Almost all states in the middle and southern parts of the United States require that before a physician performs an abortion on a minor, he or she must notify or obtain consent from at least one of the minor’s parents (Figure 1).1 In some states—Florida, for example—minors who want to obtain an abortion without parental involvement must travel hundreds of miles to a state that does not have one of these restrictive laws. The implications of this legal environment on the reproductive outcomes of minors are not well understood. Most research on parental involvement laws pertains to an earlier period, when interstate travel by minors to avoid compliance with a law was more common.2 Studies that used data on abortion by state of occurrence often found that implementation of parental involvement laws was associated with substantial decreases in teenage abortion rates.3–6 The few studies that were able to measure abortion by minors’ state of residence generally showed a small, if any, association.7–9 In a series of studies, researchers analyzed a parental involvement law under circumstances consistent with the current distribution of laws: Implementation of the Texas parental notification law in 2000 was associated with a decline in abortion rates, a rise in birthrates and an elevated likelihood of minors’ obtaining an abortion after 12 weeks’ gestation.10–12
Parental involvement laws regarding minors’ access to abortion in the United States, 2009
Source: reference 1.
Although the national map of parental involvement laws is unlikely to change appreciably in the near future, since 2003, four states—Arizona, Arkansas, Texas and Virginia—have converted their parental notification statutes to laws that require parental consent.*1,13 Such changes are motivated by a belief that a notice requirement is easier to circumvent than a consent statute, and that a stricter law will lead to fewer abortions. As Texas State Representative Phil King, who sponsored the change in his state, commented, “I think it will do what [parental notification] intended to do by bringing parents into the decision-making process, and when that happens, we’ll see a reduction in abortion and in teenage pregnancy.”14
Despite this belief by some politicians, only three studies have analyzed whether laws that require parental consent have a greater impact on the behavior of minors than laws that require parental notification, and they report widely disparate estimates and have major methodological weaknesses.6,15,16 Tomal analyzed county-level rates of abortions and births among minors and older teenagers in 11 states in 1995, and found that parental notification laws had a stronger negative association with minors’ abortion rate than did parental consent laws.15 Moreover, both types of law were associated with the abortion rates and birthrates of adolescents aged 18–19, which calls into question the association with minors’ rates. Medoff analyzed minors’ abortion rates—which he defined as abortions per 1,000 pregnancies—from 50 states at three points in time (1982, 1992 and 2000).16 Like Tomal, he found that notification laws were associated with a greater decline in abortions for minors than were parental consent laws. However, it is extremely difficult to distinguish ongoing trends in abortion rates from the influence of changes in the law that may have occurred 8–10 years earlier. Finally, New used a pool time-series cross section of state abortion rates from 1985 to 1999 based on the states that reported abortion to the Centers for Disease Control and Prevention.6 He found that parental consent laws were associated with substantially greater declines in minors’ abortion rate than were parental notification requirements. Yet despite the methodological sophistication of his analysis, the results are questionable. Because New examined abortions by state of occurrence, he could not determine whether the negative association represented a real decline in minors’ abortion rate or simply a change in where abortions were performed. Given the inconsistent findings and questionable designs of these studies, whether parental consent laws have a greater influence than notification requirements on the reproductive outcomes of minors remains an open question.
Cross-state travel by minors to avoid compliance with a law has become more difficult for those residing in the central and southern parts of the United States. The Texas study, for instance, found little evidence that minors left the state to obtain an abortion in the period after the parental notification law took effect, since most of the surrounding states also enforced a parental involvement law.10 Because of minors’ lack of access to abortion services in states that require parental involvement, the only practical option for those who do not want to involve their parents is to seek a court waiver based on a judge’s assessment of their maturity or whether parental involvement would not be in their best interest. Despite the potential importance of the judicial bypass system, no population-based studies have examined the characteristics of minors who obtain an abortion through a court waiver, or whether minors who use this option terminate later in pregnancy than those whose parents are involved in their abortion decision. Existing data, from surveys at selected abortion clinics or at a referral organization for pregnant minors, describe minors who went to court for a waiver, but three of these four surveys were conducted more than 25 years ago.17–20
This study draws on unique data on induced terminations from Arkansas and addresses three questions. First, was the state’s change from a notification statute to a consent statute in 2005 associated with a decrease in the abortion rate and an increase in the rate of second-trimester abortions among minors relative to rates among older teenagers? Analysis of changes in a law within the same state, as opposed to a comparison of the effect of laws in different states, offers a useful way to control for hard-to-measure differences between legal jurisdictions. Second, do minors who obtain an abortion through a court bypass differ in age, race or ethnicity, or other demographic characteristics from minors who obtain parental consent for abortion? Finally, do minors who use a judicial bypass have the abortion later in pregnancy than minors who have parental consent? This study is possible because the Arkansas law requires that all abortion providers report to the state health department whether the minor obtained parental consent or received a judicial waiver of consent for the abortion. This is the first effort to analyze these data.
BACKGROUND AND METHODS
On March 3, 2005, Gov. Mike Huckabee signed into law House Bill 1033, which states that no person may perform an abortion on an unemancipated minor without the written consent of a parent or legal guardian. The note providing parental consent must be notarized if the parent is not present; if the parent is present, a valid photo identification is required. Photocopies of the note and identification must be maintained by the provider for five years. The consent requirement can be waived if a judge determines that the minor is mature or that an abortion without parental consent is in her best interest. Consent can also be waived if the attending physician certifies that a medical emergency exists and there is insufficient time to obtain parental consent. The law amended a parental notification statute that had been in effect since 1989.21
Data and Analysis
The study used de-identified records of abortions performed in Arkansas between 2001 and 2007. These records contain information on age, race and ethnicity, education level, state of residence, previous live births, previous induced abortions and the gestational age of the fetus. Starting in April 2005, the files indicate whether a minor had parental consent or had obtained a waiver of consent through a judicial bypass. The Guttmacher Institute, whose periodic survey of abortion providers is widely considered to have the most complete count of abortions by state, reported that 5,540 abortions were performed in the state in 2000, and 4,710 were performed in 2005.22 The numbers collected by the Arkansas Department of Health and reported to the Centers for Disease Control and Prevention were 5,501 and 4,685, respectively.23,24
Data on 7,463 abortions among women aged 15–19 were analyzed to assess changes in abortion and second-trimester abortion rates from 2001 to 2007. Rates per 1,000 women were examined for the age-groups 15–17 (minors) and 18–19 (older teenagers); second-trimester abortions were defined as those performed after 12 weeks’ gestation. Rates among 17-year-olds and 18-year-olds were also compared in an effort to account for the baseline differences and improve the internal validity of the comparison. Population estimates are from the Surveillance Epidemiology and End Result registry.25
Linear regression analyses were used to estimate the changes in the annual abortion rate and in the rate of -second-trimester abortions among minors and older teenagers associated with the shift from a parental notification to a consent statute. Two specifications were employed for each type of abortion rate: In the first, the dependent variable was the actual rate, and in the second, it was the natural logarithm of the rate. Expressing the dependent variable in logarithms is convenient, since coefficients on the indicator variables measure relative changes in the outcome. In addition, comparisons of regressions using rates with ones using logs show whether results are sensitive to the functional form of the dependent variable.
The regression analysis is based on 14 observations: seven years of data and two age-groups. Each model included an indicator for age (minors vs. older teenagers, or 17- vs. 18-year-olds), an indicator for the years in which the consent requirement was in effect (2005–2007 vs. 2001–2004), an interaction of minor and the consent indicator, and a linear trend term. The coefficient on the interaction term measured the change in the abortion rate among minors relative to that among older adolescents in the period after parental consent was required relative to before the requirement.
Pearson chi-square tests assessed differences between selected characteristics of minors who obtained parental consent and those of minors who obtained the abortion through a judicial bypass. Logistic regressions estimated differences among subgroups of teenagers in the likelihood of having used a judicial bypass and of having had a second-trimester abortion, and linear regressions estimated differences in mean gestation at termination; these models adjusted for teenagers’ background characteristics and the year in which the abortion occurred.
If the change from a parental notification to a parental consent requirement is associated with a decline in abortions, then the abortion rate among minors will drop relative to that among older adolescents following implementation of a new law. Arkansas minors’ abortion rate did decline between 2004 and 2005, but so did the rate among 18–19-year-olds (Figure 2). Moreover, these declines appear to be part of downward trends that began in 2001. In addition, the groups showed a similar pattern in abortion rates for 2006 and 2007.
Abortion and second-trimester abortion rates among teenagers, by age, Arkansas, 2001–2007
Abortion and second-trimester abortion rates among teenagers, by age, Arkansas, 2001–2007
If a consent statute is more burdensome than a notice requirement, and if more minors use the judicial bypass procedure as a result, then the rate of second-trimester abortions might increase after a law changes. However, like the abortion rate, the rate of second-trimester abortions declined between 2004 and 2005 among all age-groups. The rate also rose between 2005 and 2006, but the increase was greater among older adolescents than among minors, contrary to expectations. Hence, the consent statute does not appear to have been associated with a differential change in the incidence or timing of abortion among minors.
Regression estimates largely confirm the foregoing evidence: The change in the law had no statistically robust association with the incidence or timing of abortion among minors (Table 1). The one exception was that the rate of second-trimester abortions rose 0.59 per 1,000 more among minors than among older adolescents in the period after the consent statute was implemented. However, the coefficient in the model using the natural logarithm of the rate, although not statistically significant, suggests a 3% decline in the rate. In the regression models comparing abortion rates between 17- and 18-year-olds, none of the coefficients were statistically significant. The lack of consistent association suggests that the change from a notification to a consent statute had no effect on either the rate or the timing of abortion among minors.
Coefficients from linear regression analyses assessing changes in measures of abortion among teenagers associated with implementation of a parental consent statute, Arkansas, 2001–2007
|Measure||Minors vs. older teenagers†||17- vs. 18-year-olds|
|Rate||1.59 (–2.13 to 5.32)||0.94 (–3.24 to 5.13)|
|Log of rate||0.02 (–0.28 to 0.32)||0.02 (–0.37 to 0.41)|
|Rate||0.59 (0.03–1.16)*||0.02 (–0.76 to 0.80)|
|Log of rate||–0.03 (–0.46 to 0.41)||–0.11 (–0.70 to 0.47)|
Judicial Bypass and Its Effects
Ten percent of the 972 abortions among minors in Arkansas were obtained through the judicial bypass process (Table 2). Among minors who used this option, 55% were 17-year-olds, 34% were 16-year-olds and the remainder were 15 or younger. The age distribution among those who obtained parental consent was substantially different—42% were 17 years old, 30% were 16 and 28% were 15 or younger. No differences were evident by gestational age, race or ethnicity, or schooling between those who obtained parental consent and those who used the bypass.
Percentage distribution of minors who obtained an abortion after implementation of the parental consent statute, by selected characteristics, according to whether they obtained parental consent or used judicial bypass
|Characteristic||Total (N=972)||Parental consent (N=876)||Judicial bypass (N=96)|
|Gestational age (wks.)|
|Behind in school†|
Residency status was strongly associated with use of the judicial bypass. Eighteen percent of all abortions among minors in Arkansas during the period in which the consent law was in effect were among nonresidents; while 17% of minors who obtained parental consent were nonresidents, 25% of minors who used the judicial bypass lived outside the state. Further examination revealed that 97% of the nonresidents in the sample were from five states: Mississippi, Missouri, Oklahoma, Tennessee and Texas (not shown). Each of these states had a parental consent requirement. Why minors would leave their own state and travel to Arkansas to obtain an abortion, given its consent statute, is at first unclear. The multivariate analysis explores this issue further.
Regression analysis found that Hispanic minors were more likely to use the bypass procedure to obtain an abortion than were whites (odds ratio, 3.7—Table 3). Age continued to be an important correlate, as minors aged 15 or younger were less likely to use a bypass than were 17-year-olds (0.2); no differences were found between 16- and 17-year-olds. Nonresidents of Arkansas had an elevated likelihood of relying on the bypass option (1.9). Finally, minors who had an abortion in 2006 were more likely than those who had one in 2005 to have obtained a bypass (2.5).
Odds ratios from logistic regression analyses assessing predictors of minors’ use of judicial bypass and second-trimester abortion, and coefficients from linear regressions assessing differences in mean gestational age at termination, by selected characteristics
|Characteristic||Judicial bypass (N=972)||Second-trimester abortion (N=970)||Difference in mean gestational age (wks.) (N=970)|
|Black||0.76 (0.46–1.25)||2.79 (1.95–3.93)*||2.10 (1.56–2.64)*|
|Hispanic||3.74 (1.52–9.23)*||3.29 (1.35–8.01)*||2.46 (1.01–3.92)*|
|Other/unknown||1.83 (0.65–5.15)||1.72 (0.69–4.30)||1.47 (0.09–2.86)|
|≤15||0.24 (0.12–0.49)*||0.93 (0.61–1.43)||0.16 (–0.48 to 0.80)|
|16||0.86 (0.53–1.38)||0.88 (0.58–1.34)||–0.24 (–0.84 to 0.36)|
|Behind in school†||0.52 (0.21–1.23)||1.17 (0.96–1.42)||0.30 (–0.05 to 0.67)|
|Had previous birth||0.40 (0.12–1.37)||0.55 (0.24–1.26)||–0.94 (–2.12 to 0.24)|
|Had previous abortion||0.50 (0.15–1.66)||0.68 (0.30–1.53)||–0.30 (–1.39 to 0.78)|
|Out-of-state resident||1.90 (1.13–3.18)*||5.31 (3.63–7.77)*||3.05 (2.40–3.71)*|
|2006||2.53 (1.31–4.89)*||0.92 (0.60–1.40)||–0.09 (–0.74 to 0.56)|
|2007||1.78 (0.89–3.58)||0.74 (0.47–1.17)||–0.24 (–0.90 to 0.43)|
|Used judicial bypass||na||0.45 (0.23–0.89)*||–1.11 (–1.97 to –0.25)*|
Minors who obtained a judicial bypass had a decreased likelihood of having had a second-trimester abortion (odds ratio, 0.5), and they terminated the pregnancy 1.1 weeks earlier, on average, than did minors who had obtained parental consent. The latter finding was unexpected, since the bypass process adds a potentially time-consuming step in the pathway to obtaining an abortion. Another noteworthy finding was that blacks and Hispanics were more likely than whites to have had a second-trimester abortion (2.8 and 3.3, respectively); moreover, blacks and Hispanics terminated their pregnancies more than two weeks later, on average, than did whites (2.1–2.5 weeks). Finally, nonresident minors had an elevated likelihood of having had an abortion in the second trimester (5.3), and on average, they had their abortions more than three weeks later than did residents. This gestational age difference is substantial, and suggests that nonresidents travel to Arkansas because late-term abortion providers may be less accessible in their home state.
This study found no decrease in the abortion rate among minors relative to that among older adolescents—or in the rate among 17-year-olds relative to that among 18-year-olds—after Arkansas changed its parental notification statute to a parental consent statute in 2005. Overall, changes in the rates of second-trimester abortions between these age-groups were also nonsignificant. The results suggest that it is the requirement of parental involvement, and not whether the policy is a notification or consent statute, that may alter minors’ reproductive outcomes. The findings are not surprising, considering that a change from a notification to a consent statute would have to induce an appreciable change in parental behavior, or in a minor’s expectation of her parents’ behavior, to lead to an additional decrease in abortions. For instance, parents who would not prevent their daughter from getting an abortion under a notification statute are unlikely to do so under a consent law. Indeed, surveys of minors indicate that relatively few fear that their parents would force them to carry an unwanted pregnancy to term.17,20,26 The most common response that minors offer for why they did not involve their parents is the desire not to upset or disappoint them. The distinction between a consent and a notification statute would appear to be secondary to these concerns.
The lack of any substantive causal difference between a consent and a notification statute on minors’ abortion rates is unlikely to alter the drive among some state politicians to convert laws requiring parental notification to laws requiring parental consent. Advocates of these laws contend that notification requirements deny parents their rightful authority over their children’s behavior. As Texas State Representative Phil King stated in the debate over his state’s change from a notification to a consent statute, “Forget how you feel about abortion, this really isn’t about abortion. It’s about families and parents being involved and having authority to make decisions about medical procedures given to their kids.”14
Data from the current study indicate that following implementation of the consent law in 2005, 10% of all abortions among minors in Arkansas were obtained through the judicial bypass process. In a 1984 survey of minors who obtained abortions in Minnesota, 43% had used the state’s waiver procedure.19 However, the sample may not have been representative of all minors, since it was obtained from a single (albeit major) clinic. Data from a survey of minors who sought a judicial waiver in Massachusetts from May 1998 to April 1999,20 combined with statewide abortion data for 2000,27 suggest that if all bypass petitions were approved, 31% of abortions among minors in the state were obtained via bypass.† Even if some minors were denied their petition, the proportion is substantially greater than that found in Arkansas. The more conservative political climate in Arkansas and its more rural population may explain some of the difference. For instance, a 1997 survey of county courts charged with processing bypass applications in Alabama and Tennessee found that clerks in 22 of Alabama’s 67 counties and in 37 of Tennessee’s 95 counties were unaware of the procedure.28 The researcher argued that minors could be discouraged from pursuing the judicial bypass in the face of uninformed officials. Both states are similar to Arkansas in terms of region, demographic characteristics and political sentiment.
Nonetheless, the 10% of minors’ abortions in Arkansas that were obtained through the bypass option represent a sizable proportion of minors who are unwilling to involve their parents. If national data pertain to Arkansas, then approximately 40% of minors who had an abortion would not have informed their parents in the absence of the consent law.26 Thus, about one-fourth of minors in Arkansas who would not have informed a parent if they had a choice obtained a court waiver. Furthermore, if some bypass requests are not granted, then the proportion is even higher.
The importance of the bypass procedure has likely increased with the distance that minors must travel to access abortion services in states without a parental involvement law. All of the states that surround Arkansas required parental involvement from 2005 to 2007. The nearest states without these laws were Illinois and New Mexico.1 Indeed, most minors in the South and Midwest would have had to travel long distances to access abortion services without parental involvement, and this underscores the need for bypass procedures that are clear, expedient and confidential. Results from the survey of court clerks in Alabama and Tennessee suggest that minors are further burdened by lax or even negligent implementation of the law.28 A potentially significant initiative would be to ensure that personnel in all county courts have accurate information about a minor’s rights and are able to expeditiously process her request for a court appearance.
This is the first study to use population-based data on abortions obtained through a judicial bypass procedure to characterize minors who used the court system to avoid parental consent, and the first to test whether the process was associated with an increase in abortions performed in the second trimester. Not surprisingly, the finding that older minors are much more likely to use the bypass option than are minors aged 15 or younger is consistent with previous studies of bypass procedures.17,19,20 It is also consistent with surveys showing that in states without parental involvement laws, older minors are less likely than younger ones to include parents in their decision.26
The present study found no difference between whites’ and blacks’ likelihood of using a judicial bypass; however, Hispanics were more likely than whites to obtain a court waiver of parental consent. In addition, Hispanic minors were likely to terminate later in pregnancy than were whites, but this association was based on a small number of Hispanics in the sample. Nevertheless, if the prohibition against abortion is greater in the Hispanic than in the non-Hispanic community (as suggested by the proportion of pregnancies that are terminated), then the more frequent use of the bypass procedure by Hispanics would be consistent with the survey findings that the most important reasons why minors do not involve their parents is the fear of disappointing or upsetting them.17,20,26
Methodologically, the results illustrate the importance of having an appropriate comparison group when evaluating the effects of parental involvement laws. Adolescents aged 18–19, for instance, are a questionable comparison group for minors because of large differences in sexual behavior. For example, the national abortion rate among older adolescents is about 2.5 times that among minors.27 In Arkansas, minors’ second-trimester abortion rate rose relative to older teenagers’ after the parental consent statute was implemented, but the association did not hold when the dependent variable was expressed in logs. The finding that the result was sensitive to the functional form may reflect the large difference between minors’ and older adolescents’ rates of second-trimester abortions in the years before the consent statute. It thus calls into question the internal validity of a design based on disparate levels of the outcome in the baseline period between the treatment and comparison groups,29 which is why results were also presented for 17- and 18-year-olds. Both the abortion rate and the rate of second-trimester abortions among 17- and 18-year-olds were more similar in magnitude in the years before consent than were those of minors and of older teenagers. Given the study’s strong design, it is also noteworthy that the consent statute was not associated either with the abortion rate or with the rate of second--trimester abortions among 17-year-olds relative to 18-year-olds, regardless of whether the dependent variable was measured in levels or in logs.
Results do not appear to be biased because of minors’ leaving the state to obtain an abortion. First, as noted earlier, Arkansas is surrounded by states that enforce a parental consent law. In the central and southern parts of the United States, access to abortion without parental involvement requires extensive travel that few minors would be willing to undertake.10,11 Second, if minors from Arkansas had left the state, the abortion rate of 17-year-olds should have declined relative to that of 18-year-olds, but no such decline was observed.
Finally, minors who used the bypass procedure terminated their pregnancy earlier than did minors who obtained parental consent, even after age and race were controlled for. This may reflect selection based on minors’ maturity. The data do not indicate how many minors attempted the bypass process. Minors who obtained a waiver may have been more composed and had more support, and may have acted upon their unintended pregnancy more quickly, than their counterparts who had to obtain parental consent.
The study has a number of limitations. First, the data are from a single state with fewer than three million residents, so findings cannot be generalized to more urban and populous states. Second, there was no information on minors who sought a bypass but were denied. Such data may have explained why minors who obtained an abortion using the bypass had the procedure earlier in pregnancy than minors who obtained parental consent. Finally, information on Arkansas minors who terminated their pregnancy in another state was unavailable. In light of the surprising number of nonresident minors who traveled to Arkansas to obtain an abortion, it would be useful to better understand their motivations for seeking the procedure in another state.
The bypass option is now used in 34 states.30 Given its importance, there is a need for further study of the characteristics of minors who seek a bypass, their experience with the process and the proportion who are granted the waiver.