 |
 |
 |
|
|
|
Carol Sakala and Maureen P. Corry
Co-published by Childbirth Connection, the Reforming States Group, October 2008
and the Milbank Memorial Fund
(To see a complete list of Milbank reports, click here.
To print this report, click here for the pdf version.)
Foreword
Evidence-based maternity care uses the best available research on the safety and effectiveness of specific practices to help guide maternity care decisions and to facilitate optimal outcomes in mothers and newborns. Although the field of pregnancy and childbirth pioneered evidence-based practice, resulting in a wealth of clear guidance for evidence-based maternity care, there remains a widespread and continuing underuse of beneficial practices, overuse of harmful or ineffective practices, and uncertainty about effects of inadequately assessed practices.
In order to inform coverage and clinical policy decision making for maternity care, Childbirth Connection (CC), the Reforming States Group (RSG), and the Milbank Memorial Fund (MMF) collaborated to write, review, and publish this report. The report presents a discussion of current maternity care in the U.S. health care system and identifies key indicators that show the need for improvement. The report further summarizes results of the many systematic reviews that could be used to improve maternity care quality, identifies barriers to the use of evidence-based maternity care, and offers policy recommendations and other strategies that could lead to wider implementation of evidenced-based maternity care in the United States. These maternity care quality concerns and opportunities for improvement are not widely recognized at this time.
Organized in 1992, the RSG is a voluntary association of leaders in health policy in the legislative and executive branches of government, from all fifty states, Canada, England, Scotland, and Australia. The Milbank Memorial Fund is an endowed national foundation, established in 1905, that works with decision makers in the public and private sectors to carry out nonpartisan analysis, study, and research on significant issues in health policy. Established in 1918, Childbirth Connection (formerly Maternity Center Association) is a national not-for-profit voice for the needs and interests of childbearing families. Its mission is to improve the quality of maternity care through research, education, advocacy, and policy.
Many members of the RSG, as well as others knowledgeable in the field, reviewed successive drafts of this report. As a result of these reviews and the authors’ subsequent revisions, we believe that the information in this report is timely and accurate. The matters that have been highlighted by the authors do not necessarily represent the policy preferences of all the members of the RSG or of the other individuals who reviewed drafts of this report.
We thank all who participated in this project.
Eileen Cody
Chair, Health Care and Wellness Committee
Washington House of Representatives
Co-Chair, Reforming States GroupKevin Concannon
Former Director
Iowa Department of Human Services
Past Co-Chair, Reforming States GroupJohn Nilson
Member of the Legislative Assembly
Province of Saskatchewan
Co-Chair, Reforming States GroupMaureen P. Corry
Executive Director
Childbirth ConnectionCarmen Hooker Odom
President
Milbank Memorial Fund
Acknowledgments
Our sincere thanks to those who generously contributed to this report. They are listed in the positions they held at the time of their participation.
The following members of the Reforming States Group demonstrated initial and continuing enthusiasm, helped us set a direction, and worked with us to increase the utility of the report for policymakers: Laurie Monnes Anderson, Chair, Health Policy and Public Affairs Committee, Oregon Senate; Lee Greenfield, Senior Policy Advisor, Hennepin County Department of Human Services and Public Health; Toni Nathaniel Harp, Chair, Committee on Appropriations, Connecticut Senate; Pamela S. Maier, Chair, Health and Human Development Committee, Delaware House of Representatives; John M. O’Bannon, Member, Health, Welfare and Institutions Committee, Virginia General Assembly; Amy R. Paulin, Chair, Committee on Libraries and Education Technology, New York State Assembly; Charles K. Scott, Chair, Labor, Health and Social Services Committee, Wyoming Senate.
The following referees honored us with a careful reading and thoughtful feedback on ways to improve the report: Leah Albers, Professor, University of New Mexico; José M. Belizán, Adjunct Professor, Institute for Clinical Effectiveness and Health Policy, Buenos Aires; Ned Calonge, Chief Medical Officer, Colorado Department of Public Health and Environment; Rosemary Chalk, Director, Board on Children, Youth, and Families, Institute of Medicine; Frank Chervenak, Professor, Chair, and Director, Maternal Fetal Medicine, Department of Obstetrics and Gynecology, Weill Medical College of Cornell University; Helen Darling, President, National Business Group on Health; Suzanne Delbanco, Chief Executive Officer, The Leapfrog Group; Marilyn DeLuca, Executive Director, Jonas Center for Nursing Excellence; Murray Enkin, Professor Emeritus, Department of Obstetrics and Gynecology, McMaster University; Eunice Ernst, Director, American Association of Birthing Centers Consulting Group; Tina Clark-Samazan Foster, Assistant Professor, Department of Obstetrics and Gynecology, Dartmouth Hitchcock Medical Center; Ellen Hodnett, Professor, Faculty of Nursing, University of Toronto; Debbie Jessup, Legislative Assistant, Office of Representative Lucille Roybal-Allard of California; Holly Kennedy, Associate Professor, University of California at San Francisco; Valerie King, Associate Director, John M. Eisenberg Clinical Decisions and Communications Science Center, Oregon Health and Science University; Andrew Kotaska, Clinical Director, Department of Obstetrics and Gynecology, Stanton Territorial Hospital; Douglas Laube, Chair and Professor, Department of Obstetrics and Gynecology, University of Wisconsin School of Medicine and Public Health; Amy Levi, Interim Program Director, Midwifery, University of California at San Francisco; Mona Lydon-Rochelle, Associate Professor, University of Washington; Charles Mahan, Professor Emeritus, Department of Community and Family Health, University of South Florida School of Public Health; David A. Paul, Chair, Delaware Healthy Mother and Infant Consortium, Christiana Neonatal Associates; Richard G. Roberts, Professor, Department of Family Medicine, University of Wisconsin; Barbara Rudolph, Director, The Leapfrog Group; Marlene Smadu, Associate Dean of Nursing, University of Saskatchewan College of Nursing, Regina Site; Stephen Thacker, Director, Office of Workforce and Career Development, Centers for Disease Control and Prevention.
We also thank Tomoko Kushnir, Joann Petrini, and Rebecca Russell of the March of Dimes Perinatal Data Center, who contributed the table on maternal and newborn health indicators at the state level.
We also wish to acknowledge Daniel M. Fox, President Emeritus of the Milbank Memorial Fund, who further assisted in strengthening the report and provided support for development, production, and distribution.
Executive Summary
Synopsis
Effective maternity care with least harm is optimal for childbearing women and newborns. High-quality systematic reviews of the best available research provide the most trustworthy knowledge about beneficial and harmful effects of health interventions. A large, growing body of systematic reviews is available to help clarify effects of maternity practices, yet these valuable resources are grossly underutilized in policy, practice, education, and research in the United States. Practices that are disproved or appropriate for mothers and babies in limited circumstances are in wide use, and beneficial practices are underused. Rates of use of specific practices vary broadly across facilities, providers, and geographic areas, in large part because of differences in practice style and other extrinsic factors rather than differences in needs of women and newborns. These gaps between actual practice and lessons from the best evidence reveal tremendous opportunities to improve the structure, process, and outcomes of maternity care for women and babies and to obtain greater value for investments. This report points the way to achieving these gains for the large population of childbearing women and newborns and for those who pay for their care.
Report Aims
This report has several aims:
- to position maternity care within the U.S. health care system and to identify key indicators that clarify the need for improvement
- to present a framework for identifying the best available research, based on the principle of effective care with least harm, and to apply the framework to maternity care
- to summarize results of many systematic reviews that could be used to improve maternity care quality, with a focus on opportunities to increase benefit and/or reduce harm for large segments of the population of childbearing women and newborns
- to identify barriers to wider implementation of evidence-based maternity care in the United States
- to identify policy and other strategies that, if adopted, could lead to wider implementation of evidence-based maternity care in the United States
Report Audiences
This report is directed toward many stakeholder groups. It is a priority to communicate about these matters with policymakers who have legislative, executive, delivery system, purchasing, and other responsibilities for maternity care. Members of the Reforming States Group, a voluntary association of state-level health policymakers, have helped ensure that the strategies for quality improvement and other sections of this report can assist policymakers with efforts to improve maternity care. The report is also directed to others who are involved with maternity care, including health professionals and health profession educators, hospital and health plan administrators, insurers, employers, researchers, childbearing women and their families, consumer advocates, and journalists.
Maternity Care in the U.S. Health Care System
Childbearing is a major life passage for over 4.3 million mothers, newborns, and families annually in the United States. Within the U.S. health care system, childbirth is the leading reason for hospitalization. About 23 percent of all individuals discharged from hospitals are mothers or newborns. The current style of maternity care is procedure-intensive, and six of the fifteen most commonly performed hospital procedures in the entire population are associated with childbirth. Cesarean section is the most common operating room procedure in the country. Only three reasons for outpatient visits involve more visits annually than maternity care (prenatal and postpartum visits combined): general medical examination, progress visit, and cough.
Financing Maternity Care
Due to the large number of births per year and this technology-intensive style of care, hospital charges for birthing women and newborns far exceed those of any other condition. Costs of this care especially impact employers and private insurers, the primary payers for 51 percent of the births, and taxpayers and Medicaid programs, primary payers for 42 percent. “Mother’s pregnancy and delivery” is the most costly hospital condition for both Medicaid and private insurers, followed by “newborn infants.” These conditions are associated with 27 percent of hospital charges to Medicaid and 15 percent of hospital charges to private insurers.
Charges for childbirth vary considerably depending on the type and place of birth. The average hospital charge in 2005 ranged from about $7,000 for an uncomplicated vaginal birth to about $16,000 for a complicated cesarean section, and charges for newborn care, anesthesia services, and the maternity provider involved additional expense. By contrast, childbirth charges in a national survey of out-of-hospital birth centers were about one-quarter of the charges of uncomplicated vaginal birth in hospitals ($1,624 in 2003, when the national average charge for uncomplicated vaginal birth in hospitals was $6,239), in addition to charges for maternity provider services.
Actual payments tend to be lower than charges, but payment data are difficult to obtain. A recent analysis of a large database of payments for all maternity services (excluding newborn care) was weighted to reflect the national population of childbearing women with commercial insurance; the report concluded that average payments for cesarean births exceeded those for vaginal births by nearly 50 percent, adding several thousand dollars to insurers’ expenditures. Another recent analysis estimated that the average total prenatal and intrapartum expenditure for women with a code for “normal pregnancy and delivery” was $7,564 (2004 dollars), with over three-quarters of the expense concentrated in the hospital stay. Although the cost of prenatal care for Medicaid and privately insured women was similar, the hospital component of care for privately insured women was about $2,000 more than the hospital component for women with Medicaid coverage.
Performance of the U.S. Maternity Care System
Many performance indicators raise concern about U.S. maternity care. A mid-course review of national Healthy People 2010 objectives for the country found that we have been moving away from targets for many maternity objectives, including low birthweight and preterm birth measures, cerebral palsy, mental retardation, and cesarean measures. Changes in measurement make it difficult to understand trends in maternal mortality, which may be rising after stagnating with no improvement at the end of the past century. The national cesarean rate has increased annually from the mid-1990s and has reached a record level each successive year of the present century. Four percent of women lack access to insurance for childbirth, and a much larger proportion transitions from being uninsured to having insurance coverage during pregnancy. Rates of specific indicators vary widely across states. In comparison with white non-Hispanic and Hispanic mothers, black mothers experience a breadth and depth of disparity in maternity care delivery and outcomes. Cross-national comparisons from the World Health Organization and the Organisation for Economic Co-operation and Development clarify that many other nations are doing a better job with measures such as perinatal, neonatal, and maternal mortality, low birthweight, and cesarean rates. Nonetheless, per capita health expenditures for the United States far exceed those of all other nations. These outcomes, together with costly, procedure-intensive care, have been called the “perinatal paradox: doing more and accomplishing less.”
Framework for Evidence-Based Maternity Care
Evidence-based maternity care uses the best available research on the safety and effectiveness of specific practices to help guide maternity care decisions and to facilitate optimal outcomes in mothers and newborns. Various care paths that might be pursued in a specific situation can involve very different benefit/harm profiles. Evidence-based maternity care gives priority to effective care with least harm.
A rigorous, well-conducted systematic review of original studies yields the most trustworthy knowledge about beneficial and harmful effects of specific interventions. Randomized controlled trials are especially valuable original studies, but have some important limitations. Other types of study designs are often needed to help answer important questions. Many factors shape both views about suitable care and patterns of care, which often do not reflect the best current research. Thus, it is always important to ensure that policy and practice are in fact guided by the best available research. Informed decision making should consider safety and effectiveness as well as values and circumstances of individual women.
Although most childbearing women and newborns in the United States are healthy and at low risk for complications, national surveys reveal that essentially all women who give birth in U.S. hospitals experience high rates of interventions with risks of adverse effects. Optimal care avoids when possible interventions with increased risk for harm. This can be accomplished by supporting physiologic childbirth and the innate, hormonally driven processes that developed through human evolution to facilitate the period from the onset of labor through birth of the baby, the establishment of breastfeeding, and the development of attachment. With appropriate support and protection from interference, for example, laboring women can experience high levels of the endogenous pain-relieving opiate beta-endorphin and of endogenous oxytocin, which facilitates labor progress, initiates a pushing reflex, inhibits postpartum hemorrhage, and confers loving feelings. Large national prospective studies report that women receiving this type of care are much less likely to rely on pain medications, labor augmentation, forceps/vacuum extraction, episiotomy, cesarean section, and other interventions than similar women receiving usual care. Such physiologic care is also much less costly and thus provides outstanding value for those who pay for it. Burgeoning research on the developmental origins of health and disease clarifies that some early environmental and medical exposures are associated with adverse effects in childhood and in adulthood. Recognition of known harms and the possibility that many harms have not yet been clarified further underscores the importance of fostering optimal physiologic effects and limiting use of interventions whenever possible.
Overused Maternity Practices
Many maternity practices that were originally developed to address specific problems have come to be used liberally and even routinely in healthy women. Examples include labor induction, epidural analgesia, and cesarean section. These interventions are experienced by a large and growing proportion of childbearing women; are often used without consideration of alternatives; involve numerous co-interventions to monitor, prevent, or treat side effects; are associated with risk of maternal and newborn harm; and greatly increase costs. Mothers, babies, and purchasers would benefit from giving priority to effective, safer care paths and using risky interventions for well-supported indications only or when other measures are inadequate. The following practices would instead be consistent with the framework of this report: avoiding induction for convenience; using labor support, tubs, and other validated nonpharmacologic pain relief measures and stepping up to epidurals only if needed; and applying the many available measures for promoting labor progress before carrying out cesarean section for “failure to progress.” Such protocols would require considerable change in many settings, but would lead to a notable reduction in the use of more consequential procedures and an increase in cost savings. Available systematic reviews also do not support the routine use of other common maternity practices, including numerous prenatal tests and treatments, continuous electronic fetal monitoring, rupturing membranes during labor, and episiotomy.
Underused Maternity Practices
Systematic reviews also clarify that many effective maternity practices with modest or no known adverse effects are underutilized. Greater fidelity in providing these forms of care would lead to improved outcomes for many mothers and babies. In pregnancy, such care includes prenatal vitamins, smoking cessation interventions, measures for preventing preterm birth, and hands-to-belly maneuvers to turn fetuses to a head-first position before birth. The many beneficial, underused practices around the time of birth include continuous labor support, numerous measures that increase comfort and facilitate labor progress, nonsupine positions for giving birth, delayed cord clamping, and early mother-baby skin-to-skin contact. Best available evidence also supports providing access to vaginal birth after cesarean (VBAC) for most women with a previous cesarean. Systematic reviews also identify many strategies for increasing both establishment and duration of breastfeeding and effective ways to treat postpartum depression.
Barriers to Evidence-Based Maternity Care
Efforts to increase access to evidence-based maternity care should address barriers to quality improvement. Barriers to evidence-based maternity care include the following:
- lack of a set of robust maternity performance measures with buy-in of key stakeholders to use them for measuring, reporting, rewarding, and improving performance
- perverse incentives of payment systems
- adverse effects of the malpractice system
- primary reliance on specialists for providing maternity care to a predominantly healthy, low-risk population
- limited reliance on best evidence in leading guidelines for maternity care
- loss of core childbearing knowledge and skills among health professionals
- limited attention to harms and iatrogenesis
- challenge of translating research into practice
- adverse effects of pressure from industry
- inadequate informed consent processes and women’s lack of preparation for making informed decisions
- limitations of views put forth in media and popular discourse
Efforts to improve payment systems, the liability system, consumer decision making processes, and other factors that impact clinical decisions should identify best evidence and develop policies, programs, and processes that align these systems with optimal care.
Policy and Other Strategies to Help Align Practice with Evidence
Members of the Reforming States Group have worked with the authors of this report to identify the following priority strategies to increase provision of evidence-based maternity care:
- increase awareness about concerns with the present maternity care system and knowledge of evidence-based maternity care by educating and advising the range of stakeholders
- support research to further evidence-based maternity care
- reform the current reimbursement system to promote evidence-based maternity care and involve federal and state payers and private insurers
- require performance measurement, reporting, and improvement
The report provides specific recommendations for operationalizing these strategies.
Introduction
This report addresses the scientific basis for maternity practice. It begins by positioning care for the large, distinctive population of childbearing women and newborns within the U.S. health care system and describing performance on several maternity care quality indicators. The report then provides a framework for understanding “evidence-based maternity care,” including the relationship between evidence about human physiology and evidence about specific maternity practices. Evidence-based maternity care uses best available evidence to identify and provide optimal maternity care, defined as effective care with the least harm. The report then identifies a series of practices that are overused, as they have an unfavorable benefit/harm profile and good evidence points to the availability of effective, safer, and less costly options for most women. The next section identifies underused practices that offer established benefit with little or no identified risk. Mothers and babies would benefit from judicious, more restrictive use of the overused practices and more extensive use of underused practices. The examples identify important opportunities for improving the quality of maternity care for large proportions of mothers and babies through provision of effective care with minimal harm. While a comprehensive, up-to-date overview of best maternity evidence is needed, such an overview is beyond the scope of this technical report. Final sections of the report describe some of the greatest challenges to reducing the evidence-practice gaps and identify policy and practice strategies that might be used to narrow the gaps.
The Committee on Quality of Health Care in America and the Institute of Medicine’s landmark 2001 report, Crossing the Quality Chasm: A New Health System for the 21st Century, outlined fundamental concerns with the quality of health care in the United States. The report identified six aims for improvement that have been widely adopted: health care should be safe, effective, patient-centered, timely, efficient, and equitable. A major theme of the report was the importance of providing care that is based on the best available scientific evidence. The report identified impediments to such care, including underuse of beneficial care, overuse of services unlikely to offer benefits, and illogical variation in care from provider to provider and place to place. The report found that these concerns have troubling implications for health outcomes and efficient use of resources.
As detailed in the following section, more than 4.3 million babies are born in the United States every year, a life passage with major consequences for mothers, newborns, and families. Within the health care system, childbirth is the leading reason for hospitalization, and charges for birthing women and newborns far exceed hospital charges for any other condition. Notably, a follow-up Institute of Medicine report, Priority Areas for National Action: Transforming Health Care Quality, identified pregnancy and childbirth as a national priority area for health care quality improvement (Adams, Corrigan, et al. 2003).
Maternity care has attributes that distinguish it from much other health care. The “Bridges to Health” model identified childbearing women and infants as one of eight population segments with distinct characteristics that must be addressed if the entire population is to achieve the Institute of Medicine’s aims for improvement (Lynn et al. 2007). Another contribution points to numerous similarities between maternity care and end-of-life care, in contrast to the delivery of health care for many other conditions (Clark 2008). Distinctive attributes of care for childbearing families include the following:
- the challenge of caring for a primarily healthy population within acute care facilities that focus on treatment of pathology
- the difficulty of predicting how childbirth will unfold and the individual nature of the experience
- the importance of the continuous attentive presence of caregivers and loved ones
- the importance of respectful care of women and families—including clear communication, high-quality information, and control over decision making—and of their positive memories of the experience
- incentives arising from service bundling and global fee payment systems that encourage use of interventions and measures to hasten and control childbirth even though such care generally is not optimal for mothers and babies
- missed opportunities to prepare women to make informed decisions during their pregnancy and well before labor
- the challenge for women of making informed decisions about many crucial care matters while in labor and constraints on their choice at that time
- the great extent to which services could be calibrated to provide more appropriate care and to increase benefit and reduce harm and waste
- concerns about the severe impact of the malpractice system on maternity services
- exclusion of this clinical area from many established quality initiatives due to their focus, for example, on Medicare beneficiaries or chronic conditions
The evidence base for care during pregnancy and childbirth has been progressively developed and refined over several decades. Three comprehensive overviews of best evidence in the field were published in 1989: Effective Care in Pregnancy and Childbirth (Chalmers, Enkin, and Keirse 1989), A Guide to Effective Care in Pregnancy and Childbirth (Enkin, Keirse, and Chalmers 1989), and Oxford Database of Perinatal Trials (Chalmers 1989–92). Through updating and further development of these or successor products, along with the work of many other organizations, agencies, and individuals, a large, growing body of systematic reviews is available to guide maternity policy, practice, education, and research.
However, comparing current maternity care practice and performance in the United States to lessons from the best available research and to performance benchmarks reveals large gaps. Consistent with common patterns of innovation in medicine (McKinlay 1981), obstetric practices such as episiotomy (Graham 1997) and electronic fetal monitoring (Graham et al. 2004; Hoerst and Fairman 2000) were adopted prior to adequate evaluation. Implementation of best evidence has proven to be extremely difficult following adequate evaluation. Therefore, many practices that are disproved or appropriate for mothers and babies only in limited circumstances are in wide use. Conversely, numerous beneficial practices are underused because they offer limited scope for economic gain, are less compatible with predominant medical values and practices, have only recently been favorably evaluated, or due to other reasons. Beyond average overall gaps between evidence and practice, use of specific maternity practices varies broadly across facilities, providers, and geographic areas. This is primarily due to differences in practice style and other extrinsic factors rather than differences in needs of mothers and newborns. These gaps between where we are and what we could achieve present opportunities to improve the structure, process, and outcomes of care for mothers and babies and to obtain greater value for investments.
The discussion of overused and underused practices focuses on some of the greatest opportunities for increasing benefit and/or reducing harm for large segments of the population of childbearing women and newborns. It is not intended to be a comprehensive review of the evidence about maternity care. The Appendix points to many excellent resources for a more comprehensive understanding of evidence-based maternity care. Most are freely available to those with Internet access. Despite the abundance of resources, there are important areas where systematic evidence is not presently available and adequate to guide practice, such as evidence about effective pre- and interconceptional care, care for childbearing teenagers, and interventions to prevent and treat alcohol abuse and depression in pregnancy.
This report was developed to inform many stakeholder groups. It is a priority to communicate with policymakers who have legislative, executive, delivery system, purchasing, and other responsibilities about these matters. Sections on barriers to optimal care and on policy and other strategies for closing evidence-practice gaps are intended to assist policymakers with efforts to improve maternity care. The involvement of policymakers from the Reforming States Group has strengthened the entire report and these sections in particular. This report is also directed to many others who are involved with maternity care, including health professionals and health profession educators, hospital and health plan administrators, insurers, employers, researchers, childbearing women and their families, consumer advocates, and journalists.
Maternity Care in the U.S. Health Care System: Prominent Position, Large Expenditures, Troubling Performance
Maternity Care Leading Reason for Hospitalization/Office Visits
With over 4.3 million births every year, childbirth is the leading reason for hospitalization in the United States, exceeding such prevalent conditions as pneumonia, cancer, heart failure, bone fracture, and stroke (Kozak, DeFrances, and Hall 2006). Figure 1 lists leading major diagnostic categories by number of hospital discharges in 2005. Combined annual discharges for childbearing women and newborns greatly surpassed those for other major categories. In the 2005 Nationwide Inpatient Sample, 23 percent of all hospital discharges (9,144,958 among 39,163,834 total discharges) were for these two major diagnostic categories (Agency for Healthcare Research and Quality 2008).
Maternity care is also a leading reason for ambulatory visits. Within the entire population in 2003–2004, maternity care (combined prenatal and postpartum visits) was the fourth most common reason for an outpatient visit, exceeded only by general medical examination, progress visit, and cough, and representing 2.8 percent of all outpatient visits (Hing 2007).
Hospital Charges for Current Style of Maternity Care Highest of all Hospital Conditions
Hospitalization is by far the largest component of health care costs, and hospital charges for the current style of childbirth are considerable. Combined hospital charges for birthing women (about $44 billion) and newborns (about $35 billion) totaled $79,277,733,843 and far exceeded charges for any other condition in 2005 (Agency for Healthcare Research and Quality 2008).
In 2005, private insurers paid for 51 percent of hospital stays for childbirth in the United States, and Medicaid paid for 42 percent of these stays, with variation in these proportions across states. These payers were responsible for markedly greater proportions of childbirth payments than for all conditions combined (Figure 2) (Agency for Healthcare Research and Quality 2008).
Thus, the financial toll of maternity care on private payers/employers and Medicaid/taxpayers is especially large. In 2005, fully 27 percent of hospital charges (or $34,164,460,561) to Medicaid and 15 percent of hospital charges (or $39,726,164,301) to private insurers were for birthing women and newborns (Figure 3). “Mother’s pregnancy and delivery” was the most expensive condition for both payers, followed by “newborn infants” (Andrews and Elixhauser 2007).
The procedure-intensity of these hospital stays helps to explain the level of expense. In 2005, 49 percent of all hospital procedures performed on all individuals aged eighteen to forty-four were obstetric procedures, and six of the fifteen most commonly performed hospital procedures in the entire population involved childbirth (Agency for Healthcare Research and Quality 2008):
- medical induction, manually assisted delivery, and other procedures to assist delivery (number 2)
- repair of current obstetric laceration (number 6)
- cesarean section (number 7)
- circumcision (number 8)
- fetal monitoring (number 13)
- artificial rupture of membranes (number 14)
Six of the ten most common procedures billed to Medicaid and to private insurers in 2005 were maternity related (Table 1). Cesarean section was the most common operating room procedure for Medicaid, for private payers, and for all payers combined (Agency for Healthcare Research and Quality 2008).
Maternity care thus plays a considerable role in escalating health care costs, which increasingly threaten the financial stability of families, employers, and federal and state budgets (Blumenthal 2006; Cowan and Hartman 2005).
Great Variance in Charges and Payments by Type and Place of Birth
In 2005, the national average hospital charge for childbirth ranged from about $7,000 to nearly $16,000, depending on whether the birth was vaginal or cesarean and, further, was coded as uncomplicated or complicated (Figure 4). A national 2003 survey puts hospital charges in further perspective. In eighty-six freestanding birth centers across the country, the average childbirth charge was about $1,600, one-quarter of the hospital charge for uncomplicated vaginal birth that year (Figure 4). These figures do not include additional hospital-only charges for newborn care and additional anesthesia services charges for most hospital births, as well as maternity provider fees for all births. The National Birth Center Study of nearly twelve thousand women found excellent outcomes and very high levels of satisfaction with birth center care (Rooks et al. 1989; Rooks, Weatherby, and Ernst 1992a, 1992b, 1992c); this comparison suggests that the level of resource use in hospitals for uncomplicated vaginal births could be much lower. At present, less than 1 percent of childbearing women in the United States experience the style of care and efficient use of resources of freestanding birth centers (Martin et al. 2007).
Actual payments for medical services are generally somewhat lower than charges; however, payment data are often not publicly available. A recently reported analysis of a database of employer-sponsored health insurance attempted to measure comprehensive payments (rather than charges) for having a baby, including hospitalization, ambulatory visits, outpatient medications, laboratory services, and radiology/imaging services. Newborn care was not included, and elimination of outliers led to further underestimation of average payments. The database included about 1 percent of U.S. births in 2004, and was weighted to reflect the national population of childbearing women covered by commercial insurance. The study found that actual payments were well below charges. It also found a large differential between vaginal and cesarean births, with average payments for cesarean births ($10,958) exceeding average payments for vaginal births ($7,737) by nearly 50 percent. The average vaginal-cesarean differential was $2,090 for hospital payments and $723 for payment of professional fees (Thomson Healthcare 2007).
Another recent national analysis used federal Medical Expenditure Panel Survey results to estimate all prenatal and in-hospital childbirth expenditures in 2004 for women with a Clinical Classification Code of “normal pregnancy and delivery.” The analysis considered all sources of payment and included all professional services, hospital charges, prescription medications, and other expenses. Expenses associated with newborns appear to have been excluded. The analysis pooled and did not distinguish vaginal and cesarean births. Investigators estimated that combined average prenatal and childbirth costs were $7,564, with delivery expenses ($5,850) involving about five times the expense of prenatal care ($1,159). Expenditures for privately insured women were higher than average ($8,366 total, $6,520 delivery), and expenditures for women with Medicaid coverage were lower than average ($6,540 total, $4,577 delivery), with differences concentrated in the childbirth component. Privately insured women paid about 8 percent of the expenses out of pocket, and Medicaid-insured women were responsible for about 1 percent of expenses (Machlin and Rohde 2007).
Overall Performance a Concern and Many Trends Headed in Wrong Direction
The U.S. Department of Health and Human Services established national Healthy People 2010 objectives for the first decade of this century. A midcourse review of progress found movement away from targets for low birthweight and very low birthweight, all preterm birth (live births before thirty-seven completed weeks of gestation), preterm births of thirty-two through thirty-six weeks of gestation, maternal labor and birth complications, initial (“primary”) and repeat cesareans in low-risk women, cerebral palsy, and mental retardation. Numerous other maternity-related goals had not reached 15 percent of their targets at midcourse, including perinatal mortality—the child mortality measure most closely associated with the quality of maternity care (U.S. Department of Health and Human Services 2006).
In the quarter-century from 1981 to 2006, the national rate of preterm birth increased by 36 percent, and the proportion of low birthweight babies increased by 22 percent (Figure 5) (Hamilton, Martin, and Ventura 2007; Martin et al. 2007). Following a steady decrease through most of the twentieth century, maternal mortality stagnated from 1982 to 1998. Changes in the measurement of maternal mortality in the United States in 1999 (implementation of International Classification of Diseases, Tenth Revision) and in 2003 (new pregnancy status question on U.S. standard certificate of death) make it difficult to compare the most recent years with the period through 1998. The national maternal mortality rate was 8/100,000 live births in 1998 and 13/100,000 live births in 2003 (Centers for Disease Control and Prevention 1998; Hoyert 2007; Miniño et al. 2007).
Following a period of modest decline, the national cesarean rate rose by 50 percent from 1996 to 2006, setting a new record each year from 2000 onward. The repeat cesarean rate rose by 28 percent from 1996 to 2005, when 92 percent of mothers with a previous cesarean had a repeat cesarean. From 1990 to 2005, the proportion of medically induced labors rose by 135 percent, from 9.5 percent to 22.3 percent (Hamilton, Martin, and Ventura 2007; Martin et al. 2006; Martin et al. 2007). Moreover, validation studies suggest that these official rates—derived from aggregate birth certificates—identify just 45 percent to 61 percent of actual instances of induced labor (Lydon-Rochelle et al. 2005; Parrish et al. 1993; Piper et al. 1993; Yasmeen et al. 2006). In just over ten years, from 1990 to 2002, with an increasing proportion of induced labors and planned cesarean sections, the most common gestational age among singleton births in the United States fell from forty to thirty-nine weeks (Davidoff et al. 2006), and current trends suggest continued foreshortening of gestational age.
In national surveys, women who gave birth in U.S. hospitals in 2005 reported high rates of numerous new-onset physical and mental health problems in the first two months after birth, with many problems persisting to six months or more postpartum (Declercq et al. 2008).
Table 2 clarifies that there is large variation in these performance indicators across states, including greater than sixfold for vaginal birth after cesarean and greater than threefold for labor induction.
In 2005, 4 percent of births were uninsured, an increase of 12 percent over the previous year (Agency for Healthcare Research and Quality 2008). Moreover, a much larger proportion of women lacks insurance prior to pregnancy than at the time of birth and transitions to insurance coverage during pregnancy. We were unable to find data describing conditions since 1999, when a study across nine states found that from 17 percent to 41 percent of childbearing women lacked insurance prior to pregnancy, with 1 percent to 4 percent remaining uninsured through to the time of birth. From 13 percent to 35 percent of mothers made the most common insurance status transition, from uninsured to Medicaid. Levels of prepregnancy uninsurance, continuous uninsurance, and transition from uninsurance to Medicaid were considerably higher for women with annual incomes below $16,000 than for women with higher incomes (Adams, Gavin, et al. 2003). Current data, along with an understanding of the impact of insurance transitions in pregnancy on access to high-quality care and health outcomes, are needed. Uninsured childbearing women face bills for maternity services when they are adjusting physically and emotionally from pregnancy and childbirth and when their infants benefit from continuity of caregiver and breastfeeding.
When comparing experiences of childbearing women with private and public payment sources or across major race/ethnicity groupings, all segments of the population appear to experience problems with access to quality care. Where differences exist, there are greatest concerns about the quality of care received by black non-Hispanic women in comparison with both white non-Hispanic and Hispanic women (Sakala and Corry 2008). Similarly, black non-Hispanic mothers experience much higher rates of preterm birth, low birthweight, and fetal, perinatal, and maternal mortality than both other groupings (Martin et al. 2006). Moreover, the midcourse Healthy People 2010 review found that disparities for black non-Hispanic women were increasing for numerous indicators, including neonatal deaths, very low birthweight infants, mental retardation, and cerebral palsy (U.S. Department of Health and Human Services 2006).
Our national maternity care performance is also disappointing when compared with other nations. In The World Health Report 2005: Make Every Mother and Child Count, the World Health Organization identified twenty-nine nations with lower estimated maternal mortality rates than the United States (14/100,000 live births), thirty-five with lower early neonatal mortality rates (4/1,000 live births), and thirty-three with lower neonatal mortality rates (5/1,000 live births) in 2000 (2005). An analysis of maternal mortality rates for 2005 identified thirty-three countries with better performance than the United States (estimated at 11/100,000 live births, but perhaps as high as 21/100,000) (Hill et al. 2007). Among the thirty member nations of the Organisation for Economic Co-operation and Development (OECD), twenty-three reported a lower low birthweight rate than that of the United States (7.9 percent) for 2003, and six had higher rates. Fourteen OECD countries reported a lower perinatal mortality rate than that of the United States (6.9 percent) for 2003, and nine had higher rates. For the same year, nineteen members reported a lower cesarean rate than that of the United States (29.1 percent), and three reported higher rates (Organisation for Economic Co-operation and Development 2007).
Although maternity-specific expenditure level data are not available across a large set of countries, the United States had by far the greatest overall health expenditure per capita across the thirty OECD countries in 2005, which was greater than twice the average expenditure of these nations. Similarly, the United States far exceeded all other OECD countries in health expenditure as a share of gross domestic product in 2005 (Organisation for Economic Co-operation and Development 2007).
These disappointing, often deteriorating outcomes in concert with procedure-intensive care and very large financial investments have been described as “the perinatal paradox: doing more and accomplishing less” (Rosenblatt 1989).
Evidence-Based Maternity Care: Effective Care with Least Harm
Note: references to systematic reviews are in italics.“Evidence-based maternity care” uses the best available research on the safety and effectiveness of specific practices to help guide maternity care decisions and facilitate optimal outcomes in mothers and newborns. Various paths that might be pursued in a given situation often have very different benefit/harm profiles. Evidence-based maternity care gives priority to care paths and practices that are effective and least invasive, with limited or no known harms whenever possible. This framework is in the tradition enjoining practitioners to “first, do no harm” and consider undesirable consequences of good intentions.
The principle of effective care with least harm has two corollaries. First, practices with established or plausible adverse effects should be avoided when best available research identifies no clear anticipated benefit to justify their use. For example, mothers reported that a substantial proportion of labor inductions and cesarean sections in 2005 were carried out because of a caregiver’s judgment and concern about a large fetus (Declercq et al. 2006; National Collaborating Centre for Women’s and Children’s Health 2008b), but a series of rigorous reviews have found that best research does not support this as a valid indication for either procedure (Chauhan, Grobman, et al. 2005; Coomarasamy et al. 2005; Pattinson and Farrell 1997; Rouse and Owen 1999). An evidence-based framework also questions the wisdom of using interventions with a marginal expected benefit that is overshadowed by greater risk of established harm. Examples of such a situation include inducing labor by various means or hastening it with synthetic oxytocin for convenience and in the absence of a clear medical rationale (Grobman 2007; Simpson and Thorman 2005).
These principles for evidence-based maternity care are especially important in consideration of the sensitive perinatal development period, the potential for long-term beneficial and adverse health effects, and the large scope for uncertainty about unintended consequences of many possible exposures, as discussed in the following section. These principles are also guides for helping purchasers obtain good value.
To implement these principles and to help guide maternity care decisions, decision makers need access to the highest quality of evidence about the safety and effectiveness of specific procedures, medications, and other interventions. They should require rigorous research results demonstrating that the care provided has been shown to work, may thus be expected to offer genuine benefit, and is a wise choice when considering associated harms and alternatives.
Basic principles for determining what constitutes best available evidence are as follows:
- Question common assumptions. Maternity care practices based on the opinions of experts or the general public or on tradition are unreliable guides for decision making. These views and patterns of care have been shaped by many factors and often do not reflect the best current research. They may lead to inadequate care, poor outcomes, and wasted resources. It is important to demand to be shown the best evidence.
- Know that many studies of interventions are unreliable guides for decision making. Careful evaluation of the quality of research using “critical appraisal” skills is essential. Many studies are flawed or limited in scope and do not provide valid answers to key questions. One newly reported study rarely offers the best, most definitive answer, and commercial interests influence many studies. It is important to ask what is already known about a particular question on the basis of the best available research, and what, if anything, a new study adds.
- Look for the “gold standard.” When available, well-designed and well-conducted systematic reviews of research should inform maternity care decisions. If systematic reviews are not available, well-designed and well-conducted studies with randomized controlled trial designs can provide the most valid answers to many questions. For many reasons, it may be important to consider other types of studies as well. (See sidebar titled “What is the ‘Gold Standard’ for Knowledge about Effects of Maternity Care?” for more about systematic reviews and original studies.)
- Make informed decisions that consider evidence about safety and effectiveness and the values and circumstances of individual childbearing women. When making maternity care decisions, it is crucial to consider the best available evidence as well as values, preferences, and individual circumstances of childbearing women who have been supported to understand this evidence. It is also important to consider the options within specific care settings, such as the skills of caregivers and available forms of care.
- Beware of misleading claims. With growing recognition of the value of evidence-based policy and practice, it is important to be wary of bandwagon slogans describing “evidence-based” products and services and of deeply flawed execution that may not in fact reflect these principles.
The Physiologic Foundation of Evidence-Based Maternity Care
Note: references to systematic reviews are in italics.Short-Term Health Benefits of Physiologic Maternity Care
In addition to evidence about interventions, an evidence-based maternity care framework must take into account evidence about the biological foundation of childbearing: how mothers’ and babies’ bodies work, and work in concert, from prenatal through postpartum periods, to accomplish growth, development, the childbirth process, the establishment of breastfeeding, and attachment as the basis of the mother-child relationship and other relationships.
The authors of A Guide to Effective Care in Pregnancy and Childbirth, the highly regarded manual on evidence-based maternity care, gave priority to the biological foundation and accorded fundamental respect to mothers and babies when they described their framework for interpreting evidence on interventions:
We worked from two basic principles: first, that the only justification for practices that restrict a woman’s autonomy, her freedom of choice, and her access to her baby, would be clear evidence that these restrictive practices do more good than harm; and second, that any interference with the natural process of pregnancy and childbirth should also be shown to do more good than harm. We believe that the onus of proof rests on those who advocate any intervention that interferes with either of these principles (p. 486 Enkin et al. 2000).
This report affirms these principles. With appropriate support and protection from external interference, childbearing women and their fetuses/newborns experience innate, mutually regulating, hormonally driven processes that have developed during human evolution. These processes facilitate the period from the onset of labor through birth of the baby and placenta, as well as the establishment and continuation of breastfeeding and the development of mother-baby attachment. Examples of steps along this path include the following (Buckley 2004a; Winberg 2005):
- the mother’s elevated levels of beta-endorphin, an endogenous opiate that relieves pain and facilitates an altered state of consciousness, similar to experiences of endurance athletes
- the mother’s rhythmic involuntary expulsion efforts shortly before birth (Ferguson’s reflex)
- the unmedicated and undisturbed infant’s drive to crawl on its mother’s chest, self-attach to the breast, and begin suckling shortly after birth
- the mother’s surge of oxytocin at the time of birth, which stimulates loving feelings and inhibits hemorrhage by contracting the uterus
- the continuing oxytocic effects with breastfeeding
When facilitated, these autonomic nervous system functions overwhelmingly succeed in conferring a cascade of physical, psychological, and social benefits for the mother-baby dyad (Buckley 2004a; Odent 2001; Winberg 2005). When caregivers recognize and give priority to these capacities, mothers and babies experience these benefits and avoid risk of known short- and long-term harms and as yet unknown harms of avoidable, medically unnecessary interventions. By mobilizing these capacities, caregivers also humanize childbirth, show respect to women and fetuses/newborns as agents of these processes, enable all involved parties to experience the remarkable competence of birthing women and newborns, strengthen mother-baby bonds, and foster a uniquely fulfilling and empowering experience (Wagner 2001).
Many historic and contemporary reports and studies confirm that the physiologic approach to childbirth, which has most consistently been provided by midwives (Brown and Grimes 1995; Hatem et al. in press; Kennedy and Shannon 2004; Waldenström and Turnbull 1998; Walsh and Downe 2004), has succeeded remarkably well in achieving positive outcomes for mothers and babies in diverse contexts. These include situations that are often viewed as involving elevated risk, such as care for women in remote and inner-city settings and care among low-income and underserved populations (Raisler and Kennedy 2005; Ulrich 1990). In addition to such physiologic care, childbearing women and newborns benefit when deprivation, disease, inadvertent use of unsafe practices, or other adverse circumstances are minimized. In all contexts, a portion of childbearing women and newborns require and gain benefit from specialized skills and knowledge and obstetric interventions that effectively address specific problems. Access to consultation, referral, shared care, transferred care, and transport is an essential complement to physiologic care.
Unlike most recipients of health care in the United States, childbearing women and newborns are primarily healthy and benefit especially from care that maintains good health (Lynn et al. 2007). It is a challenge to provide an overall estimate of the portion of this population that can benefit from more specialized care and procedures that intervene in physiologic processes. The federal Healthy People 2010 initiative provides one widely used proxy measure, which identifies low-risk women as those who are giving birth at term (thirty-seven completed weeks of gestation or beyond) with a single infant in a head-first position (U.S. Department of Health and Human Services 2000). In 2003, 82.6 percent of childbearing women met these criteria (National Center for Health Statistics 2006). By this estimate, more intensive and invasive care is appropriate for about one mother in six.
Physiologic childbirth is within reach of the great majority of mothers and babies. However, this approach is poorly recognized and supported at present in the United States and other industrial nations. External, professional-directed management of childbirth in hospitals (Table 3) typically interferes with these mother- and baby-led capacities. Results in the table from the national Listening to Mothers II survey of women who gave birth in U.S. hospitals in 2005 clarify the extent of use of obstetric interventions in this primarily healthy population. A project of Childbirth Connection, this survey was conducted by Harris Interactive in January–February 2006 among 1,573 women across the United States. The methodology was designed to describe the survey’s target population of women aged eighteen through forty-five who gave birth to single babies in U.S. hospitals in 2005, with the baby still living at the time of the survey (a detailed appendix describes the methodology). The survey covered the time from before conception through the postpartum period. It included many items that are not available through other national data sources or appear to be undercounted in those sources; many validation studies (described in the survey’s report appendix) have found that birth certificates and hospital discharge records do not capture a large proportion of actual occurrences for many data items. The Listening to Mothers II survey is thus a unique resource for describing contemporary experiences of childbearing women and newborns in the United States and for comparing care that is actually received with optimal care (Declercq et al. 2006).
As shown in Table 3, labor is literally pushed by routine or common measures applied to this primarily healthy population—measures including labor induction, labor augmentation, staff-directed maternal pushing, and forceful pressure applied by staff on women’s abdomens at the time of birth. Labor is also frequently pulled by interventions such as vacuum extraction/forceps, cesarean section, pulling on the cord to hasten birth of the placenta, and separation of babies from mothers after birth. About one-half of the items in Table 3 were experienced by a majority of women despite the overall healthy status of this population. These and other common interventions disrupt and preclude the physiologic capacities of the childbirth process (Buckley 2004b; Odent 2001) and incur a cascade of secondary interventions used to monitor, prevent, and treat the side effects of the initial interventions (Brody and Thompson 1981). As one intervention justifies or increases the likelihood of using others, the cumulative effect is to create a distorted understanding of childbirth as a time when things are likely to go wrong and intensive medical management is required (Mold and Stein 1986).
By learning from those with the skills and knowledge to enhance the innate physiologic capacities of the childbearing process, we can refrain from exposing mothers and babies to the harm and expense of avoidable interventions and use medical interventions appropriately, as needed. Table 4 compares rates of several interventions among a national group of low-risk women receiving usual care with those in a large prospective study of American women who gave birth with certified professional midwives (CPMs) in 2000 (Johnson and Daviss 2005). The usual care group is composed of all women who met Healthy People 2010 criteria for low-risk laboring woman. The contrast in experiences is striking, with national rates of intervention among low-risk women with usual care from two to sixteen times as great as the midwifery study rates. Notably, both the CPM study and an earlier large prospective U.S. study of low-risk women who also received physiologic care (Rooks et al. 1989) reported a cesarean section rate of 4 percent. By contrast, the low-risk mothers with usual care in 2000 were five times as likely to experience this procedure.
Infrequent use of interventions and other conditions of the CPM study were not associated with increased risk for study participants when compared with low-risk women giving birth in usual care hospital settings (Johnson and Daviss 2005). The low CPM study rates of intervention are benchmarks for what the majority of childbearing women and babies who are in good health might achieve.
(See the sidebar titled “National U.S. Midwifery Credentials” to learn more about the relatively new CPM credential and the other two national midwifery credentials: certified nurse-midwife and certified midwife.)
Just as mothers and babies have much to gain by forgoing avoidable drugs, surgery, and other consequential procedures (as detailed in the following subsection), so can purchasers obtain exceptional value by using scientific evidence to provide effective care with least harm. Figure 4 contrasts the average charge for physiologic care in eighty-six freestanding birth centers across the United States with national average hospital charges for childbirth. In the best case, for uncomplicated vaginal birth, hospital charges were on average four times as high as birth center charges in 2003. Hospital charges were more than nine times as high as birth center charges when the pregnancy ended with a complicated cesarean. With wider application of care that facilitates physiologic processes, it is reasonable to expect that a notable proportion of births could shift from complicated to uncomplicated status and from cesarean to vaginal birth, with considerable benefit for those who receive and pay for care.
Long-Term Health Benefits of Physiologic Maternity Care
It is critical to retain a long-term, life-course focus when planning and providing care for babies and mothers. A vast body of research is accumulating about lifelong implications for babies of the medical, physical, and social environment from conception through pregnancy and birth and into the postpartum period. This early period includes windows of heightened sensitivity for fostering many dimensions of optimal human development or generating harm. Many papers review specific topics within this work on the “developmental origins of health and disease” (see, for example Csaba 2007; Davis and Sandman 2006; Gluckman and Hanson 2006; Gluckman et al. 2005; Grandjean and Landrigan 2006; Heindel 2006; Horta et al. 2007; Ip et al. 2007; Johns, Jauniaux, and Burton 2006; Lewis, Poore, and Godfrey 2006; Olsen 2000; and Tchernitchin et al. 1999). Alterations in genes, cells, and tissues can have mutagenic, teratogenic, carcinogenic, and other adverse effects. Many health problems that manifest in later childhood or adulthood appear to have origins in this much earlier period, following impairment of immune, neurobehavioral, reproductive, metabolic, cardiovascular, and other functions. Due to delay or failure to recognize effects or to establish associations with early exposures, the concept of “silent epidemics” has been proposed as an extension of areas of current knowledge (Grandjean and Landrigan 2006). Collectively, this work suggests the importance of rigorous assessment of possible long-term effects of perinatal exposures. Given current uncertainty, it would be prudent to avoid needless exposures. (The sidebar titled “Diethylstilbestrol (DES) Clarifies Importance of Caution with Perinatal Exposures” provides an example of one of the most carefully and longest documented perinatal exposures, diethylstilbestrol, or DES.)
Growing evidence also suggests that maternity practices can have a long-term positive or negative impact on maternal well-being—for example, whether mothers use medication such as DES during pregnancy, have a cesarean, or breastfeed (Ip et al. 2007; Kennare et al. 2007; Labbok 2001; Lauver, Nelles, and Hanson 2005; Silver et al. 2006). It is a priority to understand longer-term effects of maternity interventions on mothers as well, and to consider this knowledge during decision making processes.
In addition to environmental exposures and aspects of pregnant and breastfeeding women’s nutritional status, accumulating evidence finds that medical interventions used during childbirth may be associated with long-term harms (Odent 2006). These results and other possible impacts warrant further research and assessment in systematic reviews to strengthen our knowledge about long-term effects of widely experienced exposures during apparently sensitive windows of time in the perinatal period. A new narrative review describes research to date to understand mechanisms and effects of medical and environmental exposures in the perinatal period and to distinguish perinatal exposures from teratogenic exposures during early gestation (Csaba 2007; see also Tchernitchin et al. 1999). The implication, which is consistent with the framework of this report, is that interventions should only be used when there is a well-supported clinical rationale for doing so. Further, decision making processes should take into account known harms and recognize the potential for harms that have not yet been established or well-publicized. Studies that point to potential adverse developmental consequences of intrapartum interventions consistent with the growing understanding of the developmental origins of health and disease include the following:
- Babies exposed to antibiotics during the birth process were more likely than unexposed babies to experience persistent wheezing measured at age six to seven years (Rusconi et al. 2007).
- In comparison with healthy term newborns delivered by planned cesarean, healthy term newborns who experienced labor had improved survival of white blood cells that destroy microorganisms (neutrophils) in their cord blood and better neutrophil function; this suggests that labor may be immunologically beneficial to normal newborns and may help explain excess neonatal morbidity and mortality with planned cesareans (Molloy et al. 2004).
- The initial colonization of the newborn intestine persists over a long period and has a pivotal effect on long-term health (Bedford Russell and Murch 2006; Glasgow et al. 2005; Grolund et al. 1999), and babies who experienced cesarean section, failure to breastfeed, intrapartum antibiotics, or hospital birth were less likely to have early colonization with beneficial bacteria than those who were, respectively, born vaginally, breastfed, not given antibiotics, or born at home (Penders et al. 2006).
- Cesarean section is associated with numerous adverse future harms in women, including abdominal adhesion formation and chronic pelvic pain (Almeida et al. 2002; Lyell et al. 2005; Morales, Gordon, and Bates Jr. 2007) and in mothers and babies in future pregnancies, including placenta previa, placenta accreta, placental abruption, uterine rupture, hysterectomy, small size for gestational age, low birthweight, preterm birth, stillbirth, and neonatal intensive care unit admission (Getahun et al. 2006; Kennare et al. 2007; Taylor et al. 2005). Serious maternal morbidity increases progressively as the number of previous cesareans increases (Nisenblat et al. 2006; Silver et al. 2006).
- In contrast with unmedicated babies, babies whose mothers received epidurals and/or systemic opioids during labor exhibited reduced breast-seeking and breastfeeding behaviors, were less likely to breastfeed within 150 minutes of birth, and cried more; from 90 percent to 100 percent of the unmedicated newborns exhibited all six measured breastfeeding behaviors (Ransjö-Arvidson et al. 2001). Numerous childbirth interventions decrease the likelihood of establishing breastfeeding (Forster and McLachlan 2007; Moore, Anderson, and Bergman 2007; Smith 2007), which confers many short- and long-term benefits to babies and mothers (American Academy of Pediatrics 2005; Horta et al. 2007; Ip et al. 2007; Labbok 2001; Labbok, Clark, and Goldman 2004).
- Adults who met diagnostic criteria for drug addiction were about five times as likely as sibling controls to have received three or more doses of opioid and barbiturate drugs within ten hours before birth (Nyberg, Buka, and Lipsitt 2000). When controlling for numerous potential confounders, researchers concluded that the association between pain medications and adult addiction appeared to have a dose-response effect and was not found with drugs administered more than ten hours before birth (Jacobson et al. 1990).
- After adjusting for numerous potential confounders, researchers found that men who committed suicide by violent means were about five times as likely as sibling controls to have experienced multiple trauma at birth (identified as events likely to cause pain to the baby). A sensitive window for effects (“imprinting”) is postulated as the mechanism (Jacobson and Bygdeman 1998).
Tables 3 and 4 clarify the degree to which mothers and babies are experiencing many of these practices as well as the potential for reducing use of many interventions. In view of known and suspected adverse effects of such perinatal exposures, as well as much uncertainty about unintended effects, it would be wise to learn more about these relationships; to studiously avoid maternity interventions that do not offer clear, compelling, and well-supported benefits; and to give priority to effective practices that promote, protect, and support physiologic labor.
Overused Interventions: Examples of Practices to Use Judiciously and with Careful Attention to Informed Consent
Note: references to systematic reviews are in italics.Many maternity practices originally developed to address specific problems have come to be used liberally and even routinely in healthy women (Simpson and Thorman 2005). This overuse exposes many mothers and babies to risk of harm with marginal medical benefit or none at all. This section presents evidence regarding several of these interventions and applies the principles articulated in preceding sections. Greatest attention is given to labor induction, epidural analgesia, and cesarean section, which have all increased considerably in use over the past decade in the United States. These interventions are experienced by a large proportion of childbearing women; are often applied without consideration of alternatives; involve numerous co-interventions to monitor, prevent, or treat side effects; are associated with risk of maternal and newborn harm; and greatly increase costs. As clarified in the following subsections, there are many signs that a notable proportion of use involves casual application with marginal medical benefit or none at all.
It is challenging for childbearing women to recognize that structure and process of care affect outcome; to gain access to full, high-quality information and learn about benefits and harms of common and consequential labor interventions, and of alternative measures; and to clarify their preferences, set goals, and make plans for achieving their goals. Women need opportunities to become informed about these matters and to weigh options well before labor, in addition to consistent, rigorous adherence to informed consent processes during labor. Due to personal values and preferences, women may exercise their right to informed choice and prefer a care path involving greater likelihood of harm than other possible paths. It is inappropriate, however, for clinicians, administrators, and other professionals to recommend, encourage, or give priority to use of care practices with increased risk of harm to mothers and newborns because the path is more convenient, efficient, or lucrative for professional work. Further, it is essential to improve the liability system and enable health professionals to make clinical decisions free of pressure to reduce their risk of legal liability.
Labor Induction
Labor induction is the use of drugs and/or techniques to cause labor to start, as opposed to waiting for labor to begin on its own through a complex interplay of maternal and fetal factors (Liao, Buhimschi, and Norwitz 2005). Many putative indications are used to justify labor induction, and many agents and techniques are used to carry it out.
In considering this increasingly used intervention, it is important to distinguish labor that is in fact induced from unsuccessful attempts to bring on labor (as not all attempts cause labor to begin), and to distinguish women’s attempts to self-induce from efforts of health professionals. The national Listening to Mothers II survey sheds light on these facets of labor induction among women who gave birth in U.S. hospitals in 2005 (Declercq et al. 2006). It is an important source of information because birth certificates only include one item about induction, that is, whether labor was actually brought on by any medical intervention. In addition, as referenced earlier, validation studies have found that a large proportion of cases of induced labor is not in fact recorded on birth certificates. Still, the undercounted rate of medically induced labor derived from birth certificates increased by 135 percent from 1990, when it was 9.5 percent, to 2005, when it reached 22.3 percent of all women giving birth (Martin et al. 2007).
Reflecting increasingly casual professional and social attitudes toward intervening in the process of childbirth, 22 percent of Listening to Mothers II participants indicated that they had themselves tried to start their labor. Of these, 21 percent—or 4 percent of all of the mothers—reported actually inducing labor. Leading methods used for trying to bring on labor were walking/exercise (82 percent), sexual intercourse (71 percent), and nipple stimulation (41 percent). The most common reason was fully elective—the desire to end the pregnancy (58 percent of attempts to self-induce), followed by a desire to avoid a medical induction (33 percent), interest in controlling the timing (15 percent), and their provider’s concerns about a large baby (10 percent) (sum of percentages exceeds 100 as the mothers were asked to identify all methods and reasons that applied) (Declercq et al. 2006).
Fully 41 percent of Listening to Mothers II participants reported that a health professional tried to induce their labor, with 84 percent of those—34 percent of all women—reporting that the attempt did in fact start labor. Combining self- and medical induction, 50 percent of all women were exposed to induction agents and/or techniques, and 39 percent of all labors were started by external means without waiting for labor to start on its own (Declercq et al. 2006).
Combining induced labor with cesareans that were carried out before the onset of labor, a majority of mothers (52 percent) experienced elective delivery rather than spontaneous onset of labor (Sakala 2006a), resulting in a social foreshortening of the length of human gestation. The most common gestational age at birth among single babies shifted from forty to thirty-nine weeks between 1992 and 2002 (Davidoff et al. 2006). This shift in the duration of gestation appears to be continuing despite evidence for progressive fetal development of vital organs such as the brain and lungs after thirty-seven completed weeks of gestation (Kinney 2006; Morrison, Rennie, and Milton 1995; Stutchfield, Whitaker, and Russell 2005; Zanardo et al. 2004), the current definition of full term.
We were unable to find any published study or professional statement identifying any absolute indication for inducing labor. Listening to Mothers II survey mothers who experienced medical attempts to induce labor revealed the most common reason(s) for use of this intervention. They reported a caregiver’s concern that the baby was overdue (25 percent of women whose caregivers tried to induce labor), a maternal health problem that called for quick delivery (19 percent), mother’s desire to end the pregnancy (19 percent), and a caregiver’s concern about the size of the baby (17 percent). Less common reasons included concern about infection with ruptured (broken) membranes (9 percent), concern about baby’s health (9 percent), mother’s interest in controlling timing (8 percent), and mother’s interest in giving birth with a specific provider (8 percent) (Declercq et al. 2006). As with self-inductions, many women reported use of this intervention with no expectation of a medical benefit.
Most Listening to Mothers II participants who experienced attempted medical induction were exposed to two or more methods of induction. Synthetic oxytocin was most commonly used (by 80 percent of this group), followed by breaking of membranes (49 percent), sweeping or stripping membranes loose (33 percent), and some form of prostaglandin applied near the cervix (24 percent). Forty-five percent of this group experienced both synthetic oxytocin and rupture of membranes (Declercq et al. 2006).
Wide practice variation in rates of induced labor appears to be unrelated to needs of mothers or babies. For example, an analysis of over thirty-one thousand births in 1998–1999 in sixteen hospitals in upstate New York found that 21 percent of all births were induced, and that there was no apparent reason for 25 percent of those. Rates of induction varied about fourfold across hospitals (from 10 percent to 39 percent) and about sevenfold across providers in hospitals (from 7 percent to 48 percent). There was even greater variation in the proportion that appeared to have no medical rationale across hospitals (from 12 percent to 55 percent) and across providers in hospitals (from 3 percent to 76 percent). The variation could not be explained by risk status of the women or other investigated factors (Glantz 2003).
Eleven percent of all Listening to Mothers II participants reported experiencing pressure from a health professional to undergo labor induction. Those who reported pressure were more likely to experience attempts to induce labor than those who did not report pressure (Declercq et al. 2006).
The large prospective CPM 2000 study of low-risk American women who received care with a focus on enhancing the physiology of childbirth suggests a benchmark labor induction rate that might be achieved for the majority of childbearing women who are at low risk at the end of pregnancy. In this population of 5,418 births in 2000, 9.6 percent experienced attempts to induce labor (Johnson and Daviss 2005), in contrast to the equivalent figure of 50 percent of Listening to Mothers II participants.
What do we lose when we forgo spontaneous labor and instead expose mothers and babies to induction agents and techniques and shorter gestation without sound evidence that the health benefits outweigh harms? Several examples and other possible impacts suggest that this question warrants further research and assessment in systematic reviews to improve knowledge about the impact of forgoing labor, foreshortening gestational age, and exposing mothers and babies to induction agents and techniques. Studies point to the importance of improving knowledge of potential effects, including the following:
- Synthetic oxytocin, which is widely used to induce labor, interferes with the functioning of a woman’s own oxytocin receptors (Phaneuf et al. 2000). This may adversely affect other important functions of a mother’s natural oxytocin release, such as reducing postpartum hemorrhage and contributing to attachment and the establishment of breastfeeding (Buckley 2004b).
- Prenatal methods for estimating gestational age are imprecise and have a margin of error of up to ± two weeks (Engle 2006), so elective labor induction will in many cases lead to delivery at an earlier gestational age than intended.
- Evolving understanding of normal fetal brain development has identified major changes continuing through forty-one weeks of gestation; for example, over one-third of brain volume increase takes place in the final six to eight weeks, and a fivefold increase in white matter volume occurs from thirty-five to forty-one weeks gestation. There is uncertainty about how extrauterine brain development compares to intrauterine development during similar time periods from conception (Kinney 2006).
- Induction appears to increase the likelihood of cesarean in first-time mothers, when the cervix is not ready for labor and at earlier gestational ages (Kaufman, Bailit, and Grobman 2002).
In the national Listening to Mothers II survey, childbearing women in the United States expressed a strong desire to know about all or most potential complications of labor induction before deciding to have one, yet their demonstrated knowledge of labor induction complications was quite poor, whether they had one or not (Declercq et al. 2006). This identifies the need for improved education and informed consent processes.
Induction for convenience or for a medical indication that is not supported by clear evidence may be expected to offer minimal benefit at best. It is important to identify any harm that may be associated with the extensive use of elective induction, and a systematic review of effects of elective induction is being completed. A recent narrative review (Grobman 2007) identified concerns with elective induction, including increased likelihood of the following:
- fetal monitoring
- epidural analgesia
- cesarean section in first-time mothers
- cesarean section when the cervix is not ready for labor
- assisted delivery (vacuum extraction or forceps)
- postpartum hemorrhage and transfusion
- longer intrapartum period and longer postpartum stay
- costs (with increases in multiple cost centers)
Given such concerns, it is important to avoid exposure of mothers and babies and costs to payers of labor induction that lacks a clear medical benefit. In addition to inductions for convenience, a major area for improving practice in the United States is with respect to those that are initiated for a suspected large baby (macrosomia). Best current evidence identifies no benefits for mothers and babies when labor is induced for suspected fetal macrosomia (National Collaborating Centre for Women’s and Children’s Health 2008b; Sanchez-Ramos, Bernstein, and Kaunitz 2002) and limitations of leading methods for estimating fetal size (Chauhan, Grobman, et al. 2005; Coomarasamy et al. 2005; Dudley 2004; Pattinson and Farrell 1997).
Economic analyses find that induction increases costs associated with childbirth. The costs are especially high for first-time as opposed to experienced mothers, when carried out at earlier gestational ages, and when a woman’s cervix does not show signs of readiness for labor. One estimate of the extra cost associated with induction of one hundred thousand first-time mothers with a cervix that did not have clinical signs of readiness for labor at thirty-nine weeks was $91,000,000 (Kaufman, Bailit, and Grobman 2002), an average of $910 per woman. Another analysis reported that induction added an average of 11 percent to the cost of childbirth among low-risk women (Tracy and Tracy 2003). By implementing a program to reduce inappropriate elective labor induction at eleven hospitals, a health care system estimated that the average total maternal and newborn variable cost decreased by $300 (The Commonwealth Fund 2004).
In sum, an evidence-based framework does not support elective nonmedical induction or induction for a medical rationale that is not supported by strong evidence, as these expose mothers and babies to risk without clear health benefit. Such practices are unlikely to be in the best interests of mothers and babies and increase the cost of maternity services. From a clinical perspective, the preferred alternative is “watchful waiting” for the spontaneous onset of labor and readiness to intervene should a clear justification arise. The strategies identified in the section of this report on policy recommendations might be used to address overuse.
Epidural Analgesia
Epidural analgesia, a regional form of pain medication administered into the epidural space of the spinal cord, is the most effective form of pain relief commonly available for use during labor. The rate of use of epidurals during labor has rapidly increased in recent years, and 76 percent of participants in the national Listening to Mothers II survey experienced epidural analgesia or the spinal variant in 2005 (Declercq et al. 2006).
The effectiveness of this method of pain relief comes at a cost. Labor epidurals alter the physiology of labor and increase risk for numerous adverse effects. Undesirable maternal effects include immobility, voiding difficulty, sedation, fever, hypotension, itching, longer length of the pushing phase of labor, and serious perineal tears. Undesirable fetal/newborn risks include rapid fetal heart rate, hyperbilirubinemia, increased workup for sepsis and administration of antibiotics (due to fever in mothers), and poorer performance on newborn assessment scales (Leighton and Halpern 2002; Lieberman and O’Donoghue 2002; Mayberry, Clemmens, and De 2002). The spinal variant of this regional analgesia method is associated with increased likelihood of bradycardia, or abnormally low heart rate, in the fetus (Mardirosoff et al. 2002). Under some conditions—when initiated early in labor or when used with low- as opposed to high-dose synthetic oxytocin—epidural appears to be associated with increased likelihood of cesarean section (Klein 2006; Kotaska, Klein, and Liston 2006).
Numerous co-interventions, which may further alter the course of labor and have their own side effects, are used to monitor, prevent, and treat unintended consequences of the epidural. Continuous electronic fetal monitoring, intravenous infusions, and frequent blood pressure monitoring are standard precautions with epidural analgesia that would otherwise be unnecessary in healthy women. Women with an epidural are also more likely to experience bladder catheterization, synthetic oxytocin, medication for hypotension, vacuum extraction or forceps, and episiotomy. The original and cascading interventions transform normal labor into a technology-intensive experience.
Many laboring women welcome the pain relief of epidural analgesia, but they do not appear to be well-informed about the side effects. Although childbearing women in the United States overwhelmingly want to be informed of complications of epidurals before deciding to have one (Declercq et al. 2006), their demonstrated knowledge of epidural complications in a national survey was poor, whether they used this method or not (Declercq et al. 2002). This identifies the need for improved education and informed consent processes.
Due to costs of purchasing, operating, maintaining, and providing this package of interventions, epidurals substantially increase costs of childbirth. In one analysis, epidural was associated with as much as a 32 percent increase in the cost of care among low-risk first-time mothers and a 36 percent increase in cost among low-risk experienced mothers (Tracy and Tracy 2003).
Both pharmacologic and nonpharmacologic alternatives to epidurals are available. Although systemic opioids and self-administered nitrous oxide gas both provide less complete pain relief than epidural analgesia, most women who used them rated them in a national survey as very or somewhat helpful for pain relief (Declercq et al. 2002). Both methods have less adverse impact on the course of labor and on mothers. Opioids have the established and undesirable residual side effect of sedation, which can result in depressed newborns; when nitrous oxide is discontinued, the effects appear to cease immediately (Bricker and Lavender 2002; Kronberg and Thompson 2005; Rosen 2002). Twenty-two percent of Listening to Mothers II participants used narcotics, and 3 percent used nitrous oxide in 2005 (Declercq et al. 2006).
Many Listening to Mothers II survey participants gave favorable ratings to a broad range of drug-free pain relief methods (Declercq et al. 2006). Some of the most favorably rated, however, were underutilized (due to lack of access and of high-quality information and other reasons):
- tubs—very or somewhat helpful according to 91 percent who used them, but used by just 6 percent
- use of hot or cold objects—very or somewhat helpful: 81 percent, but used by just 6 percent
- showers—very or somewhat helpful: 78 percent, but used by just 4 percent
- birthing balls—very or somewhat helpful: 67 percent, but used by just 7 percent
Systematic reviews of drug-free measures, including hypnosis (Cyna, McAuliffe, and Andrew 2004), immersion in water (Cluett et al. 2002; Simkin and O’Hara 2002), acupuncture (Lee and Ernst 2004), and other methods (Simkin and O’Hara 2002), have found that these measures are helpful for many women, are associated with decreased use of medications, and appear to have excellent safety profiles. In addition, access to a companion who is present exclusively to provide continuous support throughout labor is associated with substantially decreased use of pain medication and increased satisfaction with the childbirth experience in comparison with usual care, and has no known adverse effects (Hodnett et al. 2007; Simkin and O’Hara 2002). A classic study to understand factors that are associated with women’s experience of labor pain found that women’s degree of confidence in their ability to handle labor was most important and had a bigger impact than such matters as their childbirth preparation, fear of pain, cervical dilation, and frequency of uterine contractions (Lowe 1989).
Labor support from a trained doula or other companion and many drug-free techniques and comfort measures appear to enhance normal labor physiology rather than disrupting it, which contributes to optimal outcomes (Simkin and Ancheta 2005). Many measures for increased comfort can be used in combination—for example, labor support and hydrotherapy. In contrast to narcotics and epidurals, all can be readily discontinued with little or no residual effect if inadequate, and others can readily be tried.
The evidence-based framework in this report suggests that optimal outcomes in mothers and babies and best value for payers would result from using safer, less invasive physiology-enhancing methods for comfort and labor pain relief as first-line care for most women, and using more consequential methods if women find that a series of simpler ones have been inadequate. First-line epidural analgesia would be optimal in selective situations, such as the small proportion of women with extreme fear of labor pain. Such a strategy would elevate appreciated and effective but underused pain relief methods and lead to more conservative use of narcotics and epidurals. It would require that women and health professionals clearly understand the pros and cons of the leading pharmacologic and nonpharmacologic methods of pain relief, that women have opportunities well before labor and again during labor to learn about and discuss these matters, that health professionals are educated to provide a range of methods, and that women have access to these methods and are supported in their decisions about pain relief.
Cesarean Section
Delivery by cesarean section is a clearly beneficial and even life-saving procedure for mother and/or baby in selected circumstances. Absolute indications for cesarean section include prolapsed umbilical cord (cord precedes the baby’s head through birth passage), placenta previa (placenta has grown over the opening of the cervix), placental abruption (placenta has separated from uterus before birth of baby), and persistent transverse lie (fetus is fixed in a horizontal position).
The absolute indications for cesarean section apply to a small proportion of births, yet rates of cesarean section are steadily increasing in the United States and many areas of the world (Betrán et al. 2007). Figure 6 illustrates recent U.S. trends for the overall cesarean rate, the first-time or “primary” cesarean rate, and the rate of vaginal birth among women with a previous cesarean (vaginal birth after cesarean, or VBAC, rate). When first measured nationally in 1965, the U.S. cesarean rate was 4.5 percent (Taffel, Placek, and Liss 1987). Since 1996, it has risen steadily from 20.7 percent to the provisional 2006 rate of 31.1 percent, a 50 percent increase (Hamilton, Martin, and Ventura 2007). A new record level has been reached every year in the present century, and the trend is for continued increase. In 2008 an estimated one mother in three is giving birth by cesarean in the United States. This reflects both a steady rise in primary cesareans and a sharp 72 percent decline in vaginal births among women with a past cesarean, from 28 percent in 1996 to 8 percent in 2005 (Martin et al. 2007).
Contrary to recent trends, national Healthy People 2010 objectives call for a substantial decrease in the cesarean rate and an increase in the rate of vaginal birth after cesarean from 2000 to 2010 (U.S. Department of Health and Human Services 2000). Recent analyses substantiate the World Health Organization’s recommendation that optimal national cesarean rates are in the range of 5 percent to 10 percent of all births and that rates above 15 percent are likely to do more harm than good (Althabe and Belizán 2006). Participants in two large prospective studies of American women experienced cesarean rates that were compatible with this recommendation: both low-risk populations experienced cesarean rates of 4 percent and no observed increase in harms through use of care that enhanced physiologic labor (Johnson and Daviss 2005; Rooks et al. 1989).
What is lost with unnecessary deviation from physiologic labor through planned prelabor cesareans or cesareans initiated during labor? Several examples and other possible effects suggest that this question warrants further research and assessment in systematic reviews to strengthen our knowledge about the impact of forgoing labor, deliberately foreshortening gestational age, and/or exposing mothers and babies to cesarean section. Studies point to the importance of improving knowledge of potential effects, including the following:
- When babies do not experience labor, they fail to benefit from physiologic changes that precede spontaneous onset of labor to help clear fluid from their lungs, and from further clearance during the process of labor, which appear to protect against serious breathing problems in newborns with the sudden transition to extrauterine life (Jain and Eaton 2006).
- Following the sterile intrauterine environment, passage through the vagina increases the likelihood that the newborn intestines will be colonized with beneficial bacteria and reduces colonization with harmful bacteria, in comparison with cesarean delivery (Penders et al. 2006); initial colonization influenced by mode of birth endures over time (Bedford Russell and Murch 2006; Grolund et al. 1999) and may help to explain the association of cesarean birth with asthma and allergy (Renz-Polster et al. 2005; Salam et al. 2006).
- As methods of estimating fetal gestational age are imprecise (Engle 2006), planned cesareans may inadvertently lead to iatrogenic prematurity. In Florida, between 1995 and 2003, 50 percent of the increase in the preterm birth rate among single births was associated with increasing numbers of cesarean births. Further examination of the relationship between cesarean birth and late preterm birth (thirty-four to thirty-six weeks of gestation) among births of single babies to Florida women with low documented medical risk revealed that cesarean without labor (suggesting planned cesarean) was associated with a 53 percent increase in the estimated risk of a late preterm birth, while cesarean with labor was not associated with increased risk of late preterm birth (women with a previous cesarean and with fourteen potential risk factors for cesarean were excluded) (Goodman, Sappenfield, and Thompson 2007), which may help explain why the recent increase in preterm birth has been concentrated in the late preterm weeks (Russell et al. 2007).
- In comparison with vaginal or intended vaginal birth, delivery by elective cesarean is consistently associated with increased risk of respiratory morbidity in near-term newborns and full-term newborns (Hansen et al. 2007).
Although many health professionals, journalists, and others have proposed that the rising cesarean rate is largely a consequence of women’s requests for planned cesarean without a medical rationale, surveys of mothers themselves find that this phenomenon is very limited (Declercq et al. 2006; Kingdon, Baker, and Lavender 2006; McCourt et al. 2007). Similarly, increased genuine need for cesarean in the population of childbearing women—associated, for example, with more multiple births and childbearing among older women who are more likely to have chronic medical conditions—appears to play a limited role in recent trends, as the cesarean rate is rising for all classes of women, at all levels of risk, including those with no indicated risk at all. The increase reflects changing professional standards, with growing casual acceptance of cesarean surgery, lowered thresholds for applying traditional indications, and the appearance of new and unsupported justifications such as “baby seems large” (Declercq, Menacker, and MacDorman 2006; Declercq et al. 2006). Consistent with this supply-side interpretation, fully 25 percent of Listening to Mothers II participants who had both primary and repeat cesareans reported having experienced pressure from a health professional to have a cesarean (Declercq et al. 2006). This style of professional practice is efficient and lucrative for professionals and hospitals (Sakala 2006a) and is widely viewed as reducing risk for malpractice claims and suits (Lockwood 2004). There is considerable practice variation in the use of cesarean, and higher rates are associated with inappropriate use in healthy women (see sidebar “Is the Most Resource-Intensive Care the Best Care?”).
As major surgery, cesarean section has potential for great harm when overused. Most comparisons of effects of cesarean and vaginal birth are based on single studies, focus on a small set of outcomes, and fail to bring into view the full range of effects that are relevant to decision making. A systematic review that aimed to identify all known harms that differ in likelihood by mode of birth found a large inventory of differences that strongly favored vaginal birth (Sakala 2006b). A booklet to help women become informed about these matters is based on the review and describes the full range of outcomes (summarized in the following paragraphs) and the added likelihood of experiencing them when having a cesarean or, in several cases, a vaginal birth; it has been endorsed by over thirty organizations and is freely available online (Childbirth Connection 2006).
Short-term harms to mothers that were more likely with cesarean section included
- maternal death
- emergency hysterectomy
- blood clots and stroke
- surgical injury
- longer hospitalization and more likely rehospitalization
- infection
- poor birth experience
- less early contact with babies
- intense and prolonged postpartum pain
- poor overall mental health and self-esteem
- poor overall functioning
Possibly due to postsurgical adhesion formation, cesarean mothers were also more likely to experience the longer-term problems of chronic pelvic pain and bowel obstruction. The review found that cesarean born babies were more likely than vaginally born babies to experience
- respiratory problems
- surgical injuries
- failure to establish breastfeeding
- asthma in childhood and adulthood
The review also identified many adverse effects impacting a woman’s future reproductive life and mothers and babies in future pregnancies, including greater likelihood of
- involuntary infertility
- reduced fertility due to decreased desire to have more children
- cesarean scar ectopic pregnancy
- placenta previa
- placenta accreta
- placental abruption
- uterine rupture
- hemorrhage
- low birthweight
- preterm birth
- stillbirth
- maternal death
The likelihood of many of these conditions was found to increase as the number of previous cesareans increased (Childbirth Connection 2006; Sakala 2006b).
A scarred uterus appears to be less likely to provide a hospitable environment for the developing fetus than an unscarred uterus and may contribute to placental insufficiency (Smith, Pell, and Bobbie 2003). Many women with an interest in vaginal birth after cesarean (VBAC) are unable to find a health provider or hospital willing to support this choice (Declercq et al. 2006), and many are thus forced to accept the risks associated with repeated cesareans. (See sidebar titled “The Evidence about Vaginal Birth after Cesarean (VBAC)” in the following section.)
The review that aimed to identify the full range of harms that differ in likelihood by mode of birth identified several maternal outcomes that favored cesarean section: increased perineal pain, urinary incontinence, and anal incontinence (Sakala 2006b). Research does not yet exist to clarify the degree to which these conditions, which generally are mild and abate in the months after birth, are associated with vaginal birth per se or with the common use of practices that increase the likelihood of injury, such as episiotomy, staff-directed pushing, supine birthing position, and staff-applied abdominal pressure to push babies out (Albers and Borders 2007). A single outcome in babies favored cesarean section, brachial plexus shoulder nerve injury, which is primarily transient and occasionally permanent in limiting use of the affected arm (Childbirth Connection 2006; Sakala 2006b). By one estimate, one permanent brachial plexus injury occurs in ten thousand vaginal births (Chauhan, Rose, et al. 2005).
A series of recent studies has confirmed this broad range of excess risk associated with cesarean section even when conducted under optimal conditions for limiting harm in healthy, low-risk women without medical or obstetric conditions having planned, nonurgent cesareans (Declercq et al. 2007; Hansen et al. 2007; Kolas et al. 2006; Liu et al. 2007; MacDorman et al. 2008; Tracy, Tracy, and Sullivan 2007). In continuing reports of large studies, downstream adverse reproductive effects were more likely in women with a history of cesarean than in women who had vaginal births (Kennare et al. 2007; Silver et al. 2006), and repeat cesareans were associated with significant cumulative abdominal adhesion formation and adverse reproductive effects (Morales, Gordon, and Bates Jr. 2007; Nisenblat et al. 2006).
Although childbearing women in the United States expressed in a national survey a strong desire to know about complications of cesarean section before deciding to have one, their demonstrated knowledge of cesarean complications was quite poor, whether they had one or not (Declercq et al. 2006), pointing to the need for improved education and informed consent processes.
As shown in Figure 4, the average hospital charge for an uncomplicated cesarean is almost twice as high as the average hospital charge for an uncomplicated vaginal birth. The average charge for an uncomplicated cesarean is about seven times the average charge for a physiologic vaginal birth, as carried out in out-of-hospital birth centers across the country. Most expensive of all are hospital charges for cesareans with complications, which averaged $15,960 in 2005. These figures do not include charges for anesthesia services and newborn care services in hospitals and maternity provider services for all births. The substantial cesarean-vaginal differential was also found in a systematic review of individual economic analyses (Henderson et al. 2001). Most studies in the latter were based on charge data from the United States. The differential was also found in a new analysis of payments from a commercial database of privately insured women (Thomson Healthcare 2007). The analyses are limited to the initial hospitalization period and exclude rehospitalization and other subsequent medical costs, as well as indirect costs of recovery from surgery to the woman and her family.
Adjusting the Thomson Healthcare (2007) figures for national health expenditure inflation rates of 6.5 percent in 2005 and 6.7 percent in 2006 (Centers for Medicare and Medicaid Services 2008), and applying them to the 31.1 percent cesarean rate in 2006, we estimated 39 percent of all payments for childbearing women were for women with cesareans. If, by contrast, the U.S. cesarean rate had been 15 percent in 2006, an estimated point at which harms begin to exceed benefits (Althabe and Belizán 2006), just 20 percent of payments for mothers’ care would have been devoted to women with cesareans, with a net reduction in expenses of more than $2.5 billion.
As over nine out of ten births following a previous cesarean are repeat cesareans at this time in the United States, the cost of the initial cesarean is magnified over time. Capital costs to reconfigure facilities for more planned births and surgical births with longer lengths of stay increase overall expense and generate pressure to sustain this level of revenue and style of practice involving high rates of surgery and weekday deliveries.
In addition to the strategies described in the section on policy recommendations, the following strategies are associated with reduced likelihood of cesarean section:
- in clinical settings, multifaceted interventions, including audit and feedback, were effective in reducing cesarean rates (Chaillet and Dumont 2007), for example, a hospital program combining stringent requirements for a second opinion, objective criteria for the most common indications, review of all cesareans, and reporting of rates of individual physicians providing childbirth services (Myers and Gleicher 1988)
- for pregnant women, providing access to and seeking settings and caregivers with conservative practice styles and low overall rates of cesarean section (Kennedy and Shannon 2004; Reime et al. 2004)
- for women in labor, working with caregivers to delay going to the hospital until labor is well established (Jackson et al. 2003; Klein et al. 2004)
- for women in labor, having a companion (such as a doula, friend, or family member) who is not a member of the hospital staff and is present during labor exclusively to provide continuous support (Hodnett et al. 2007)
- for maternity care providers, retaining and applying skills to facilitate vaginal birth, including a broad range of strategies that foster progress and comfort during labor (Simkin 2002; Simkin and Ancheta 2005), manually turning babies that are not in a head-first position (Collaris and Oei 2004; Hofmeyr and Kulier 1996), skillful vaginal breech birth (Hofmeyr and Hannah 2003), skillful vaginal twin birth (Hogle et al. 2003), and vaginal birth after cesarean (Guise, Berlin, et al. 2004; Guise, McDonagh, et al. 2004; Mozurkewich and Hutton 2000; National Collaborating Centre for Women’s and Children’s Health 2004)
- in facilities, avoiding whenever possible interventions that can increase the likelihood of cesarean section, including continuous electronic fetal monitoring (Alfirevic, Devane, and Gyte 2006), labor induction, especially in first-time mothers with an “unfavorable” cervix (Kaufman, Bailit, and Grobman 2002), and early epidural (Klein 2006)
- in facilities, limiting cesarean section to clearly established indications and addressing inappropriate use of unsupported indications, such as “large baby” (Chauhan, Grobman, et al. 2005; Coomarasamy et al. 2005; Pattinson and Farrell 1997; Rouse and Owen 1999), twin birth, preterm birth, and babies that are small for gestational age (Hogle et al. 2003; National Collaborating Centre for Women’s and Children’s Health 2004)
Brief Notes about Some Other Overused Maternity Interventions
Continuous Electronic Fetal Monitoring
It is important during labor to periodically monitor the fetal heart rate as a way to check on the baby’s well-being. Electronic fetal monitoring (EFM) is the predominant means of doing this in the United States. Ninety-four percent of women who experienced labor in U.S. hospitals in 2005 reported using EFM, and among those, 93 percent were monitored either continuously (76 percent) or for most of the time (17 percent) during labor. Just 3 percent were monitored using a handheld device alone (Declercq et al. 2006).
A recently updated systematic review pooling studies that compared continuous EFM with intermittent EFM monitoring found that continuous EFM did not reduce the likelihood of perinatal death or cerebral palsy, but increased the likelihood of cesarean section and vaginal birth assisted with vacuum extraction or forceps. Other adverse effects of continuous EFM were impairment of mobility, increased discomfort, and focus on the machine rather than the woman. The sole advantage documented of continuous EFM was a slight reduction in newborn seizures, with no known long-term impact on babies. The rarity of this event would require 661 women to be continuously monitored to avert one seizure. Similar results were found for lower-risk and higher-risk subgroups (Alfirevic, Devane, and Gyte 2006). Although expected benefits for continuous EFM have been disproven, the practice has become the standard of care. Intermittent monitoring with various devices is more consistent with an evidence-based maternity care framework.
Two systematic reviews have also assessed the impact of a baseline period of fetal monitoring shortly after hospital admission with the machine used for continuous monitoring. The most recent review found no benefit for newborns and increased likelihood of both cesarean section and assisted delivery among low-risk women experiencing such baseline monitoring (Gourounti and Sandall 2007). The earlier review reported a nonsignificant trend toward cesarean and assisted delivery; increased likelihood of use of epidural analgesia, continuous EFM, and fetal blood sampling; and no newborn benefit in randomized controlled trials of low-risk women experiencing the baseline test; and other types of studies were difficult to interpret (Blix et al. 2005). Both reviews concluded that there is no support for using this admission test with low-risk women.
Rupturing Membranes
Breaking the membranes containing the fetus, amniotic fluid, and umbilical cord with a tool similar to a crochet hook (amniotomy) is a common procedure for inducing labor and—after labor has begun—for hastening labor. Forty-seven percent of Listening to Mothers II participants reported that their caregivers had ruptured their membranes after labor had begun (Declercq et al. 2006). A recent systematic review concludes that there is no evidence of shorter labor, increased maternal satisfaction, or improved newborn outcomes with amniotomy after the start of spontaneous labor, whether the labor is progressing well or is prolonged. The researchers found a possible increase in cesarean section with this procedure and identified concerns about adverse effects on the fetal heart rate and the serious problem of umbilical cord prolapse and compression (Smyth, Alldred, and Markham 2007).
Episiotomy
Episiotomy is a cut made to enlarge the vaginal opening just before birth. Although the rate of use has declined in recent years, 25 percent of women with vaginal births continued to experience this intervention in 2005 (Declercq et al. 2006).
A recent systematic review reaffirmed longstanding evidence: the routine or liberal use of this practice does not confer benefits and rather exposes women to risk of harm. Depending on circumstances, the literature reviewed found that routine episiotomy was associated with an increase in the following conditions: perineal injury, need for stitches, experience of pain and tenderness, healing period, likelihood of leaking stool or gas, and pain with intercourse (Hartmann et al. 2005). The review authors recommended that with judicious use, the rate of episiotomy could be below 15 percent of all vaginal births in the United States. Benchmark episiotomy rates of 2 percent or less have recently been reported in large studies of American women with physiologic care (Albers et al. 2005; Johnson and Daviss 2005).
Certain Prenatal Care Practices
The United Kingdom’s National Institute for Health and Clinical Excellence carries out systematic reviews and develops guidelines for clinical practice. A broad, in-depth report that was updated in 2008 and not limited to studies from the United Kingdom concluded that the following practices should not be included in prenatal care, as they have either been disproven or there is inadequate evidence to support their use (National Collaborating Centre for Women’s and Children’s Health 2008a):
- routine iron supplementation
- routine ultrasound after twenty-four weeks
- routine fetal movement counting
- routine chlamydia screening
- routine hepatitis C screening
- routine toxoplasmosis screening
- routine bacterial vaginosis screening
- routine preterm labor screening
- routine ultrasound to estimate fetal size if large baby is suspected
- routine vaginal examination to assess gestational age, predict preterm birth, or estimate a tight passage during birth
Underused Interventions: Examples of Practices to Use Whenever Possible and Appropriate
Note: references to systematic reviews are in italics.This section highlights effective, noninvasive forms of care with modest or no known adverse effects and low plausibility of serious unknown harms. They are suitable for routine use. It is reasonable to anticipate that greater fidelity in providing these forms of care to childbearing women and newborns would lead to considerable improvement in outcomes. The inventory here is not meant to be exhaustive but rather illustrative of the broad range of generally safe, effective interventions that are underused, could offer benefits to a large segment of the childbearing population, and should be more widely available. In selecting these examples, we have also given preference to measures that can prevent problems (primary prevention) and measures that can help resolve problems (secondary prevention).
Midwives and Family Physicians
In the United States, midwives are the lead maternity caregivers for 8 percent to 9 percent of women during pregnancy and childbirth (Declercq et al. 2006). Of the three national midwifery credentials, certified nurse-midwives (CNMs) are regulated in all states, certified midwives (CMs) are regulated in several states, and certified professional midwives (CPMs) are regulated in about one-half of the states, with efforts under way to develop legislation in the remaining states.
Several systematic reviews have summarized the evidence for midwifery care relative to physician-led or shared care. A meta-analysis of fifteen studies of care by CNMs in U.S. settings found that when differences in process and outcome were identified, they favored CNMs with the exception of increased likelihood of spontaneous perineal tears, primarily smaller first-degree tears, compatible with considerably reduced rates of episiotomy (a second-degree incision) in CNM groups. Other pooled differences in studies that controlled for risk status of mothers included less use of analgesia, anesthesia, intravenous fluids, electronic fetal monitoring, artificially ruptured membranes, and forceps; greater likelihood of spontaneous vaginal birth; and reduced low birthweight in midwifery groups (Brown and Grimes 1995). A systematic review comparing midwifery care in freestanding birth centers to obstetrician-led care in hospitals found that differences favored the midwifery groups, including reduced likelihood of episiotomy and cesarean section (Walsh and Downe 2004). Another systematic review compared midwifery-led care from prenatal through postpartum periods in a diversity of delivery settings with usual care in the locality of the study. Differences favored women who received midwifery care, who were less likely to experience labor induction, labor augmentation, electronic fetal monitoring, pain medications, assisted vaginal birth, and episiotomy, and were more likely to be satisfied with all phases of their care (Waldenström and Turnbull 1998). A Cochrane review comparing midwifery-led care to other models will be published in 2008 a
