The 2005 cesarean delivery rate in the United States was more than 30%, the highest U.S. rate ever reported.1 Indeed, the cesarean delivery rate has been steadily increasing over the last decade.2 Part of this increase may stem from the movement to perform a cesarean delivery in the absence of a medical or obstetric indication, termed a cesarean delivery on maternal request. Recent estimates place the rate of cesarean delivery on maternal request between 4% and 18% of all cesarean deliveries and between 2% and 5% of all deliveries in the United States.2,3
Cesarean delivery on maternal request is a controversial topic. On one hand, ethicists propose that it is a woman’s right.4 Additionally, advocates of cesarean delivery on maternal request promote it as a way to protect the pelvic floor,5,6 cite the convenience of a scheduled cesarean delivery compared with the unpredictability of the timing and length of labor and suggest that it may decrease the risk of unexplained stillbirth and neonatal morbidity associated with chorioamnionitis, cord prolapse, and fetal heart rate abnormalities. However, opponents stress the risks of elective cesarean delivery to both the mother and fetus. Not only is cesarean delivery a major surgical procedure (with associated morbidity and mortality), but it increases risk for complications in subsequent pregnancies including uterine rupture, placenta previa and accreta, placental abruption, and related maternal morbidity.7–11 Furthermore, elective cesarean delivery may increase the risk of iatrogenic neonatal prematurity and length of neonatal hospital stay.2
In 2006, the National Institutes of Health (NIH) convened a State-of-the-Science Conference on cesarean delivery on maternal request. Because the risks of placenta previa and accreta increase with each cesarean delivery, the panel concluded that it was not recommended for a woman desiring “several” children.2 The recent American College of Obstetricians and Gynecologists (ACOG) Committee Opinion echoed this recommendation.12 For cesarean delivery on maternal request to be an appropriate option, women should be able to accurately predict their final parity.
Our goal for the current study was to obtain a preliminary estimate of the accuracy with which women, at the time of their first pregnancy, are able to predict their final parity. Because almost 50% of pregnancies and over 30% of births are unplanned,13 we hypothesized that a substantial number of women would have more children than their planned maximum. We are unaware of any past or current studies addressing this question and undertook this study as an initial look into the accuracy of parity prediction.
MATERIALS AND METHODS
We conducted a cross-sectional study of women aged 18–50 years, drawn from a panel of Internet users administered by Survey Sampling International, who voluntarily agreed to receive invitations to fill out questionnaires. E-mail invitations were sent to a stratified random sample of panel members with the goal of approximating the U.S. census on education level, race, and income in the final subject pool. To ensure demographic diversity (but not representativeness as Internet users who sign up to take surveys may have distinct characteristics) and offset large expected variations in response rates (especially for African Americans and Hispanic Americans), we established target response levels for each racial/ethnic group. We also drew three distinct age samples within each race (one third aged 18–30, one third aged 30–39, and one third aged 40–50) to offset lower response rates from younger sample members. The number of e-mail invitations in each demographic subsample was dynamically adjusted until all target response levels were achieved. Upon completion, participants were entered into a drawing administered by Survey Sampling International for modest cash prizes. We obtained exempt status approval from the University of Michigan Medical School Institutional Review Board as anonymous survey research. This survey was done in conjunction with a survey on amniocentesis preferences.14 The questions regarding parity came first in the survey, followed by amniocentesis preference questions, and finally demographic questions.
Women were asked whether they had given birth. Those who had given birth recorded their pregnancies and births. For all live births, participants provided the year and type of delivery and the gender of the child. Finally, women indicated whether they planned to give birth to more children.
Respondents who had previously given birth were asked to recall their original estimate of the maximum number of children to which they planned to give birth. The specific question wording was: “Please think back to when you were pregnant with your first child. At that time, what was the maximum number (or the most) children you thought you would give birth to in your life?” Response options were 1 child, 2 children, 3 children, and 4 or more children.
Women whose parity history indicated a larger number of children than originally estimated were classified as having underestimated their parity. These women received an additional section of the survey which first noted this discrepancy and then asked them (using yes/no response categories) to indicate possible reasons they had more children than planned. Options given were: 1) I had an unplanned pregnancy, 2) I wanted a child of the opposite sex, 3) I had twins, triplets, or more, 4) I divorced/separated from my previous partner and wanted children with my new partner, 5) one or more of my children died and I decided to have another, 6) I changed my mind, and 7) my husband or partner wanted to have more children.
Standard demographics were collected. Finally, number of siblings and number of partner’s siblings were recorded.
Based on the statements of NIH and the ACOG that cesarean delivery on maternal request is not recommended for women desiring several children,2,12 we analyzed the data for women who planned a maximum of two or fewer and three or fewer children.
We used χ2 for categorical dependent and independent variables and logit for dichotomous dependent variables and categorical or continuous independent variables to compare characteristics of accurate and inaccurate predictors. Intraclass correlation was used to estimate the correlation between actual parity and predicted parity. All analyses were performed using STATA 9.0 (StataCorp, College Station, TX).
A total of 1,785 women reached the survey website and viewed the first content page. Of these, 10 were excluded for reporting ages outside of the requested sample range and 56 were excluded because they did not report information regarding their parity and/or future childbearing plans. Of the remaining 1,719 women, 626 (36.4%) were nulliparous, 267 (15.5%) were parous and planning more children, and 826 (48.1%) were parous women who reported they had completed childbearing. Women were, on average, 35 years old (range 18–50). Seventy-three percent of respondents described themselves as Caucasian, 12% as African American, and 12% as Hispanic (any race). Our sample was educationally diverse, with 27% having completed a Bachelor’s or higher college degree, but also 19% with only a high school education or less.
The percentage of women who planned to have a minimum of one child but a maximum of two differed significantly by birth status. Thirty-six percent of nulliparous women (average age 28 years) estimated they would have a maximum of two children, while 39% of parous women planning more children (average age 29 years) and 56% of parous women who completed childbearing (average age 39 years) reported that, at the time of their first pregnancy, they planned a maximum of two children (overall P<.001).
To estimate the accuracy of the parity predictions, it was necessary to limit analyses to women who reported that they had completed childbearing. The most relevant group to examine is women who had predicted (just before their first delivery) that they would have a maximum of two or fewer children. Thus, the subsequent analyses are limited to women who reported that (1) they had completed childbearing and (2) at the time of their first pregnancy resulting in a birth, they planned a maximum of two or fewer children. This resulted in a sample of 458 women. Of those women who planned a maximum of two or fewer children, 30% were primiparous and 70% were multiparous. As expected for women who do not plan on having more children, the mean age of this subsample was older than that of the overall sample (39 years compared with 33 years, P<.001). Most of the women in the subsample were married (74%), 79% described themselves as Caucasian, 14% as African American, and 13% as Hispanic (any race). Fifty-one percent of these respondents had some college education, with 21% having completed a Bachelor’s or higher college degree, but also 27% with only a high school education or less. Ten women reported a multiple gestation; we adjusted for multiple gestation as a possible reason for inaccuracies in prediction by subtracting one from the total number of children for women who had twins and two for the women who had triplets (as it is the number of pregnancies resulting in delivery that we are actually interested in as opposed to the total number of children which was used as a proxy for the total number of deliveries).
There was poor correlation between planned maximum number of children and actual parity (r=0.04, 95% confidence interval [CI] 0.00–0.13). Only 44% (95% CI 40–49%) of these women accurately predicted their final parity, while 39% (95% CI 34–43%) underestimated their final parity. Of the under-predictors who had estimated a maximum of two children or fewer (n=179), 50% had three children and 21% had four or more children.
Demographic characteristics for accurate and inaccurate predictors can be found in Table 1. Accurate predictors were older at the time of first birth, had fewer siblings, and were less likely to be Hispanic compared with inaccurate predictors. There was no difference in current age, marital status, number of partner’s siblings, race, education or income between accurate and inaccurate predictors.
Among those who underestimated their final parity (n=194), reasons for having more children than initially planned included having unplanned pregnancy (66%), experiencing a change of mind (34%), partner desiring more children (30%), wanting a child of the opposite gender (23%), desiring children with a new partner (22%), or experiencing death of a child (6%).
Given the controversy of whether the recommended maximum number of cesarean deliveries for women planning cesarean delivery on maternal request should be two or three, we replicated the above analysis with women who planned a maximum of three or fewer children (n=648). Again, we found poor correlation between planned maximum number of children and actual parity (r=0.07, 95% CI 0.00–0.20) and 32% (95% CI 27–34%) under-predicted their final parity. Accurate predictors were also older at the time of first birth (24 years compared with 21 years, P<.001) and had fewer siblings (2.6 versus 3.2, P=.001). Again, among the underpredictors, the most common reason cited for underprediction was having an unplanned pregnancy (67%).
Our results indicate that at the time of first pregnancy, women are poor predictors of their final parity. Almost 40% of women planning a maximum of 2 children underestimate their final parity. Twenty percent of these report having four or more children and thus potentially would have had four or more cesarean deliveries if they opted for cesarean delivery on maternal request at the time of their first delivery. Consistent with our hypothesis, 66% of women cited unplanned pregnancy as a reason for underprediction.
When evaluating the appropriateness of advocating for cesarean delivery on maternal request, a particularly crucial component is the risks associated with multiple cesarean delivery. Many studies have shown that there are incremental risks associated with each subsequent cesarean delivery.10,11 This has led the NIH and ACOG to recommend it only for women planning a maximum of two or three children.2,12 Yet, if a woman chooses to have a cesarean delivery with her first delivery, she is likely to have a cesarean delivery with subsequent pregnancies. This expectation is based on decreased vaginal birth after cesarean delivery rates, which are likely due to both patient and provider preference and hospital policies limiting vaginal birth after cesarean delivery.15 Women who request cesarean delivery for the first pregnancy seem unlikely to opt for a vaginal delivery with subsequent pregnancies if reasons are fear of labor, protecting the pelvic floor, or controlling timing of delivery. For these reasons, women who have a cesarean delivery on maternal request and have more than one child are likely to have multiple cesarean deliveries. Our findings, that almost 40% of women underestimate their final parity, add further concern regarding the negative health benefits a woman might experience if she undergoes cesarean delivery on maternal request for her first birth. Thus, it is imperative that long-term risks, including placenta previa, accreta, and percreta be included when counseling women regarding delivery choice, as these risks increase with each subsequent cesarean delivery, regardless of women’s predictions of their final parity.
Our study had several limitations. Although we believe a long-term prospective study is ideal, currently (for our research team), the cost of following a cohort of women over 25 years (from first pregnancy to end of fertility) is prohibitive. We designed this cross-sectional study to obtain an estimate of the accuracy with which women, at the time of their first pregnancy, are able to predict their final parity. The cross-sectional design introduces the possibility of recall bias in the observed differences between planned number of children and final parity. It is likely that if recall bias was present, women’s estimates would more closely reflect their parity16,17. This would then inflate the percent of accurate predictors and likely underestimate the percent of underpredictors. In addition, we relied on each woman’s report that she had completed childbearing. Certainly some of these women may experience unplanned birth in the future, such occurrences would only serve to increase the proportion of women who underpredicted their final parity. Finally, because our sample was drawn from a panel of Internet users, our results may not generalizable to all populations. Certainly, in countries in which cesarean delivery on maternal request is widespread, women who pursue this option are often in the upper income levels or have achieved high educational status, characteristics similar to the Internet user population. However, our sample was ethnically diverse, and the racial make-up mirrored the racial composition of the United States. In addition, our sample had a wide range of educational attainment.
Given the numerous assumptions required for a discussion regarding the risks and benefits of cesarean delivery on maternal request, including the often inaccurate discussion of parity prediction, we urge both practitioners and patients to tread cautiously. Practitioners should educate women about the inaccuracies in parity prediction. Part of any discussion regarding cesarean delivery on maternal request should prompt a discussion regarding the incremental risks of three, four or more cesarean deliveries, even if the patient is sure that she plans only two children. From our study, we know that accurate predictors are older at the time of first birth. Certainly, this criterion is helpful if a woman is delivering her first child at age 40 and unlikely to have many future pregnancies. This predictor becomes less helpful if she is 33 or 35 years old. More research is needed to confirm our findings and to delineate further characteristics of women who are accurate in predicting final parity. In the future, we may be able to use certain characteristics to help counsel our patients regarding the best mode of delivery.
Our findings are an important addition to the ongoing international debate regarding cesarean delivery on maternal request and raise questions about the appropriateness of performing this procedure, even on women planning a maximum of two to three children.
2. National Institutes of HealthNIH State-of-the-Science Conference: cesarean delivery on maternal request. Obstet Gynecol 2006;107:1386–97.
3. Wax JR, Cartin A, Pinette MG, Blackstone J. Patient choice cesarean: an evidence-based review. Obstet Gynecol Survey 2004;59:601–16.
4. Minkoff H, Powderly KR, Chervenak F, McCullough LB. Ethical dimensions of elective primary cesarean delivery. Obstet Gynecol 2004;103:387–92.
5. Eason E, Labrecque M, Marcoux S, Mondor M. Anal incontinence after childbirth. CMAJ 2002;166:326–30.
6. Rortveit G, Daltveit AK, Hannestad YS, Hunskaar S, Norwegian ES. Urinary incontinence after vaginal delivery or cesarean section. N Engl J Med 2003;348:900–7.
7. Clark SL, Koonings PP, Phelan JP. Placenta previa/accreta and prior cesarean section. Obstet Gynecol 1985;66:89–92.
8. Miller DA, Chollet JA, Goodwin TM. Clinical risk factors for placenta previa-placenta accreta. Am J Obstet Gynecol 1997;177:210–4.
9. Juntunen K, Makarainen L, Kirkinen P. Outcome after a high number (4-10) of repeated caesarean sections. BJOG 2004;111:561–3.
10. Silver RM, Landon MB, Rouse DJ, Leveno KJ, Spong CY, Thom EA, et al. Maternal morbidity associated with multiple repeat cesarean deliveries. Obstet Gynecol 2006;107:1226–32.
11. Grobman WA, Gersnoviez R, Landon MB, Spong CY, Leveno KJ, Rouse DJ, et al. Pregnancy outcomes for women with placenta previa in relation to the number of prior cesarean deliveries. Obstet Gynecol 2007;110:1249–55.
12. Cesarean delivery on maternal request. ACOG Committee Opinion No. 394. American College of Obstetricians and Gynecologists. Obstet Gynecol 2007;110:1501.
13. Henshaw SK. Unintended pregnancy in the United States. Fam Plann Perspect 1998;30:24–9.
14. Zikmund-Fisher BJ, Fagerlin A, Keeton K, Ubel PA. Does labeling prenatal screening test results as negative or positive affect a woman’s responses? Am J Obstet Gynecol 2007;197:528.e1.
15. Centers for Disease Control and Prevention. Births: final data for 2004. Natl Vital Stat Rep 2006;55 (1).
16. Tourangeau R, Rips LJ, Rasinski K. The psychology of survey response. New York (NY): Cambridge University Press; 2000.
17. Loftus EFSK, Klinger MR, Fiedler J. Memory and mismemory for health events. New York (NY): Russell Sage Foundation; 1992.
© 2008 by The American College of Obstetricians and Gynecologists.