|Home | About | Journals | Submit | Contact Us | Français|
The incidence of unintended pregnancy is among the most essential health status indicators in the field of reproductive health. One ongoing goal of the U.S. Department of Health and Human Services is to reduce unintended pregnancy, but the national rate has not been estimated since 2001.
We combined data on women’s pregnancy intentions from the 2006—08 and 2002 National Survey of Family Growth (NSFG) with a 2008 national survey of abortion patients, data on births from the National Center for Health Statistics, induced abortions from a national abortion provider census, miscarriages estimated from the NSFG, and population data from the U.S. Census Bureau.
Nearly half (49%) of pregnancies were unintended in 2006, up slightly from 2001 (48%). The unintended pregnancy rate increased to 52 per 1,000 women aged 15—44 years old in 2006, from 50 in 2001. Disparities in unintended pregnancy rates among subgroups persisted, and women who were 18—24 years old, poor, or cohabiting had rates two to three times that of the national rate. The unintended pregnancy rate declined notably for teens 15—17 years old. The proportion of unintended pregnancies ending in abortion decreased from 47% in 2001 to 43% in 2006, and the unintended birth rate increased from 23 to 25 per 1,000 women 15—44 years old.
Since 2001, the U.S. has not made progress reducing unintended pregnancy. Rates increased for nearly all groups and remain high overall. Efforts to help women and couples plan their pregnancies, such as increasing access to effective contraceptives, should target groups at greatest risk of unintended pregnancy, particularly poor and cohabiting women.
Preventing unintended pregnancy is a personal goal for most couples, and reducing the national level of unintended pregnancy is one of the most important reproductive health goals identified by the United States Department of Health and Human Services . Women who have an unintended pregnancy are also at risk for unintended childbearing, which is associated with a number of adverse maternal behaviors and child health outcomes, including inadequate or delayed initiation of prenatal care, smoking and drinking during pregnancy, premature birth, and lack of breastfeeding, as well as negative physical and mental health effects on children [2–9].
While the unintended pregnancy rate in the United States decreased between the late 1980s and mid-1990s , it stalled by 2001, the last year for which estimates are available . Recent decreases in births and abortions have occurred among some population subgroups (e.g., teens) , but it is unclear if unintended pregnancy rates have also changed. The recent release of new data on pregnancy intentions has made it possible to determine the incidence of unintended pregnancy for 2006. We calculated unintended pregnancy rates for all women of reproductive age and for key population subgroups, including race and ethnicity and relationship status, because previous studies indicate strong associations among unintended pregnancy and these groups . We also present information on outcomes of unintended pregnancy, including the percentage of unintended pregnancies ended in abortion and the rate of births that followed unintended pregnancy. These estimates are some of the most essential indicators in the field of reproductive health, and periodic trend assessments provide valuable information for public health officials and policy makers who monitor progress towards reducing unintended pregnancy.
For all U.S. women and by key population subgroups (age, educational attainment, race and ethnicity, income, relationship status, parity and religious affiliation), we determined the number of pregnancies that ended in birth, induced abortion and miscarriagea; calculated the proportion of each of these outcomes that were unintended; and then divided the total number of unintended pregnancies by the population of women aged 15—44 years old to obtain an unintended pregnancy rate per 1,000 women.
We relied on data from the National Center for Health Statistics (NCHS) [13–15] to obtain the number of U.S. births that occurred in 2001 and 2006 overall and by the mother’s age, educational attainment, race and ethnicity, relationship status (not including cohabitation), and parity (2006 only). We distributed births by other subgroups (including cohabiting status) using the National Survey of Family Growth (NSFG), a nationally representative survey of U.S. women aged 15—44 years old conducted by the NCHS.
Women’s pregnancy intentions were obtained from the NSFG, which asked women a series of retrospective questions to determine whether each of the pregnancies she had had was intended or unintended at the time it occurred. Intended pregnancies were those that occurred to women who wanted a baby at the time they became pregnant or sooner, or were indifferent about conceiving; unintended pregnancies were conceptions that were mistimed (i.e., the woman wanted to become pregnant at some point in the future, but not when she conceived) or unwanted (i.e., she did not want to become pregnant at the time of conception nor in the future). We focused on the births in the five years preceding the 2006—08 (n=2,044) and 2002 (n=2,618) interviews.
The total number of surgical and medication abortions performed in 2001 and 2006 came from a census of U.S. abortion providers  conducted by the Guttmacher Institute. Counts by age came from the Centers for Disease Control and Prevention’s 2001 and 2006 abortion surveillance reports [17,18], and estimates for all other subgroups were based on interpolations of distributions from two nationally representative Abortion Patient Surveys (APS) conducted by the Guttmacher Institute in 2000 (n=10,683)  and 2008 (n=9,493) .
Abortions are underreported in the NSFG. Therefore, pregnancy intentions among women obtaining abortions for both 2006 and 2001 were based on distributions from the 2008 APS, which, for the first time, asked women the same set of questions that were used in the NSFG. Use of these data enabled us to identify the proportion of abortions that followed intended pregnancies, rather than assuming that all abortions followed unintended pregnancies, an approach used in previous analyses.b
There is no “gold standard” count of miscarriages. Official statistics are limited to fetal deaths at 20 weeks of gestation or later , so miss those that occur earlier in pregnancy. We estimated the number of miscarriages for 2006 by calculating the ratio of miscarriages to births  overall and by subgroup that occurred in the seven years preceding the last two NSFG rounds (2002 and 2006—08) and multiplying that ratio by the total number of U.S. births in 2006 overall and by subgroup. Women in their teens and those 40 or older had relatively fewer pregnancies, so we increased the sample size by including data from a third round of the NSFG (1995) to improve the validity of the estimate.c To estimate the number of miscarriages for 2001, we applied the same ratio calculated from all three NSFG surveys combined to the 2001 birth counts.
Information on the intendedness of pregnancies ending in miscarriage came from miscarriages in the five years preceding the 2006—08 (n=560) and 2002 (n=729) NSFG interviews. In previous analyses, we relied directly on women’s reports of intendedness, but subgroup sample sizes for 2006 were inadequate. Because miscarriages are pregnancies that would otherwise end in either birth or abortion, we would expect that the proportion of miscarriages that were intended would fall between the proportion of births that were intended and the proportion of abortions that were intended. For the entire NSFG sample, this assumption was accurate.d Therefore, for subgroups, we calculated the proportion of miscarriages that were intended by constraining it to fall between the proportion of births and abortions intended.e
Denominators for pregnancy, birth and abortion rates for all women aged 15—44 years old and by age and race and ethnicity were obtained from population estimates published by the U.S. Census Bureau . Population distributions by educational attainment, poverty and relationship status came from the Annual Social and Economic Supplements of the Current Population Survey (CPS). The population distributions for women by cohabitation status, religious affiliation and parity were based on interpolations of the 1995, 2002 and 2006--08 NSFGs. Distributions by education were limited to the population of women 20 and older, who were likely to have completed or mostly completed schooling.
When calculating the percentage of unintended pregnancies that ended in abortion, we excluded miscarriages from the denominator in order to better represent pregnancies with outcomes decided by the woman.
There were 6.7 million pregnancies in the United States in 2006 (Table 1), up from 6.4 million in 2001 (data not shown). Some 3.2 million pregnancies were unintended in 2006, compared with 3.1 million in 2001 (data not shown). The percentage of pregnancies that were unintended increased slightly between 2001 (48%) and 2006 (49%), and the unintended pregnancy rate also increased during this time period: In 2006, there were 52 unintended pregnancies for every 1,000 women aged 15—44 years, compared with 50 in 2001. When looking at unintended pregnancy by timing, 29% of all pregnancies were mistimed and 19% were unwanted (data not shown). In other words, about 5% of women of reproductive age had an unintended pregnancy in 2006. When looking at unintended pregnancy by timing, 29% of all pregnancies were mistimed and 19% were unwanted (data not shown). The intended pregnancy rate stayed nearly the same, and the overall pregnancy rate increased.
The proportion of pregnancies that were unintended generally decreased with age, with more than four out of five pregnancies unintended among women 19 years and younger. Between 2001 and 2006, this percentage decreased for women aged 15—17 years and increased or stayed nearly the same for all other women. The unintended pregnancy rate was the highest for women 20—24 years, due to an increase between 2001 and 2006.
Women with the fewest years of education had the highest unintended pregnancy rate, and rates decreased as years of education attained increased. Unintended pregnancy rates increased the most among women with no college experience.
Black women had the highest unintended pregnancy rate among all racial and ethnic subgroups, more than double that of non-Hispanic white women. Rates changed little between 2001 and 2006.
Poor and low-income women’s unintended pregnancy rates increased substantially, while the rate for higher-income women decreased. The rate for poor women was more than five times the rate for women in the highest income level. While there was little difference by education among women in the highest income bracket (Fig. 1A), minorities had the highest unintended pregnancy rates regardless of income level (Fig. 1B).
Unintended pregnancy rates increased among cohabitors and formerly-married women. Cohabiting women exhibited both the highest rate and the greatest increase among all individual subgroups measured in this analysis. Rates were even higher among cohabiting women who were under 25 years old (Fig. 2A) or poor or low-income (Fig. 2B).
Women with one previous birth had an unintended pregnancy rate that was roughly twice as high as the rate for women who have never given birth and women with two or more births.
Women with no affiliation reported the highest unintended pregnancy rate, followed by Catholics, Protestants, and women with other affiliations.
Forty-three percent of unintended pregnancies ended in abortionf in 2006, a decline from 47% in 2001 (Table 2). In 2006, the unintended birth rateg was 25 per 1,000 women aged 15–44 years, up from 23 in 2001.
Between 2001 and 2006, the proportion of unintended pregnancies ending in abortion increased for women aged 15—17 years and declined or stayed the same for all other women. The greatest declines were exhibited among women 18—24 years. As a result, the unintended birth rate decreased for women 15—17 years, and increased the most for women aged 18—24 years. Rates for women 18—24 years were more than twice the national rate.
Women with some college but no degree were most likely to end an unintended pregnancy by abortion; these women were also more likely to still be enrolled in school. Those without a high school diploma were most likely to continue an unintended pregnancy, and had an unintended birth rate that was almost twice the national rate and nearly four times the rate for college graduates.
The proportion of unintended pregnancies ending in abortion decreased across all racial and ethnic subgroups, with black women most likely to end an unintended pregnancy by abortion. Hispanic women had the highest unintended birth rate, and minority women had rates that were more than twice that of white women.
Compared with higher-income women, poor and low-income women were less likely to end an unintended pregnancy by abortion. Consequently, poor women had a relatively high unintended birth rate. While lower-income women experienced an increase in the unintended birth rate, this rate remained relatively stable for women in the highest income category.
Married and cohabiting women were much less likely than other women to end an unintended pregnancy by abortion. The rate of unintended births among cohabiting women increased sharply, and was more than three times the rate for other women.
Women with exactly one previous birth were least likely to end an unintended pregnancy by abortion, and their unintended birth rate was more than twice that of the other groups’ rates.
Women with no affiliation were most likely to end an unintended pregnancy by abortion; they also had the highest unintended birth rate, followed closely by Catholics and Protestants. Evangelicals were least likely to terminate an unintended pregnancy.
The U.S. unintended pregnancy rate increased slightly between 2001 and 2006, a worrisome trend, and remains significantly higher than the rate in many other developed countries . Population shifts—e.g., increases in groups with high rates, such as poor and minority women—may have contributed to the overall increase. In addition, the overall increase could have occurred if the trend toward later childbearing  has led to a longer period before childbearing when relatively less-effective methods are used , and a shorter period post-childbearing when use of highly effective long-term methods is more common.
During the same period, the overall proportion of women ending an unintended pregnancy by abortion decreased. These changes may have been due to decreased access to abortion in some areas, increased stigmatization of abortion, or both.
Among all the subgroups for which we present data, only women aged 15—17 saw notable improvements since 2001; both their unintended pregnancy rate and unintended birth rate declined by roughly one quarter.
Many disparities among subgroups, already large, grew. In particular, cohabiting women exhibited very high and increasing unintended pregnancy and unintended birth rates. Like married women, cohabiting women are regularly sexually active, but are less likely than married women to desire pregnancy, and thus are at a very high risk for unintended pregnancy. They are, however, more likely to carry a pregnancy-including an unintended pregnancy—to term than unmarried noncohabiting women, perhaps because they have more partner support. In addition, the decline in the proportion of unintended pregnancies ending in abortion may have been related to increased normalization of childbearing among these couples. These findings represent consequences of broad demographic trends—specifically, fewer married women and a greater proportion of childbearing to unmarried women—and also help to explain those trends by showing that cohabiting couples, regardless of marital status, have high pregnancy rates and that a large proportion of those pregnancies are unintended.
Poor and low-income women also experienced some of the greatest increases and highest rates of unintended pregnancy. This finding is consistent with numerous studies that document the association between disadvantage and higher risk for unintended pregnancy due to risky sexual behavior, including less vigilant contraceptive use patterns [27–29]. The upward trend in their unintended pregnancy rate has continued for over a decade . While reasons behind these behaviors are not fully understood, they are related to the significant life challenges facing many of these women [30,31].
The disparities by parity are probably explained by the desire for families with two children. In other words, the high intended and unintended rates for women with one birth compared with childless women or those with two or more births may be due to the fact that women reporting only one birth may be more likely to have a second birth, but are less likely to progress to a third birth . At the same time, their high unintended pregnancy rate suggests that mothers have difficulties timing births, and their high unintended birth rate suggests less concern about continuing an unintended pregnancy compared with other women.
This is an aggregate-level analysis incorporating data from multiple datasets, which makes statistical testing difficult. One test that can be performed is a comparison based on a subset of our data: the proportion of pregnancies ending in birth (i.e., excluding abortions, which are underreported, and miscarriages) that were unintended in 2006 and 2001. The overall percentage increase, from 35% to 36%, was not significant, although the increase among women 20—24 years, from 45% to 53%, was significant at the p<.10 level. Nonetheless, we do see substantively significant changes in unintended pregnancy rates in several subgroups. This argues that the limited tests on a subset of our key statistic do not capture the whole picture, and their results should not be considered conclusive.
In conclusion, the United States did not make progress toward its goal of reducing unintended pregnancy between 2001 and 2006. To better understand what drove these rates up, we are currently conducting a demographic analysis of changes in population composition and reproductive health behaviors that have historically affected them. But given the nation’s increasingly high unintended pregnancy rate and the fact that 11% of the population at risk does not use birth control , reducing the unintended pregnancy rate requires that we focus on increasing and improving contraceptive use among women and couples who want to avoid pregnancy. Increased use of long-acting and cost-effective contraceptive methods such as the IUD could play an important role in such an effort. In particular, the age at which childbearing begins has increased , and the length of time from first intercourse to first birth is on average 8 years; this is a period of potential risk for women and couples, and should be seen as an appropriate time to use long-acting methods. The American Congress of Obstetricians and Gynecologists has indicated that such methods should be “first-line” choices for young women, and coupling IUDs with condoms for additional protection may have the potential to reduce unintended pregnancy even further [33,34]. Although these methods are highly cost—effective over time, even women with health insurance may have difficulty paying for these methods because some plans do not cover the high upfront costs or other charges women often incur to use them . Research indicates that when financial barriers are completely removed and comprehensive information is provided on all methods, women choose long-acting, highly effective methods in large numbers .
The authors would like to thank Stanley Henshaw, Rachel Jones and Megan Kavanaugh for reviewing the manuscript, and Jacqueline Darroch and Susheela Singh for providing guidance on study methodology. This study was supported by award R01HD059896 from the Eunice Kennedy Shriver National Institute of Child Health and Human Development. The content is solely the responsibility of the authors and does not necessarily represent the official views of NICHD or NIH.
aMiscarriage refers to spontaneous fetal loss or stillbirth.
bThis change resulted in lower estimates for 2001 than were previously reported .
cThe ratio of miscarriages to birth has not changed much between 1995 and 2006, so use of earlier 1995 data should not be problematic.
dIn 2006, 57% of miscarriages followed intended pregnancies, compared with 64% of births and 5% of abortions.
eFor example, in 2006, the proportion of miscarriages that were intended within each subgroup was calculated as A + (0.884 × [B – A]), where A is the proportion of abortions in that subgroup that were intended, B is the proportion of births in that subgroup that were intended, and 0.884 is (57% – 5%)/(64% – 5%), based on the overall proportions for the sample population mentioned in the previous footnote.
fAs described above, this calculation excludes miscarriages.
gThe term “unintended birth rate” is shorthand for the rate of births that followed unintended pregnancies.