Maternal Mortality

Quality Improvement

Maternal mortality is a tragedy for the woman and her family, as well as for health care providers and society. Reports of trends or causes of maternal mortality invariably receive much attention in the media and on the Internet. Although the maternal mortality rate is now low in resource-rich countries, these deaths reflect much larger numbers of mothers with near misses or severe morbidity. In addition, it is likely that further reductions are possible in many high-resource countries, including the United States. Also, maternal mortality is a summary indicator of maternity care, and combined with infant mortality it is a measure of the quality of a country’s maternal-child health care system. In addition to being a significant public health metric, maternal mortality has become a driver for improving the safety and quality of maternity services at the state and national levels.

In this chapter, maternal mortality is examined through multiple lenses—public health, medical, and quality improvement—but it can also be examined through political, personal, and medicolegal lenses. There are various definitions of maternal mortality, each with advantages and limitations. There have been significant changes in how data are collected and analyzed, with corresponding changes in rates reported. The leading causes of pregnancy-associated deaths in resource-rich countries such as the United States are reviewed here, including a profile of women at greatest risk for maternal death. The concept of preventability has been variously defined and differs by cause. Opportunities to reduce maternal mortality are reviewed, with special focus on anticipation, prevention, and rapid response to obstetric emergencies. The roles of maternal-fetal medicine subspecialists and other obstetrics leaders in promotion of quality and safety are described.

Measurement/Definitions

It is difficult to accurately measure maternal mortality. One problem is the use of varied definitions. The most commonly used definitions are those developed by the World Health Organization (WHO). ^, Maternal death is defined as the “death of a woman while pregnant or within 42 days of termination of pregnancy, irrespective of the duration and site of pregnancy, from any cause related to or aggravated by the pregnancy or its management, but not from accidental or incidental causes.” A pregnancy-related death is the “death of a woman while pregnant or within 42 days of termination of pregnancy, irrespective of the cause of death.” Late maternal death is the “death of a woman from direct or indirect causes, more than 42 days but less than 1 year after termination of pregnancy.” The maternal mortality ratio is the “number of maternal deaths per 100,000 live births,” and the maternal mortality rate is the “ratio of maternal deaths to the woman-years of exposure for women aged 15–49 years.” Although maternal mortality is reported per 100,000 live births, it is a ratio rather than a rate because some of the deaths occur in women with nonviable pregnancies (e.g., ectopic pregnancies, miscarriages, terminations, stillbirths), which are not in the denominator of live births. Because the proper denominator, the total number of pregnant women, is unknowable (there is no system for collecting early pregnancy losses), the countable number, live births, is used as an approximation, which leads to the correct term, a maternal mortality ratio. Most investigators and public health agencies utilize maternal mortality ratios, and occasionally maternal mortality rates, when considering maternal deaths.

Definitions from the Centers for Disease Control and Prevention (CDC) subdivide maternal deaths into pregnancy-related and pregnancy-associated. Pregnancy-related obstetric deaths are those resulting from obstetric complications of the pregnant state (pregnancy, labor, and puerperium) as a consequence of interventions, omissions, or incorrect treatment or from a chain of events resulting from any of these. Pregnancy-related deaths also include deaths from aggravation of an unrelated condition by the physiologic effects of pregnancy. In contrast, pregnancy-associated deaths are those occurring during pregnancy or within a year of the end of pregnancy, regardless of cause. Often, when trying to determine whether a death was pregnancy-related, maternal mortality review committees ask, “If this woman had not been pregnant, would she have died?” As the US maternal population has aged and developed more underlying conditions (e.g., morbid obesity), the number of pregnancy-associated deaths has increased. In practice, the distinction between pregnancy-related and pregnancy-associated is more useful for understanding causes than for clinical care. Deaths of unknown cause are not classified as either pregnancy-related or pregnancy-associated.

Another issue is accurate ascertainment and reporting of maternal deaths. As a rare event, it is necessary to use vital statistics. In the United States this is done through death certificates filed in states and compiled into national data through the National Vital Statistics System. The death certificate has been reclassified using disease codes from the 10th revision of the International Statistical Classification of Diseases and Related Health Problems (ICD-10). Cases with obstetric (or “O”) codes are classified as a maternal death (excepting O96, late obstetric death). This standard approach to data compilation is performed annually by the US National Center for Health Statistics (NCHS), which produces the official maternal death statistics for states and for the nation. This methodology is also used for time trends and for international comparisons.

It may be difficult to ascertain whether an individual was recently pregnant. For example, a death 5 weeks after delivery due to cardiomyopathy may be ascribed to heart disease and not counted as a maternal death. This assuredly leads to underreporting of maternal deaths. In recognition of the limitations of ICD-10 coding, the United States has a complementary system for evaluating maternal deaths, which is based in the Division of Reproductive Health at the CDC. The Pregnancy Mortality Surveillance System was started in 1985 in consultation with the American College of Obstetricians and Gynecologists (ACOG) and state health departments. By linking birth and death certificates, they can identify considerably more deaths than those reported on death certificates alone.

In an additional effort to increase accuracy, the World Health Assembly in the International Classification of Diseases , 10th Revision (ICD-10) recommends that a pregnancy checkbox be included on death certificates. This practice increases ascertainment but may lead to overreporting due to accidental or incorrect marking of the check box. In an analysis of 467 potential pregnancy-related deaths that were identified by either the pregnancy check box or through linkage of birth and death certificates, the checkbox had a sensitivity of 62% and a positive predictive value of only 68%. Moreover, not all countries or even all states in the United States use the checkbox. Numerous other factors contribute to variation in reporting of maternal deaths, including what are considered “incidental causes,” and historically there have been questions as to whether events such as suicide or motor vehicle crashes should be considered a maternal death. More recently, however, accidental drug overdose and suicide have emerged as leading causes of maternal death in many states. ^{, , ,} Also, the standardized maternal death reporting form from the CDC now queries about this cause of death and asks mortality review committees to deliberate whether there were factors indicating that the death was related to the pregnancy. Finally, there are well-known challenges of accurate data collection in low-resource settings wherein many individuals may not intersect with hospitals or health care systems.

Epidemiology

It is estimated that there were 303,000 (95% confidence interval [CI], 291,000–349,000) maternal deaths worldwide in 2015. This is a maternal mortality ratio (MMR) of 216 per 100,000 live births (95% CI, 207–249), which is a relative decrease of 43.9% from the 385 (95% CI, 359–427) noted in 1990. The MMR varies considerably by region. In 2015, it was 12 per 100,000 (95% CI, 11–14) in high-income countries but was as high as 546 per 100,000 (95% CI, 511–652) in sub-Saharan Africa.

The dramatic fall of MMR in the United States during the 20th century has been heralded as a success for both public health programs and obstetrics providers ( Fig. 47.1 ). The decline, from rates approximating 900 per 100,000 births in 1901 to 9 per 100,000 in 1991, in the United States is similar to observed rates of decline in other resource-rich countries. This success has been attributed to many factors: the movement of most births to hospitals, improved hygiene and aseptic technique, common use of prenatal care, screening for preeclampsia, the introduction of blood transfusions and antibiotics, widespread availability of obstetric anesthesia, an increase in training and expertise of obstetrics providers, and an improvement in the overall health of the population. Although all of these may be important, the period of greatest decline was in the 1930s and 1940s, when many hospital-based advances were being introduced. Although the United States in the 1930s was not resource-rich, the advent of state and city maternal mortality review committees focused attention on causes of and solutions for maternal mortality. Developing community consensus that involves collaboration between the public health system, hospitals, and health care providers, together with forming a local maternal mortality review committee (MMRC), should be considered the model when maternal mortality and morbidity are addressed.

More recently, however, the national US metrics for maternal mortality have been rising. In 2014, the MMR in the United States was 23.8 per 100,000 live births. ^, In large part, the increase is due to improved ascertainment. However, there also has been an increase in population risk factors such as maternal age and obesity. Even when increased ascertainment is accounted for, maternal mortality in the United States is substantially higher than that in other resource-rich countries.

State and national MMRs are underestimated when using only death certificate codes compared with analyses using enhanced surveillance techniques, such as linked birth and death certificates. In the United States between 1999 and 2007, a series of changes were made to improve the accuracy of maternal mortality data. ICD-10–coded death certificates had more categories for maternal death than those coded by the Ninth Revision (ICD-9). A new standard death certificate was introduced in 2003 that offered a pregnancy checkbox to identify women who were pregnant at the time of death or had delivered within the prior 42 days (or, in some states, 365 days). Given the small numbers of maternal deaths, modest improvements in case ascertainment would be expected to lead to increased rates. Clinical factors that contributed to the higher risk for maternal mortality in this time interval likely included increases in maternal age and obesity, rises in cesarean delivery rates, and increased rates of underlying medical comorbidities among reproductive-age women. Thus it is difficult to determine the relative contributions of improved data collection and of clinical and patient factors to the increased MMRs ( Fig. 47.2 ).

In 2012, because of uncertainties about data collection (e.g., different states have different rules), the NCHS suspended national reporting of MMR for several years in the hope that these issues could be clarified. Following an extensive review to improve data quality, and after adoption of new coding procedures based on this review, NCHS released an MMR for 2018, which was 17.4 deaths per 100,000 live births. Recently a pan-European surveillance program for maternal mortality and severe maternal morbidity also concluded that current population-based data reports were beset by problems, with under-ascertainment of mortalities, poor data collection systems, and lack of acceptance of the uniform definitions of severe maternal morbidity. The Euro-Peristat Scientific Committee strongly recommended a return to confidential MMRCs, which would allow for detailed objective review of maternal deaths with the opportunity to guide public health interventions to eliminate preventable deaths.

Recently, several investigators assessed the contribution of improved data collection to the increase in the US MMR. MacDorman and colleagues evaluated vital statistics mortality data from all US states in relation to the type of reporting used (e.g., type of pregnancy checkbox questions). They showed that most of the reported increase in MMR was due to increased ascertainment. Joseph and coworkers also used vital statistics to numerically underscore this point. According to their data, the MMR increased from 7.55 in 1993 to 21.5 in 2014 (relative risk [RR] comparing 2014 with 1993 = 2.84; 95% CI, 2.49–3.24). However, the increased risk was mainly the result of increases in death associated with two ICD-10 codes for renal disease and “other maternal diseases classifiable elsewhere.” If these deaths were excluded, the increase in MMR disappeared (RR = 1.09; 95% CI, 0.90–1.25). Furthermore, regression analysis that adjusted for improvements in surveillance also abolished the temporal increase in the US MMR (adjusted RR = 1.06; 95% CI, 0.90–1.25). Most of the increase is related to the pregnancy checkbox. Between 1998–2002 and 2008–12, the MMR increased from 9.0 to 22.4 in states using the checkbox compared to 8.6 to 9.9 in states that did not adopt the checkbox. These investigators determined that an estimated 90% of the observed change in MMR was due to change in identification rather than a true change in rates. They also noted that up to 28.8% of the change may have been due to maternal death misclassification. A recent report utilizing data from the CDC Pregnancy Mortality Surveillance System demonstrated a relatively stable US MMR, increasing from 16 per 100,000 live births in 2006–10 to 17 per 100,000 in 2011–13, with an increasing proportion of deaths related to cardiovascular and other medical conditions.

Regardless of the causes, the US MMR is unacceptably high. In the United Kingdom, an excellent system for confidential maternal mortality reviews (the Confidential Enquiry into Maternal Deaths ) results in nearly 100% ascertainment and has produced triennial reports since 1952. The latest reported MMR in the United Kingdom (covering the period 2012–14) was 8.5 per 100,000, considerably lower than that seen in the United States for the same time period (and the United States does not have complete ascertainment). The most recent estimates for 2015 from the WHO were 9 per 100,000 in the United Kingdom compared to 14 per 100,000 in the United States.

Demographics

Notably, MMRs in Black individuals ( Fig. 47.3 ) are nearly four times higher than those among White, Hispanic, or Asian individuals, and these ratios have increased as overall rates have risen. The magnitude of disparity between Black individuals and other racial and ethnic groups is higher in the MMR than in any other health indicator. Although the high MMR in Black individuals is clearly multifactorial, no effective reduction strategies have been introduced (see below). Similar data are noted in the United Kingdom; the most recent data noted an RR of 4.19 (95% CI, 2.69–6.35) for maternal mortality among Black women compared to White women. This rise in overall MMR, and in the MMR among Black women in particular, has sparked national debate, with attention focused on the quality of maternity care, changes in risk factors among pregnant women, and the improved ability to collect accurate maternal mortality data in recent years.

Most US reviews of maternal mortality note that women who die in childbirth share several key demographic features. On average, they are more disadvantaged, older, and heavier; they have less engagement in prenatal care; and they are much more often Black than those who do not die in childbirth. Although these characteristics serve to identify the highest risk populations, they should not divert attention from the occurrence of maternal deaths in all social strata or among women in their teens, women who are slender, women who have complete prenatal care, and women of any race. For example, the MMR for women in their 20s was 12.4 per 100,000 live births, but for women between 35 and 39 years old it was doubled (24.1), and for women 40 years old or more, it was 54.9. In the UK Confidential Enquiry, similar trends (but lower overall rates) were seen: MMRs of 5.95 for 20–24 years of age, 13.62 for 30–35 years, and 17.88 for greater than 40 years.

In a review of all maternal deaths in California using a linked data set and medical record abstraction, women older than 35 accounted for 17% of the births but fully 37% of the pregnancy-related mortalities. The story for obesity is similar. In the United Kingdom, 27% of maternal deaths were in women with a body mass index greater than 30 (i.e., obese), and in California, 22.9% of the pregnancy-related deaths were in women with a body mass index greater than 30 (as compared to 15.7% in the general obstetric population.) Advancing maternal age and obesity are associated with increased rates of underlying medical problems, especially hypertension, diabetes, and cardiac disease. Formal studies of the increased risk for death with advancing maternal age are lacking, but preexisting disease is likely to be the main driver. In addition, there are concerns about overreporting deaths in this age group in the United States, as previously described.

Role of State Maternal Mortality Review Committees

State-level MMRCs can provide detailed data on maternal deaths through case reviews of autopsy reports, medical records, police reports (if applicable), birth certificates, and death certificates. In-depth reviews from MMRCs can help classify maternal deaths appropriately as pregnancy-related or pregnancy-associated and allow for discovery of opportunities for prevention and education. To this end, multidisciplinary engagement is critical, with involvement from public health personnel, obstetricians, maternal-fetal medicine subspecialists, anesthesiologists, psychiatrists, midwives, nurses, pathologists, coroners, and others. Statutory authority to request records for maternal deaths and legal protection for committee members against subpoena related to the reviewed cases is also necessary to establish an effective maternal mortality review at the state level.

As states continue to incorporate MMRCs into public health efforts, standardizing definitions and categorization of maternal deaths will be increasingly important. Through the Building U.S. Capacity to Review and Prevent Maternal Deaths project, the CDC Foundation created the Maternal Mortality Review Information Application (MMRIA or “Maria”) to improve standardization of maternal death reporting. The MMRIA data collection tool leads committees through six key decisions ( Box 47.1 ). Committees first determine whether the death was pregnancy-related, pregnancy-associated, neither, or unable to determine. The MMRC can then determine the most likely cause of death, which is classified with CDC maternal mortality codes according to cause. If possible, the committee determines whether the death was preventable and whether there was a good chance, some chance, no chance, or unable to determine degree of preventability. With this structured form, MMRCs can also evaluate whether contributing factors to the maternal death were found at the patient/family, community, provider, facility, or system level. General recommendations and the anticipated impact of those recommendations if implemented can then be made by the MMRC based on detailed case review. Codifying and standardizing reporting at the state level will ultimately allow for collaboration and identification of opportunities for prevention at the national level. The capacity to aggregate data from multiple states into a national report using the MMRIA form has been demonstrated with aggregation of data from 14 state-level MMRCs by the CDC.