Trends In Primary Cesarean Section Rates Among Women With And Without Perinatal Mood And Anxiety Disorders

Abstract

Reducing the rate of cesarean sections among women considered at low risk for delivery by that method is a goal of Healthy People 2030. Prior research suggests that perinatal mood and anxiety disorders increase the risk for cesarean section, but data are limited. This cross-sectional study of commercially insured women examined the relationship between perinatal depression and anxiety disorders and primary (first-time) cesarean section rates, using administrative claims data for US in-hospital deliveries from the period 2008–17. Of the 360,225 delivery hospitalizations among 317,802 unique women, 24.0 percent included a delivery by primary cesarean section and 3.1 percent carried a diagnosis of depression, anxiety, or both made during the index pregnancy. Using an adjusted generalized estimating equation, we found that the predicted probability of primary cesarean section was 3.5 percentage points higher, on average, among women with these disorders compared with those without. Our findings confirm the importance of pursuing research to identify mechanisms by which perinatal depression and anxiety disorders increase the risk for primary caesarian section among women otherwise considered low risk for delivery by that method, as well as effective interventions.

Perinatal mood and anxiety disorders affect as many as one in five reproductive-aged women in the United States.^1–4 The prevalence of perinatal mood and anxiety disorders, which includes depression, anxiety, or both, more than doubled between 2006 and 2015 among childbearing women in the US,⁵ with suicide representing a leading cause of maternal mortality.⁶ During the time frame that the prevalence of perinatal mood and anxiety disorders increased, the prevalence of suicidal ideation and intentional self-harm in the year preceding or after a birth also rose.⁷ Furthermore, perinatal mood and anxiety disorders have numerous other negative consequences for women and their children, including adverse birth outcomes, higher rates of maternal morbidity,⁵ and lost earnings.⁸ In the US, estimated costs of untreated perinatal mood and anxiety disorders exceed $14 billion from conception through the first five years of a child’s life and may affect long-term use of health care resources for both women and their children.⁸

A few studies report higher rates of cesarean section among women with perinatal mood and anxiety disorders. Whether perinatal mood and anxiety disorders increase the likelihood of cesarean section is crucial to establish and understand, as cesarean deliveries have short- and long-term health consequences for women and infants.⁹ The largest US study to date, using data from the National Inpatient Sample, reported that between 2006 and 2015, US cesarean section rates were higher among women diagnosed with perinatal mood and anxiety disorders during their delivery hospitalization compared with women without such a diagnosis (37.7 versus 32.7 per 100 deliveries, respectively).⁵ However, this study had several notable limitations. Among these was an inability to identify women with perinatal mood and anxiety disorders before the delivery, in addition to lack of adjustment for other conditions known to increase the risk for cesarean section, such as abnormal placentation, multiple gestations, obesity, maternal age, and chronic conditions.⁹ Furthermore, the analysis was unable to distinguish primary cesarean sections from repeat cesarean sections; this is an important distinction because a history of prior cesarean section is considered a medical indication for a repeat cesarean section. Globally, other small, single-site studies have also reported higher rates of cesarean section among women with perinatal mood and anxiety disorders,^10–14 but findings have been inconsistent.¹⁵

Mechanisms that might account for an underlying relationship between cesarean section and perinatal mood and anxiety disorders remain largely a matter of conjecture. Suggested mechanisms or pathways include medical practice patterns used to manage maternal distress during pregnancy or labor, such as elective labor inductions or elective cesarean sections, or underlying associated placental dysfunction resulting in a higher probability of fetal distress during labor.^12,14,16 Clarifying the relationship between perinatal mood and anxiety disorders and primary cesarean section rates would enhance our understanding of the consequences of perinatal mood and anxiety disorders for health outcomes. In turn, that knowledge would have the potential to inform efforts to develop and evaluate effective treatments and policy interventions.

Reducing the rate of primary (first-time) cesarean sections in the US is a 2030 Healthy People objective and has been endorsed by numerous public health and professional societies.⁹ As previously noted, focusing specifically on primary cesarean sections is important to gaining insights into how that could be accomplished because once the decision is made to perform a cesarean section, the likelihood of successful vaginal deliveries in future pregnancies is relatively low. A vaginal delivery can rupture the scar created by the cesarian section. Although uncommon, uterine rupture also may occur, with potentially catastrophic consequences for mother and baby. As a result, more than 90 percent of US women undergoing a cesarean section for their first birth will go on to have a repeat cesarean section.⁹

Accordingly, we sought to clarify the relationship between perinatal mood and anxiety disorders and primary cesarean section among commercially insured women who are considered at low medical risk for primary cesarean section. We hypothesized that commercially insured women with perinatal mood and anxiety disorders would have higher rates of primary cesarean section compared with women without those conditions.

Study Data And Methods

In this analysis, we describe the relationship between perinatal mood and anxiety disorders and primary cesarean section rates in the period 2008–17 in a large cohort of commercially insured women delivering at hospitals in the US.

Data Source

The Maternal Behavioral Health Policy Evaluation study is a retrospective, observational cohort study using administrative claims data from Optum’s deidentified Clinformatics Data Mart database. These data consist of deidentified medical and pharmacy commercial insurance claims from January 1, 2008, to March 31, 2017, and include information on medical diagnoses and procedures, hospitalizations, inpatient and outpatient visits, and patient demographic characteristics. The database does not include information about what medical conditions or scenarios were considered during labor and delivery, so we could not reliably identify specific medical indications for cesarean section. The broader aims of the Maternal Behavioral Health Policy Evaluation study include assessing the effect of federal health insurance policy changes on health outcomes among a cohort of commercially insured delivering women with and without perinatal mood and anxiety disorders. The study site’s Institutional Review Board deemed the study not regulated because it contains deidentified data (HUM00164685).

Study Sample

Our analytic sample of delivery hospitalizations was drawn from a cohort of more than seven million women aged 15–44 years (the conventional definition of “reproductive age,” and the definition used by our data source) who were enrolled in employer-based health insurance. Delivery hospitalizations between 2008 and 2017 were identified using standardized International Classification of Diseases, Ninth Revision, Clinical Modification, and International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (ICD-9-CM and ICD-10-CM), diagnosis and procedure codes.¹⁷ Our unit of analysis was deliveries; therefore, women who delivered more than once during our study period could be included more than once during the observation period.

We further restricted the study sample to people with continuous enrollment in a single employer-based health plan for at least one year before and one year after delivery, people without any diagnosis or procedural codes indicating a prior cesarean section delivery, and people characterized to be at low medical risk for delivery by that method. We searched for evidence of a prior cesarean section in two ways. First we identified any delivery hospitalizations before index delivery date with Current Procedural Terminology (CPT) codes for cesarean section. Next we searched for codes indicating a cesarean section before the index delivery—for example, O34.21x (maternal care due to uterine scar from previous cesarean section) or 654.21/3 (previous cesarean section) or CPT codes 59610, 59612, 59614, 59618, 59620, or 59622. To identify women at elevated medical risk for delivery by cesarean section we used a commonly used claims data-derived maternal risk-assessment tool, the obstetric comorbidity index.^18,19 We defined low-risk people as those with an obstetric comorbidity index score equal to zero.

We identified women with perinatal mood and anxiety disorders using Healthcare Cost and Utilization Project algorithms for anxiety and depression based on ICD-9-CM and ICD-10-CM codes.^20,21 Women were flagged as having had perinatal mood and anxiety disorders if codes were identified at either two outpatient encounters or one inpatient stay during the twelve months preceding date of delivery or during the delivery hospitalization. We identified delivery type (vaginal or cesarean section) during the index delivery hospitalization, using ICD-9-CM and ICD-10-CM procedure codes, CPT codes, and diagnosis-related group codes (online appendix exhibit A1).²²

Analytic Methods

Our unit of analysis was delivery hospitalization. All outcomes and individual characteristics were attributed to a person’s delivery year. We used CPT codes for cesarean section during the delivery hospitalization as our primary outcome. We summarized demographic characteristics both for all people yearly and separated by those with and without perinatal mood and anxiety disorders. We plotted the proportion of cesarean section deliveries yearly for people with and without perinatal mood and anxiety disorders.

We used a generalized estimating equation, clustering at the individual level, to assess how the probability of a primary cesarean section from 2008 to 2017 varied by the presence or absence of perinatal mood and anxiety disorders. This model adjusted for delivery year, number of dependents in household, household income relative to federal poverty level (≤400 percent of poverty, >400 percent of poverty, unknown), age group (15–19, 20–24, 25–29, 30–34, 35–39, 40–44), race and ethnicity (White, Asian, Black, Hispanic, unknown or missing), insurance plan type (point of service, health maintenance organization or exclusive provider organization, preferred provider organization, indemnity or other), region (Great Lakes and Northern Plains, Mountain, Northeast, Pacific, Southeast, unknown), and health reimbursement arrangement or health savings account status (yes, no). We used annual guidance for poverty-level computations from the Department of Health and Human Service’s Office of the Assistant Secretary for Planning and Evaluation,²³ the household’s income, and the number of covered dependents to categorize women into two groups relative to the federal poverty level: 400 percent of poverty or less and more than 400 percent of poverty. We used 400 percent of poverty as a boundary because it both represents the approximate median household income in our sample and is the cutoff for subsidies or tax credits for health insurance under the Affordable Care Act.

We used predictive margins to compare the probability of cesarean section between those with and those without perinatal mood and anxiety disorders. We used two-sided statistical tests and 95% confidence levels for all statistical analyses. We performed all claims data management in SAS, version 9.4, and statistical analyses in Stata, version 14.1.

Limitations

The strengths of this study include its observation of a very large, national sample. These data allowed us to observe diagnoses and health service use throughout pregnancy and national trends over time.

Our study also has notable limitations. First, it is possible that other risk factors for primary cesarean section, such as maternal obesity,²⁴ fetal size, and physician-related factors,^9,25,26 do not reliably appear in claims data. The rationale underlying clinical decisions, including indication for cesarean section, is also not available. We also did not examine the screening patterns of perinatal mood and anxiety disorders. Furthermore, we could not observe services that insurance does not cover or services not submitted to the health plan. Finally, our sample included only commercially insured women and might not be generalizable to other populations of pregnant women. This study’s findings may not reflect the experiences of uninsured or Medicaid-covered women.

Study Results

Our analytic sample included 360,225 delivery hospitalizations occurring among 317,802 unique low-risk women. (Some women had multiple delivery hospitalizations during the study.) Of the delivery hospitalizations, 24.0 percent were primary cesarean deliveries and 3.1 percent had a diagnosis of perinatal mood and anxiety disorder prior to delivery or during the delivery hospitalization (data not shown). Women with perinatal mood and anxiety disorders tended to be older, be White, and reside in a household with income more than 400 percent of poverty compared with those without perinatal mood and anxiety disorders (exhibit 1). Complete patient characteristics are available in appendix exhibit A2.²²

Exhibit 1:

Characteristics of women at time of delivery, with and without perinatal mood and anxiety disorders

Maternal characteristic at time of delivery	2008, first study year		2012, midstudy		2017, last study year
	PMAD, % (n = 805)	No PMAD, % (n = 39,198)	PMAD, % (n = 1,032)	No PMAD, % (n = 38,590)	PMAD, % (n = 1,349)	No PMAD, % (n = 28,231)
Age group, years
15–19	1.5	2.7	2.6	2.7	1.2	1.5
20–24	5.8	8.0	9.1	11.0	10.0	10.2
25–29	23.9	27.2	22.2	24.7	20.5	21.9
30–34	35.7	35.3	33.4	37.1	37.8	39.6
35–39	19.8	20.3	21.2	18.3	22.6	21.4
40–44	13.4	6.6	11.4	6.2	7.9	5.3
Race and ethnicity
Asian	2.4	7.7	2.5	7.7	2.7	8.0
Black	5.8	9.3	4.7	9.2	6.5	7.6
Hispanic	7.7	13.5	7.0	12.6	6.3	13.5
White	81.0	66.5	84.5	68.3	81.6	66.4
Unknown or missing	3.1	3.0	1.4	2.1	2.9	4.6
Household income, as percent of federal poverty level
>400%	53.8	52.5	51.3	45.8	40.3	38.8
≤400%	25.8	26.6	31.0	34.0	34.8	35.7
Missing	20.4	20.9	17.7	20.2	24.8	25.5
Delivery type
Vaginal	72.2	77.1	73.7	78.1	77.2	82.3
Primary cesarean section	27.8	22.9	26.3	21.9	22.8	17.7

SOURCE Authors’ analysis of data from the Clinformatics Data Mart database. NOTES These data represent characteristics of women at time of unique delivery. Because some women experienced more than one delivery, the number of deliveries is greater than the number of women. PMADs is perinatal mood and anxiety disorder.

Between 2008 and 2017, annual primary cesarean section rates for women with perinatal mood and anxiety disorders remained higher than those for women without perinatal mood and anxiety disorders. The predicted probability of a cesarean section was 3.5 (95% CI: 2.7, 4.3) percentage points higher on average among women with perinatal mood and anxiety disorders compared with those without perinatal mood and anxiety disorders (p < 0.001) (data not shown). Trends in annual adjusted cesarean section rates between 2008 and 2017 did not differ between women with perinatal mood and anxiety disorders and those without (p = 0.809), and annual rates were consistently higher among women with perinatal mood and anxiety disorders.

Annual primary cesarean section rates fell during the study period for both groups (exhibit 2). The predicted probability of a primary cesarean section fell from 26.8 percent (95% CI: 25.9 percent, 27.6 percent) in 2008 to 21.8 percent (95% CI: 20.9 percent, 22.6 percent) in 2017 among women with a perinatal mood and anxiety disorder diagnosis (p < 0.001). Among those without a perinatal mood and anxiety disorder diagnosis, the predicted probability of primary cesarean section decreased from 23.2 percent (95% CI: 22.8 percent, 23.6 percent) to 18.2 percent (95% CI: 17.8 percent, 18.7 percent) between 2008 and 2017 (p < 0.001).

Exhibit 2. Adjusted trends in primary cesarean section among women with and without perinatal mood and anxiety disorders, 2008–17.

Source/Notes: SOURCE Authors’ analysis of data from the Clinformatics Data Mart database. NOTES Adjusted for delivery year, number of dependents in household, household income as percentage of federal poverty level, age group, race and ethnicity, plan type, region, and health reimbursement arrangement or health savings account status. Error bars represent 95% confidence intervals. PMAD is perinatal mood and anxiety disorders.

Discussion

The finding that perinatal mood and anxiety disorders were associated with higher rates of primary cesarean section in a population of women otherwise at low risk for delivery by cesarean section extends our knowledge of the relationship between perinatal mood and anxiety disorders and a critical birth outcome. By focusing on primary cesarean section rates among women at low medical risk for delivery by that method, this study provides stronger evidence than previous studies that perinatal mood and anxiety disorders may increase the risk for cesarean section. Furthermore, these findings suggest that primary cesarean section rates, when the mother is otherwise at low risk for that procedure, may be a feasible and meaningful negative outcome to include when evaluating the effectiveness of perinatal mood and anxiety disorder treatment and other interventions intended to mitigate negative pregnancy outcomes.

In the US, approximately 25 percent of low-risk, first-time mothers deliver via cesarean section.^9,27 Cesarean sections, although potentially lifesaving, incur risks for women and infants and increase costs for women, payers, and health care systems.²⁸ Primary cesarean sections increase the risk for maternal complications during the index surgery, including serious infections, hemorrhage, and intraoperative injuries. Primary cesarean sections also increase the risk for complications in future pregnancies, including uterine rupture and placental abnormalities, both of which can have catastrophic consequences for both mother and infant.^9,29,30 Nearly all women who undergo a primary cesarean section will have a subsequent repeat cesarean section.⁹

Healthy People 2030 identifies reduction of primary cesarean sections among women otherwise at low risk for that procedure as a critically important target. The stated aim is to reduce primary cesarean deliveries in this population by about 9 percent, or 2.3 percentage points.²⁷ Similarly, a joint statement from the Eunice Kennedy Shriver National Institute of Child Health and Human Development, the Society for Maternal-Fetal Medicine, and the American College of Obstetricians and Gynecologists highlights the importance of identifying and reducing unnecessary primary cesarean deliveries.⁹

Numerous factors determine primary cesarean section rates, including medical considerations, labor management style, physician incentives, medical legal concerns, and patient expectations.⁹ Primary cesarean sections in response to maternal request are discouraged but are estimated to occur in approximately 2.5 percent of all births.³¹ This study could not examine these factors. Future research that examines the underlying rationale for the decision to perform a cesarean section, including patient preferences, is needed.

This study’s findings build on prior research suggesting that perinatal mood and anxiety disorders may increase the risk for cesarean section by focusing on primary cesarean section among women otherwise at low medical risk for delivery by that method. Kimberly McKee and colleagues used the National Inpatient Sample to examine the relationship between perinatal mood and anxiety disorders and adverse birth outcomes, including events of severe maternal morbidity and mortality, during nearly eight million delivery hospitalizations in the US.⁵ That study also found that perinatal mood and anxiety disorders were associated with an increased rate of cesarean section compared with women without perinatal mood and anxiety disorders (32.7 percent versus 37.7 percent, respectively). Women with perinatal mood and anxiety disorders also had higher rates of severe maternal morbidity and mortality than women without perinatal mood and anxiety disorders (2.3 versus 1.5 per 100 delivery hospitalizations, respectively).⁵ However, analysis was limited by the fact that National Inpatient Sample data only include information coded during the index delivery hospitalization; diagnoses made during the antenatal period, including perinatal mood and anxiety disorders, medical conditions, or history of prior cesarean section, are not included. In contrast, a meta-analysis of four small studies, none of which was conducted in the US, failed to find an association between maternal anxiety and cesarean section, but there was significant heterogeneity across the studies. None of these studies examined primary cesarean section rates.¹⁵ This current study advances our knowledge by focusing on primary cesarean section among women medically at low risk for delivery by that method and including diagnoses made during the prenatal period.

Identifying and evaluating meaningful clinical, systems, or policy interventions aimed at reducing health consequences of perinatal mood and anxiety disorders, including rising rates of suicidality and self-harm during pregnancy,⁷ requires a better understanding of the underlying mechanisms. Unfortunately, we do not yet understand the biological, sociological, and system-level mechanisms underlying the relationship between perinatal mood and anxiety disorders and primary cesarean section. Several hypothesized but largely unstudied biologic pathways could increase the risk for primary cesarean section among women with perinatal mood and anxiety disorders. For instance, women with perinatal mood and anxiety disorders have elevated levels of stress hormones, which may alter placental development,¹⁶ resulting in placental dysfunction during labor—a situation that increases the risk for cesarean section. Alternatively, maternal distress generally—or fear of labor specifically—could influence a health care provider’s decision to schedule a woman for labor induction, a practice proven to increase the risk for primary cesarean section for all women compared with waiting for spontaneous labor to occur.^32–34 According to the American College of Obstetricians and Gynecologists, primary cesarean section because of maternal request alone is uncommon in the US and should be addressed with adequate pain management during labor, education, and emotional support.³¹ Achieving a better understating of these mechanisms should be a focus of future research. Health record data would be the preferred source of patient-level information in future studies, as it provides data on medical considerations that might influence the decision to perform a cesarean section, including fetal distress, prolonged or obstructed labor, or the rare presence of an absolute contraindication to vaginal delivery.

Conclusion

During the decade between 2008 and 2017, primary cesarean sections occurred at clinically and statistically significant higher rates among otherwise low-risk, commercially insured women with perinatal mood and anxiety disorders compared with among women without perinatal mood and anxiety disorders. This observation is problematic because cesarean deliveries cause short- and longer-term risks to both mother and child and typically rule out vaginal delivery in future pregnancies. Further studies should aim to uncover underlying mechanisms and to develop effective interventions.

BIOS for 2021-00780 (Dalton)

Bio1: Melissa K. Zochowski is a research specialist in the Department of Psychiatry, University of Michigan, in Ann Arbor.

Bio2: Giselle E. Kolenic is a statistician in the Department of Obstetrics and Gynecology, University of Michigan.

Bio3: Kara Zivin is a professor in the Department of Psychiatry, University of Michigan.

Bio4: Anca Tilea is a data and analytics manager in the Department of Obstetrics and Gynecology, University of Michigan.

Bio5: Lindsay K. Admon is an assistant professor in the Department of Obstetrics and Gynecology, University of Michigan.

Bio6: Stephanie V. Hall is a doctoral student in the Department of Psychiatry, University of Michigan.

Bio7: Agatha Advincula is a student intern, Benjamin Franklin Scholars, University of Pennsylvania, in Philadelphia.

Bio8: Vanessa K. Dalton (daltonvk@umich.edu; daltonvk@med.umich.edu) is a professor in the Department of Obstetrics and Gynecology, University of Michigan.