Anesth Analg 2001;93:1-3
© 2001 International Anesthesia Research Society
EDITORIALS
Maternal Mortality in the United States: Where Are We Going and How Will We Get There?
Joy L. Hawkins, MD, and
David J. Birnbach, MD*
University of Colorado Health Science Center, Denver Colorado, and *St. Luke’s-Roosevelt Hospital Center, College of Physicians and Surgeons of Columbia University, New York, New York
Address correspondence to David J. Birnbach, MD, Department of Anesthesiology, St. Luke’s-Roosevelt Hospital Center, College of Physicians and Surgeons of Columbia University, Hospital Center, 1000 Tenth Avenue, New York, NY 10019.
Maternal mortality is considered a basic health indicator that reflects the adequacy of health care (1). Although the maternal mortality rate (MMR) in the United States (US) is approximately 7.5 per 100,000, most studies suggest that the actual number of maternal deaths is larger because of the continuing problem of under reporting (2). Unfortunately, the goal of reducing the MMR to 5 per 100,000 suggested by the Surgeon General in 1980 has not been attained. Maternal mortality has decreased over the last half of the 20th century but preventable cases continue to occur. Thus, despite numerous improvements in health care, poor outcome in the parturient remains a major public health concern that follows us into the 21st century. Although the majority of the approximately 600,000 annual maternal deaths takes place in third-world countries, western Europe and the US are not immune. As we enter the new millennium, we should ask why these deaths continue and what can be done individually and as a profession to decrease the incidence?
In this issue, Panchal et al. (3) review 13 yr of maternal mortality in the State of Maryland using a state-maintained database. By analyzing patient demographics and diagnosis and procedure codes for women who died during their admission for childbirth, these authors identified some medical and demographic risks associated with maternal mortality in their state. This is useful information for physicians and hospitals in Maryland and may also provide a model for other states to examine their maternal mortality data and thus help in efforts to develop important preventative intervention strategies.
Panchal et al. (3) found a state delivery mortality ratio (maternal deaths per 100,000 live births) of 16.4, with marked year-to-year variability ranging from 5.9 to 29.6. This compares with the Centers for Disease Control and Prevention (CDC) estimate of a national maternal mortality of 7.5. Other risk factors identified for maternal mortality in their study were African-American race (odds ratio [OR], 5.4), racial category other than African-American or White (OR, 12.2), Cesarean delivery (OR, 5.3), delivering in a "minor teaching hospital" (OR, 3.1), and being transferred from another hospital (OR, 6.2). As has been reported in other series, the MMR for women aged <34 yr was 13.9 as compared with 23.9 with advanced maternal age. The five most common diagnoses associated with maternal mortality were preeclampsia/eclampsia, postpartum hemorrhage, pulmonary complications, cerebrovascular event, and embolism. Anesthesia-related complications, although not on the top of the list, still accounted for more than 5% of the deaths.
The Ninth Revision of the International Classification of Diseases (ICD-9), commonly used in similar studies, defined maternal death as "the death of a woman while pregnant or within 42 days of termination of pregnancy, from any cause related to or aggravated by the pregnancy or its management, but not from accidental or incidental causes." Recently, the tenth revision (ICD-10) has revised that definition to include "the death of a woman while pregnant or within 42 days of termination of pregnancy, irrespective of the cause of death." It was hoped that the introduction of this new definition would allow better classification, especially in cases in which the cause of death in a pregnant patient was uncertain. Unfortunately, multiple definitions increase the confusion and have helped turn this subject into a quagmire. Results from studies using different definitions cannot be compared and incomplete databases provide information that is at best inaccurate and at worst wrong. When evaluating maternal deaths, under-reporting is the rule, rather than the exception, and may reach as much as 75% (4). Even in the UK, which has the best system for collecting this data, it has been suggested that underestimates are approximately 30% (5).
Despite the importance of collecting and analyzing data regarding maternal mortality, in the US this data is only provided to the CDC by individual states on a voluntary basis. The Maternal Mortality Collaborative established voluntary surveillance of maternal deaths in 1983 in the hopes of improving collection and analysis of these cases. This valuable resource provides useful data, but it also has a number of limitations. For example, it has been estimated that 30% of deaths are missed because of miscoding of death certificates, not noting that a woman was pregnant. The CDC rarely has access to medical records, which prevents them from gaining important insights into the details surrounding the maternal death. In addition, the medicolegal climate in the US often makes practitioners reluctant to provide any further information, especially in cases where there has been litigation.
Although the CDC data are not as complete as they could be, it is still of great benefit. However, Panchal et al. (3) convincingly argue that state-maintained databases could provide even more accurate information. These authors speculate that their delivery mortality ratio might be larger than the reported national number as a result of better detection. They were able to search the database by Diagnosis Related Grouping codes for Cesarean delivery and vaginal delivery with the disposition of death. Under Maryland law, all nonfederal hospitals must report information on care delivered during all hospital admissions, so the act of voluntary reporting is removed from the process. Because states have different issues to address, they might wish to implement different preventative strategies. However, as the authors point out, there are still a number of gaps in knowledge even in this type of database. For example, it is still very difficult to get any information on out-of-hospital births. Home and birthing center births are not captured in the MD database, although most births occur in hospitals and a serious complication would almost surely involve transfer into a hospital.
Furthermore, the system used in this study identifies only the deaths that occurred during the hospital admission for delivery. Readmissions for complications related to pregnancy but appearing after delivery (such as infection, embolism, or late-onset preeclampsia/eclampsia) would be missed. The CDC definition of pregnancy-related death is one "caused or contributed by pregnancy, occurring during pregnancy or within one year after the end of pregnancy." This definition captures deaths occurring after long illnesses such as those resulting from peripartum cardiomyopathy or anesthesia-related complications such as those resulting in neurologic injury. Deaths attributable to early pregnancy complications, such as abortion or ectopic pregnancy, would also be missed in the Maryland database.
Specific cases cannot be reviewed because the database is anonymous. There is no way to delve into the circumstances leading to the death and thus make accurate conclusions. For example, Cesarean delivery was a significant risk factor in this study, but the authors were unable to ascertain the actual cause of death. Was the death caused by the process that necessitated the Cesarean delivery (e.g., bleeding) or by the act of doing the surgery (e.g., anesthesia-related)? Similarly, it is impossible to tell what the anesthesia-related complications were, or whether they were related to general or neuraxial anesthesia techniques, or even the mode of delivery. Because no details are available concerning anesthetic management, we are unable to evaluate specific practice patterns or answer controversial questions. For example, if intravascular injection occurred after an epidural, was a test dose used and if so, which one? Were there differences between epidurals and spinals? Was there adequate staffing? A recent study has reported that the increased numbers of preventable maternal deaths in Japan were attributed to one physician functioning as both the obstetrician and anesthesiologist (6). Information regarding a correlation between safety and staffing would be of great benefit to hospitals and individual departments. In addition, anesthesiologists might benefit from learning the details regarding specific pulmonary complications. For example, how often did aspiration occur? It is tantalizing, but also dangerous, to speculate because the conclusions could be totally inaccurate.
What of miscommunication between anesthesiologists, obstetricians, and nurses? This data provided by Panchal et al. (3) does not allow us to know how many deaths occurred because the patient was ill-prepared for either surgery or anesthesia. If a failed intubation occurred, was there a Laryngeal Mask Airway or Combitube® (Kendal Sheridan Catheter Corp., Argyle, NY) immediately available? Was a jet ventilator available? Was the American Society of Anesthesiologists algorithm for failed intubation followed? If a cardiac arrest occurred, was the appropriate resuscitative equipment available? These are important questions that cannot be answered using this database. More important than assessing blame, the specifics of the cases and their management are necessary to make reasonable recommendations for prevention of similar bad outcomes in the future. For example, if a patient with placenta previa dies during an elective fourth repeat Cesarean delivery as a result of uncontrolled bleeding, we would probably not make the assumption that it was unavoidable if we knew that the case was performed on a weekend by junior faculty who did not have blood products or backup available. And finally, databases such as Panchal et al.’s (3) must be viewed with extreme caution because there is no way to track the "near misses"–the cases where a bad outcome was narrowly averted and the patient rescued by an intervention or a change in management.
Few studies have evaluated the role of quality improvement/quality assurance programs in reducing maternal mortality. The Institute of Medicine report stressed the importance of reporting medical errors. However, for that to occur we need a process whereby those problems can be reported without fear of discovery (7). In addition, physicians need to feel that self-reporting of medical mistakes will lead to improved patient care without adverse consequences. To reduce MMR and to prevent the regional, racial, and ethnic variations in maternal mortality that exist, new solutions are necessary. Acknowledging that maternal mortality continues to occur in the US is not adequate. Rather, we must know why these deaths occur and how we can fix the system so that the preventable cases no longer ensue. Accurate knowledge is not yet available, but is necessary if we are to redesign hospital systems, better train personnel, and develop preventative strategies. A new approach to this problem, including the use of an in-depth peer-review process, may help in this undertaking (1).
Great Britain has published the Confidential Enquiries into Maternal Deaths for many years. They include a detailed description of the case and what went wrong. They also provide recommendations and practice guidelines to practitioners for improvement of care. Why can’t the US get as much information about our negative outcomes so that we can all learn how to provide better care for our patients? Although this study by Panchal et al. (3) has limitations, it is an important step that will hopefully stimulate individual anesthesiologists in the US and their states to pursue better maternal mortality data collection and analysis. The loss of a young healthy mother is surely one of life’s greatest tragedies and we therefore should work tirelessly to improve our anesthetic care so that preventable cases become a memory of a bygone century.
References
-
Panting-Kemp A, Geller SE, Nguyen T, et al. Maternal deaths in an urban perinatal network, 1992–1998. Am J Obstet Gynecol 2000; 183: 1207–12.[Medline]
-
Berg CJ, Atrash HK, Koonin LM, Tucker M. Pregnancy-related mortality in the United States, 1987–1990. Obstet Gynecol 1996; 88: 161–7.[Web of Science][Medline]
-
Panchal S, Arria AM, Labhsetwar SA. Maternal mortality during hospital admission for delivery: a retrospective analysis using a state-maintained database. Anesth Analg 2001; 93: 134–41.[Abstract/Free Full Text]
-
Jacob S, Bloebaum L, Shah G, Varner M. Maternal mortality in Utah. Obstet Gynecol 1998; 91: 187–91.[Medline]
-
UK Health Departments. Report on confidential enquiries into maternal deaths in the UK 1991–1993. London: Her Majesty’s Stationery Office, 1996.
-
Nagaya K, Fetters MD, Ishikawa M, et al. Causes of maternal mortality in Japan. JAMA 2000; 283: 2661–7.[Abstract/Free Full Text]
-
Richardson WC, Berwick DM, Bisgard JC. The Institute of Medicine report on medical errors. N Engl J Med 2000; 343: 663–4.[Free Full Text]
Accepted for publication February 27, 2001.
This article has been cited by other articles:
|
|
|
|
 
H. R. Munoz, L. I. Cortinez, M. E. Ibacache, and F. R. Altermatt
Remifentanil Requirements During Propofol Administration to Block the Somatic Response to Skin Incision in Children and Adults
Anesth. Analg.,
January 1, 2007;
104(1):
77 - 80.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
