| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
ORIGINAL RESEARCH |
From the Department of Obstetrics and Gynecology, The University of Illinois at Chicago, Chicago, Illinois; and the Department of Obstetrics and Gynecology, Chinle Indian Health Services Hospital, Chinle, Arizona.
Address reprint requests to: Sarah J. Kilpatrick, MD, PhD, University of Illinois, Department Obstetrics and Gynecology, 820 S. Wood Street, Chicago, IL 60612; E-mail: sarahk{at}uic.edu.
 |
ABSTRACT |
|---|
TOP
 ABSTRACT MATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
METHODS: All pregnancy-related deaths were reviewed for cause, potential preventability, and identified preventability factors for 1998–1999 at a Zambian hospital and for 1992–2000 at an American hospital network. Preventability factors were categorized as system, provider, or patient. The maternal mortality ratio (MMR) was determined for each hospital. Causes of death, rates of preventability, and preventability factors were compared.
RESULTS: There were 108 and 33 deaths making the MMRs, 1540 and 20.4 per 100,000 live births, at the Zambian and American hospitals, respectively. Causes of death were significantly different between hospitals (P < .001). Infection, the leading cause of death in the Zambian hospital, was associated with over half of direct and indirect deaths. Hemorrhage was the leading cause of direct deaths (28%) in the American hospitals, whereas cardiac and intracerebral events were associated with 42% each of indirect deaths. Eighty-two percent of deaths were deemed preventable at the Zambian hospital compared with 42% at the American hospitals (P < .001). In 73% of the Zambian preventable deaths, system factors were identified as likely contributing factors, whereas provider factors were so identified in 86% of the preventable American deaths (P < .001).
CONCLUSION: The MMRs in each hospital were higher than their corresponding national MMRs and rates of likely preventable deaths were unacceptably high. Attention, education, and intervention must be focused on system and provider factors to reduce worldwide maternal mortality.
The maternal mortality ratio (MMR), defined as maternal deaths per 100,000 live births, continues to be the health indicator that shows the largest discrepancy between the developed and less-developed world. In the United States, the MMR is 7.5 deaths per 100,000 live births, more than 100 times less than some countries in sub-Saharan Africa.1,2 The MMR for sub-Saharan Africa is 980 and for Zambia is estimated to be 649.3,4 Despite the worldwide attention drawn to the problem through the Safe Motherhood Initiative launched in Nairobi in 1987 and the Centers for Disease Control and Prevention’s (CDC’s) Healthy People 2000 goal to reduce the MMR to 3.3, the MMR has not decreased over the past 20 years in the United States or in Africa.3,5 Each year 585,000 women continue to die from complications of pregnancy and childbirth worldwide, and 37% of these are from Africa.3 Although the vast majority of maternal deaths occur in the developing world, until factors contributing to maternal death are better understood, it is unlikely that a significant reduction in maternal death will occur in either industrialized or developing countries.
Recent data suggest that national MMRs, obtained from vital statistics reports, underestimate the true rate of maternal deaths, in some estimates by as much as 30–60%.6–9 Improved ascertainment of maternal death is needed not only to identify the true magnitude of the problem, but also to systematically identify factors contributing to the preventability of these deaths. In addition, studies in both the developed and developing world suggested that 37–90% of all maternal deaths might be preventable.10–13 Detailed analyses of maternal deaths must be done at the local or referral hospital to better delineate the cause and potential preventability of these deaths.
The purposes of this study were to compare the causes and potential preventability of maternal deaths in two disparate hospital settings: a tertiary care center in an American urban area and a level two referral hospital for the Central Province of Zambia. To do this we performed detailed analyses of maternal deaths within the two settings.
|
MATERIALS AND METHODS |
|---|
|
TOP
ABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Kabwe General Hospital (KGH) is a 400-bed level-two referral hospital for the Central Province of Zambia and serves as a referral site for Kabwe, Chibombo, Kapiri Mposhi, Mkushi, and Serenje districts. All maternal deaths occurring at KGH in 1998 and 1999 were reviewed by one of the authors (KEC) who was a staff obstetrician-gynecologist at KGH at that time. Since September 1998, a maternal death form was routinely filled out on all maternal deaths. Information collected included date, time, and origin of referral, hospital course, as well as time and analysis of death. Before September 1998, deaths were identified through the death certificate books kept on each ward and by scanning nurses’ report books compiled daily. Every attempt was made to be as thorough as possible in determining the cause of death, but it was not always possible to confirm some suspicions, especially in presumed human immunodeficiency virus (HIV)-positive patients, as testing was not available except for blood donors. A patient was diagnosed with acquired immune deficiency syndrome (AIDS) if she had an AIDS-defining illness or if she met WHO criteria.
The University of Illinois at Chicago’s (UIC) Perinatal Network consists of ten hospitals and is one of six perinatal networks in the Chicago area. Between the years 1992 and 2000 all maternal deaths within UIC’s perinatal network were identified. A longer period was studied in the UIC network because the number of maternal deaths was too small for meaningful analysis with a shorter period. Cases were reviewed at the time of death by the quality assurance committee at the hospital of death and re-reviewed by a maternal mortality peer review committee at UIC consisting of representatives from the perinatal outreach division, an epidemiologist-statistician, and a perinatologist. Using a modification of the CDC’s National Pregnancy Mortality Surveillance System, all available data were analyzed by this committee for each case.14,15 These data included hospital records, death certificates, autopsy reports when available, and quality assurance committee reviews. After review, a consensus decision was made as to the primary cause of death. Each maternal death was then further classified as pregnancy related (direct and indirect) or non–pregnancy related. Non–pregnancy-related deaths were not included in this report. Institutional review board approval was obtained at UIC.
For both UIC and KGH, once each death was reviewed and classified, a decision was made as to whether the death was potentially preventable and whether alterations in patient, provider, or system factors could have possibly prevented the mortality. Specifically, at UIC, the maternal mortality peer review committee, and at KGH, the physician reviewer, determined if alterations in patient behavior, provider factors, or system factors could have likely prevented the death. Provider factors were judged within the framework of standard of care for each institution. System factors were factors not related to provider decisions. Examples of both are listed in Results. In addition, maternal demographic and pregnancy outcome data were collected. A death was listed as postpartum if the patient was in the third trimester (
25 weeks) and died after delivery and after leaving the delivery room or operating room. Causes and preventability of deaths were compared between the centers and Statistical Package for Social Sciences was used for analyses using 
2 and Fisher exact test, where appropriate.
|
RESULTS |
|---|
|
TOP
ABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
In the UIC network, the maternal age distribution of the deaths was as follows: 64% of women were aged 20–34 years, 24% were 35 and older, and 12% were less than 20 years. Sixty-seven percent of the women were black, 18% were white, 6% were Hispanic, and 9% were other/unknown. The age distribution within KGH was not different than that for UIC: 63% of the women were 20–34, 13% were 35 and older, and 24% were less than 20 (P = .15). All women were of African descent, and the majority of them (58%) came from within the district (either directly from home or from a district health center), whereas the remaining 42% were referred from outside the district. Thirty-eight percent of patients died within 24 hours of arrival, reflecting their critical condition on admission.
As seen in Table 1
, the fetal and neonatal outcome was significantly different between the two hospitals (P = .002). The largest differences were seen in the frequencies of live births and undelivered pregnancies. Of the undelivered women at KGH, two of the direct deaths were due to hypertension, and three were due to hemorrhage. Malaria and anemia accounted for 22 of 27 indirect deaths. The causes of death for the four undelivered women at UIC were hypertension, pulmonary embolus, hypoxia with sickle cell disease, and intracranial hemorrhage. The gestational age at death was not different between the hospitals (Table 2).
|
|
and 4, respectively. At UIC, 61% of the deaths were directly related to pregnancy (20 of 33), and 39% (13 of 33) were indirect deaths. The distribution of deaths was not significantly different than that at KGH, with direct deaths accounting for 42% (45 of 108) versus 58% (63 of 108) for indirect deaths (P = .056). Although hemorrhage was common at both sites, ruptured uterus accounted for the majority of hemorrhages at KGH, whereas postpartum hemorrhage (PPH) accounted for the majority of hemorrhage deaths at UIC. The largest difference between the two sites in direct deaths was accounted for by infection. It was the least common cause of direct death at UIC but the most common cause at KGH. Indirect causes of maternal death differed greatly by site, where the majority of deaths were due to aneurysm/intracranial hemorrhage or cardiac disease at UIC, whereas at KGH the majority were due to infection (Table 4). At KGH, malaria and AIDS were the most common infections.
|
|
Provider factors played a significant role in the majority (86%; 12 of 14) of preventable deaths at UIC. These provider factors included delayed diagnosis and treatment, inadequate treatment, inappropriate discharge from the hospital, delay in taking patient to the operating room, delayed consultation, and lack of complete knowledge of the patient’s past medical history. Patient factors in the UIC network also played an important role in contributing to maternal death in two cases (14%). In both of these cases the patients refused to accept blood transfusion and died as a result. Although system issues were never thought to be the primary factor in preventable deaths within the UIC network, in 5 (36%) cases system issues were identified that may have improved patient care. For example, in one case having the wrong name on the on-call board delayed reaching the appropriate provider, and in another case an incorrect chest x-ray result was communicated to the managing physician.
The pattern of preventability was significantly different at KGH where system, rather than provider, factors played the primary role in the majority of preventable deaths (P < .001). System factors were thought accountable for 73% (64 of 88) of preventable deaths at KGH. These factors included chronic shortages of supplies and drugs such as blood products, antibiotics, sutures, and intravenous fluids. Additional system factors included late referral due to poor referral system, lack of functioning ambulance or any transport mechanism, no telephones or radios to alert hospitals when help was needed, and inadequate staffing and equipment in the operating room, including lack of anesthetist and equipment. Similar to UIC, patient factors accounted for 17% (15 of 88) of preventable deaths including refusal of blood due to religious beliefs, no prenatal care, home delivery, use of local remedies to induce labor, and refusal to go to the hospital. Interestingly, provider factors including delay in diagnosis or treatment played a primary role in only 10% (9 of 88) of preventable deaths at KGH. At KGH, several patients had more than one avoidable factor.
|
DISCUSSION |
|---|
|
TOP
ABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Birth outcomes differed greatly between the two centers. Although two-thirds of women within the UIC network had a live birth, less than one-third of women in Kabwe did so. It is likely that the decreased live births and high undelivered rates reflect system issues at KGH. Specifically, the inability to intervene quickly for fetal or maternal distress may allow more fetal and maternal deaths before delivery is possible.
The pattern of causes of maternal death, including cardiac disease as a major cause of maternal mortality, within UIC is similar to other detailed reports of US maternal deaths.8,20 In contrast, at KGH, the leading cause of maternal death was infection. Malaria was the single most commonly identified infectious agent in the indirect deaths, but HIV was likely in all deaths categorized as chronic illness although it cannot be proved because of lack of diagnostic testing. None of the patients who died from malaria received malaria prophylaxis. Testing and treatment for anemia, HIV counseling and testing, and national malaria prophylaxis should be a priority at all health centers providing maternal health services to reduce maternal deaths in Zambia. Interestingly, this pattern of infection as the leading cause of maternal death is similar to that seen in the United States before 1930, and it was not until major system changes, including public health measures, were instituted that infection as a cause of maternal death radically declined.9,21
One of the most interesting findings centered on potential preventability. Within the UIC network, 42% of all deaths were deemed potentially preventable, while at KGH 82% of deaths may have been preventable. At UIC, in 86% of preventable deaths, a provider factor was cited as the primary agent possibly contributing to the death. Although there was no appreciable pattern of cause of death in the 14 preventable deaths, 12 had provider factors related to not appreciating severity of symptoms and/or delayed diagnosis. It is clear that early recognition and timely management of these disease entities, which commonly complicate pregnancy and the puerperium, are essential in decreasing maternal deaths and should be the focus of education and training to aid in reducing maternal mortality in the United States.
A very different profile of preventability was seen at KGH, where system factors played a dominant role. Issues such as a poor referral system, lack of transport, and the subsequent critical condition of the patients on arrival were major problems at KGH but had little effect on mortality within UIC where a regionalized perinatal system with an established transport process has long been available. Even when women arrived at KGH the care was felt to be substandard due to systems factors including shortages of supplies, delays in surgical treatment, and lack of anesthetist availability. An internal audit done in April 1999 (unpublished data by KEC) showed the average delay for performing a cesarean delivery at KGH was 2 hours 30 minutes, and during evenings and weekends, the delay was over 3 hours. Similar high rates of preventability and types of preventability factors have been found in other countries, including data identifying system issues such as lack of transportation, delayed access to care, and suboptimal clinic and hospital management as contributing to maternal deaths.11–13
Although the MMR was significantly higher and significantly more deaths were deemed preventable at KGH than UIC, the MMR and preventability rates at both sites were unacceptably high and require further investigation. To make serious efforts to reduce maternal deaths worldwide, preventability issues must be honestly evaluated by regionally customized detailed analyses of maternal deaths. Within a US tertiary referral network provider, factors may have played a significant role in maternal deaths. Efforts should be directed at more education of primary and tertiary providers on recognition of severe morbidity in obstetric patients. At a Zambian referral hospital, the majority of deaths may have been preventable with system factors identified as playing a significant role. Attention needs to be applied to the development of infection prevention programs, efficiently functioning referral systems, and well-equipped health facilities. Our data should draw attention to the need for more research on potential preventability of maternal deaths worldwide.
|
Footnotes |
|---|
Received January 8, 2002. Received in revised form March 4, 2002. Accepted March 21, 2002.
|
REFERENCES |
|---|
|
TOP
ABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
2. United Nations. The world’s women 1970–1990. The trends and statistics. New York: United Nations, 1991.
3. World Health Organization. Revised 1990 estimates of maternal mortality. A new approach by WHO and UNICEF. Geneva: WHO, 1996.
4. Central Statistical Office (Zambia) and Ministry of Health and Macro International Inc. Zambian Demographic and Health Survey of 1996. Calverton, Maryland, 1997: 133–4.
5. National Center for Health Statistics 1992. Healthy people 2000 review. US Department of Health and Human Services. Maryland: Public Health Service, 1993.
6. Atrash, HK, Alexander S, Berg CJ. Maternal mortality in developed countries: Not just a concern of the past. Obstet Gynecol 1995;86:700–5.[Abstract]
7. Berg CJ, Atrash HK, Koonin LM, Tucker M. Pregnancy-related mortality in the United States, 1987–1990. Obstet Gynecol 1996;88:161–7.[Abstract]
8. Horon I, Cheng D. Enhanced surveillance for pregnancy-associated mortality—Maryland, 1993–1998. JAMA 2001; 28:1455–9.
9. Gosh MK. Maternal mortality. A global perspective. J Reprod Med 2001;46:427–33.[Medline]
10. Panting-Kemp A, Geller SE, Nguyen T, Simonson L, Nuwayhid B, Castro L. Maternal deaths in an urban perinatal network, 1992–1998. Am J Obstet Gynecol 2000; 183:1207–12.[Medline]
11. Nagaya K, Fetters M, Ishikawa M, Kubo T, Koyanagi T, Saito Y, et al. Causes of maternal mortality in Japan. JAMA 2000;283:2661–7.
12. Wessel H, Rietmaker P, Dupret A, Rocha E, Onattingus S, Bergstrom S. Deaths among women of reproductive age in Cape Verde: Causes and avoidability. Acta Obstet Gynecol Scand 1999;78:225–32.[Medline]
13. Fawcus S, Mbizvo M, Lindmark G, Nystrom L. A community-based investigation of avoidable factors for maternal mortality in Zimbabwe. Stud Fam Plann 1996;27: 319–27.[Medline]
14. Atrash HK, Koonin LM, Lawson HW, Franks AL, Smith JC. Maternal mortality in the United States, 1979–1986. Obstet Gynecol 1990;76:1055–60.
15. Ellerbrock TV, Atrash HK, Hogue CJR, Smith JC. Pregnancy mortality surveillance: A new initiative. Contemp Obstet Gynecol 1988;31:23–34.
16. Vangeenderhuysen C, Banos JP, Mahaman T. Preventable maternal mortality in an urban area in Niamey (Niger). Sante 1995;5:49–54.[Medline]
17. Orach CG. Maternal mortality estimated using the Sisterhood method in Gulu district, Uganda. Trop Doct 2000; 30:72–4.[Medline]
18. Differences in maternal mortality among black and white women—United States, 1990. MMWR Morb Mortal Wkly Rep 1995;44:6–7; 13–14.
19. State-specific maternal mortality among black and white women—United States, 1987–1996. MMWR Morb Mortal Wkly Rep 1999;48:492–6.[Medline]
20. Koonin LM, MacKey AP, Berg CJ, Atrash HK, Smith JC. Pregnancy-related mortality surveillance—United States, 1987–1990. MMWR Morb Mortal Wkly Rep 1997;46: 17–36.
21. Gibbs RS. Impact of infectious diseases on women’s health: 1776–2026. Obstet Gynecol 2001:97:1019–23.
This article has been cited by other articles:
|
|
 |
|
  C. J. Berg, M. A. Harper, S. M. Atkinson, E. A. Bell, H. L. Brown, M. L. Hage, A. G. Mitra, K. J. Moise Jr, and W. M. Callaghan Preventability of Pregnancy-Related Deaths: Results of a State-Wide Review Obstet. Gynecol., December 1, 2005; 106(6): 1228 - 1234. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Drife Thromboembolism: Reducing maternal death and disability during pregnancy Br. Med. Bull., December 1, 2003; 67(1): 177 - 190. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
