 2 to 6 months postpartum, therefore, it is possible that recall bias after delivering a neonate may influence their answering the PRAMS question about wantedness. This study only included people with a live birth. Therefore, results may be influenced by selection bias as people with unintended pregnancies who did not have a live birth (either from a termination or miscarriage) were not included in our study population.

While the reproductive rights landscape has changed in the post‐Dobbs era, we believe that this analysis of the association between state‐level reproductive rights and unintended pregnancy prior to Dobbs v. Jackson is crucial to understanding ongoing legislative trends and how they may be associated with health outcomes. PRAMS is one of the few data sources allowing for cross‐sectional comparison of a majority of states in the United States. PRAMS had not released data from 2022 and 2023 (post‐Dobbs) at the time of this study's completion. By examining pre‐Dobbs data, we hope to demonstrate the importance of ongoing evaluation of state‐level reproductive rights legislation as our data suggest that state‐level reproductive rights even before Dobbs v. Jackson were associated with differential rates of unintended pregnancy.

Unintended pregnancy has previously been defined as a key public health measure because of its historical link to restricted reproductive autonomy and association with adverse birth outcomes [24]. However, arguments have been made against the use of unintended pregnancy as a public health outcome and indicator of reproductive autonomy. A focus on unintended pregnancy may implicitly suggest a woman's failure to contracept, rather than address the structural barrier she may face to access and use contraception [39]. Unintended pregnancy likely does not comprehensively capture the continuum of pregnancy desire and monitoring unintended pregnancy as a way to decrease abortions further stigmatizes abortion as an unsavory outcome when it instead may represent the ability to autonomously make reproductive decisions [39]. We acknowledge these inherent limitations of measuring unintended pregnancy and intend to further demonstrate how structural factors (such as state reproductive health policies) are associated with pregnancy desire when individual‐level factors are controlled.

Our findings reflect an earlier, pre‐Dobbs era. Distinctions between protected and restricted states with the Dobbs decision have only become more partisan, likely eroding the number of states previously rated as having “some protection.” In the 2024 election cycle, access to reproductive rights was discussed as an important ballot issue. As research continues to demonstrate the growing state differences in health and life expectancy [2], we could see even further divergence in reproductive outcomes between states with protective and restrictive reproductive rights. This is likely to only exacerbate existing disparities in unintended pregnancy and adverse pregnancy outcomes—as well as social outcomes as reproductive autonomy is associated with educational attainment, employment, and financial opportunities [40]. Our findings suggest a need to closely monitor the ongoing implications of state‐level restrictive reproductive rights for maternal and child health.

CONFLICT OF INTEREST STATEMENT

The authors report no conflict of interest.

Supporting information

Supporting information