Interim Report on the Confidential Enquiry into Maternal Deaths in South Africa

March 1998

By the National Committee for the Confidential Enquiry into Maternal Deaths

Contents

Forward

1. Overview
2. Hypertensive disease in pregnancy
3. Non-pregnancy related infections and AIDS
4. Obstetric haemorrhage
5. Early pregnancy loss
6. Pregnancy related sepsis
7. Pre-existing maternal disease
8. Obstetric anaesthesia and analgesia
9. Methodology of data collection and analysis
10. Conclusion

    Appendix 1 - Global Data

    Appendix 2 - Definitions and method of assessment

    References

Forward

Pregnancy and labour are natural processes and few women should suffer severe complications. Unfortunately maternal mortality is a major problem in developing countries including South Africa. The National Department of Health is therefore to be congratulated in establishing a National Committee into the Confidential Enquiries into Maternal Deaths (NCCEMD) and ensuring that maternal deaths are notifiable by law.

One of the objectives of the NCCEMD is to publish reports on maternal deaths at frequent intervals to enable the providers of health care to review their current provision of services and arrangements for audit as well as reassessing local guidelines for the management of common causes of death. The NCCEMD, however, has only begun collecting data on the maternal mortality from August 1997 in a pilot fashion. Nonetheless, the committee feels it is necessary to publish an interim report to show emerging trends and to highlight areas of safe motherhood practice which may be improved in the short term.

Finally, I would like to thank all the members of the NCCEMD, the members of all the Maternal, Women and Child Health directorates and all assessors for going beyond the call of duty in getting maternal mortality into the National Agenda and in helping not only in the collection of data but also in its collation and assessment. I believe it is only when we can work together that we will reduce maternal mortality by half in the next 5 years.

Prof. Jack Moodley,
Chairman: NCCEMD

NCCEMD

Chapter 1 - Overview

Introduction

This interim report incorporates the data from the Provinces who had submitted cases to the NCCEMD by 14/2/98. Several Provinces were involved in piloting the process of notification. Some started in August 1997 and gradually more came on board. For this reason, the interim report is heavily biased by cases from KwaZulu/Natal (KZN) and Gauteng and there is under-reporting deaths from the more rural areas of South Africa (see Appendix 1 - Table 11.1). Some conditions, especially ectopic pregnancies, septic abortions, thrombo-embolism and anaesthetic deaths are also probably being under-reported. However, a clear pattern is emerging and the NCCEMD felt it necessary to publicise the initial findings so that action can be initiated to start solving the problems.

This overview will identify the major issues that have emerged. More detailed reports are found in the subsequent chapters and the global information in the Appendix 1. Appendix 2 deals with the definitions and method of assessment.

Findings

Major causes of death

The big five causes of maternal death are hypertensive conditions in pregnancy (20%), non-pregnancy related infections (18%) mainly acquired immune deficiency syndrome - AIDS, obstetric haemorrhage (14%), early pregnancy losses (12%) mainly septic abortions, and pre-existing maternal diseases (11%) mainly cardiac disease. See Appendix 1 - Table 11.3.

Women at special risk of maternal death

From the initial findings some pregnant women are more at risk than others. They are:

• Women over 35 years of age. They are especially vulnerable to complications of hypertension, obstetric haemorrhage, and heart complications if they have pre-existing valvular heart disease.
• Younger women are becoming victims of AIDS.
• The women with pre-existing valvular heart disease are at special risk of heart failure and subsequent death.
• Women who have had a previous caesarean section are at special risk of severe haemorrhage as a result of rupture of the uterus.
• Women who have had a caesarean section in this pregnancy are at particular risk of developing severe sepsis after the operation.

Health seeking behaviour of pregnant women

A feature in the behaviour of some of the pregnant women was the delay in seeking help. This was especially found in relation to women dying of septic abortions (58%) and in women who died of complications of hypertension (44%). Part of the problem for the women who died of septic abortion was the lack of accessibility to termination of pregnancy services.

Problems at the primary health care level

Detection of women with valvular heart disease and the appropriate management of women with hypertension in pregnancy were the general problems detected at the antenatal clinics.

Problems at the secondary care level

Some clear problems with the emergency management of critically ill women were seen at secondary level hospitals. They were as follows:

• The diagnosis of women with septic abortions, and postpartum sepsis,
• delay in referring women with complications of hypertension, valvular heart disease, and abruptio placentae to tertiary institutions
• a specific lack of protocols and training in the following conditions;

• Complications of hypertension
• Severe haemorrhage
• Abortions
• Anaesthesia for pregnant women
• Management of heart failure due to valvular heart disease
• Recognition and management of sepsis

Problems at all levels of care

Making good legible notes was a problem detected at all levels of care. Poor communication between clinics and hospitals, hospitals and hospitals and health workers and the clients was also found to be a recurrent problem. This was particularly evident in the transfer of patients from one health service to another. The clinical notes of the patient did not accompany the patient to the new health service.

Recommendations

Certain specific problem areas have been identified which have special reference to medical personnel and administrators. They are:

1. Referral hospitals for hypertension. It was clear that in each health region a hospital needs to be identified that is equipped and trained to manage the complications of hypertension. A clear set of referral criteria and referral patterns needs to be put in place in each region.

2. Pneumonia and AIDS. The presence of pneumonia in a women under 25 years of age should immediately alert the medical staff to the possibility of the women being HIV positive and she should be counselled and tested for HIV.

3. Recognition of obstructed labour. Emphasis needs to be placed on the recognition of cephalopelvic disproportion. The use of the partogram, and the implicit management protocols which follow, must be re-emphasised at all levels of care.

4. Management of labour in women with previous caesarean sections. Women with a previous caesarean section must undergo their labours in a secondary hospital and a clear protocol must be in place for the management of these women at the secondary level hospitals.

5. Availability of blood. A look needs to be taken at the availability of emergency blood at institutions providing delivery services.

6. Termination of Pregnancy services. The patchy implementation of the Choice of Termination of Pregnancy Act of 1996 needs to be rectified. The public needs to know that termination of pregnancy is available and specifically at which sites it is available.

7. Multidisciplinary care. The value of multidisciplinary care needs to be re-emphasied in the health services and at the training institutions. This is especially important in relation to women with pre-existing heart disease.

8. Anaesthetic services. More attention needs to be placed on the training of doctors to provide safe anaesthesia for pregnant women.

9. Prophylactic antibiotics. It has been clearly shown that the use of prophylactic antibiotics prior to caesarean sections decreases the post-operative morbidity. This should be standard policy at all places performing caesarean sections.

10. Family planning service. The women over 35 years of age are at special risk of dying during pregnancy. The attention of the family planning services needs to be focused on this issue and the clients counselled accordingly.

A more detailed analysis of the major problems follows in Chapters 2-8.

The next report will be more comprehensive and accurate. A more detailed analysis of the major problems will be provided and specific areas in which more research is required will be identified.

The NCCEMD will start getting the initial important messages out to the public and to the medical profession. To enhance public awareness of the Confidential Enquiry and to ensure better reporting of deaths a pamphlet will be written and distributed. This will coincide with the Safe Motherhood Initiative month. The messages to the health professionals will be distributed to the Provincial MCWH Units where they will organize workshops and symposiums in the Regions and Districts to get the messages across.

Chapter 2 - Hypertensive Disorders in Pregnancy

Introduction

Hypertension in pregnancy accounts for 10% of all maternal deaths world-wide (WHO statistics:- approximately 50,000 deaths annually). In this interim report of the NCCEMD of 133 maternal deaths, 27 (20%) women died from causes directly due to hypertensive disorders of pregnancy, and a further 4 died in cases where abruptio placentae and hypertension occurred simultaneously.

Of the 27 deaths, 14 women died from eclampsia, 10 from proteinuric hypertension and 3 from HELLP syndrome (Table 2.1). Intracerebral haemorrhage was the most likely cause of death in 7 cases of eclampsia.

Demographic Data

Table 2.2 shows the demographic data. Eight women were 35 years and greater and 6 women had a parity of 4 and more.

Table 2.3 shows the distribution of maternal deaths reported from the various provinces. Most of the deaths were reported by Gauteng (8/42 - 19%) and KwaZulu-Natal (11/43-26%).

Avoidable Factors, Missed Opportunities and Sub-standard Care

In 43% of the maternal deaths due to complications of hypertension, there were avoidable factors, missed opportunities or sub-standard care (Table 2.4). They were distributed as follows:

• Patient orientated factors: These were directly associated with poor utilisation of Antenatal Care Services. Seven women did not attend for antenatal care at all during their pregnancy and 5 had infrequent visits.

• Administrative Problems: Problems related to administration were mainly lack of trained staff and communicating problems between health centres and these contributed to 9 deaths.

• Antenatal Care: Reviewing care given during the antenatal period, it was discovered that, identification of the problems was either not done or incorrectly done resulting in inappropriate management. Incomplete assessment and continuous monitoring without intervention also contributed to 4 deaths. Delay in referring cases to the referral institutions was a problem which contributed to 3 deaths.

• Management Plans and Protocols: Inappropriate management plans and poor utilisation of standard protocols resulted in 26 deaths. There was delay in referring four cases and incorrect diagnosis of three cases at secondary level contributed to the death of three women. Standard protocols were not followed in management of 19 women resulting in avoidable deaths in 19 women.

Summary

Hypertension was the commonest cause of maternal deaths in this interim report. It appears that women over 35 years at particular risk. Intracerebral haemorrhage is the most likely cause of death in eclampsia. Protocols of management for hypertensive disorders of pregnancy are either unavailable or are not being utilized; in addition there were delays in referrals, lack of communication between the different levels of health care and lack of trained staff.

Recommendations

1. The provision of Antenatal Care Education for Health Care personnel and patients emphasizing the dangers of hypertension in pregnancy.

2. The provision of protocols to all the antenatal clinics and hospitals for the management of hypertension in pregnancy and its complications and monitoring of their use in the clinics and hospitals

3. The establishment of a clear system of referral and equipping the designated regional hospitals with the appropriate expertise and equipment.

Table 2.1. Breakdown of specific causes of deaths due to Hypertensive Disorders of Pregnancy

Table 2.2. Demographic details

Table 2.3. Distribution of Maternal deaths due to complications of hypertension in the Provinces

Table 2.4. Avoidable factors, missed opportunities and sub-standard care

Chapter 3 - Non-pregnancy related infections and AIDS

Introduction

Worldwide 2,3 million people died from AIDS in 1997. In South Africa the seroprevalence data for 1997 showed that 14% of the population was HIV positive. This varies throughout the country with some areas like King Edward VIII Hospital having a HIV positive rate of 28%. In this interim report 24 women (18%) died of non-pregnancy related infections (Table 3.1). AIDS contributed 17 cases, and in 11 of these case pneumonia was the final cause of death.

Demographic data

The women who died of AIDS tended to be younger than the average, and having a lower parity. See Table 3.2.

Avoidable factors, missed opportunities and sub-standard care

Very few instances of sub-standard care were noted, however, one third of the women attended antenatal care infrequently.

Summary

The over-riding message that presents itself is that AIDS is with us, and it alone comprised 13% of all maternal deaths. More deaths may have been missed as the testing for HIV was very infrequent and the condition was probably under-diagnosed. The lack of detection of sub-standard care in pregnant women with AIDS reflects our lack of knowledge on how to management a pregnant HIV+ woman to decrease morbidity and mortality rather than good care.

Recommendations

1. This information of the importance of AIDS in maternal mortality needs to be fed into public health education, and these education efforts need to be intensified.

2. Doctors should be advised to test for HIV in a young woman (<25 years) with pneumonia that does not respond rapidly to treatment.

3. Protocols for the management of pregnant HIV positive women need to be developed. Issues such as screening for HIV, antibiotic prophylaxis, the use of antivirals and particularly the appropriate intrapartum management, which might differ from HIV negative women, need to be investigated.

Table 3.1. Breakdown of the specific causes of death due to non-pregnancy related infections

Table 3.2. Demographic data

Chapter 4 - Obstetric Haemorrhage

Introduction

Obstetric haemorrhage caused the death of 20 (14,2%) of the 133 deaths. This was the second most common cause of maternal deaths. Sufficient data was received on 20 cases on which this chapter is based. The primary cause of death was antepartum haemorrhage in 7 cases and postpartum haemorrhage in 13 cases.

Demographic data

The average age of these women were 27 (± 7,5) years and the parity 2 with a range from zero to 8. There were 4 women in the age group 35 years or more and 6 women had a parity of 5 or more. The distribution of these deaths is shown in Table 4.1.

Primary and final causes of death

Seven women died due to antepartum haemorrhage (Table 4.2). Abruptio placentae in women whose prenancies were complicated by hypertensive disorders (4 cases), occurred most commonly as the primary cause of death in this group.

Thirteen women died due to postpartum haemorrhage (Table 4.2). Rupture of the uterus with a previous caesarean section and rupture of an unscared uterus (4 cases each), occurred most commonly as the primary causes of death in this group.

The final causes of deaths in this group are summarised in Table 4.3.

Avoidable factors, missed opportunities and substandard care

Patient related avoidable factors occurred with a low frequency in the group with haemorrhage (Table 4.4). Administrative problems occurred more commonly with the availability of trained medical personnel cited as the most common problem (7 cases).

Sub-standard care during the intrapartum period (9 cases) occurred more frequently, compared to the antenatal and postpartum period (5 and 4 cases respectively). A lack of adequate resuscitation was the most common defect in the management of these women. The circulation was not adequately supported in 11 cases, blood not given where required in 9 cases and an inadequate amount of crystalloids given in 6 cases. In addition problems with emergency admission occurred with a high frequency at secondary hospitals.

Recommendations

Deaths due to postpartum haemorrhage are the most preventable of all the causes of maternal deaths. Patient related avoidable factors did not play an important role, with only one woman of the 20 in this group not attending antenatal care.

Deaths due to antepartum haemorrhage will be more difficult to prevent as women with hypertensive disorders that developed abruptio placentae was the most common problem is this group. However, adequate resuscitation and referral to at tertiary hospital must receive the highest priority in the management of these cases.

The following avoidable factors must be addressed as a matter of urgency to reduce deaths due to postpatum haemorrhage:

1. Health administrators must ensure:

• The availability of trained staff at all times in all centres conducting deliveries.

• The availability of blood for transfusion at all secondary hospitals.

• Protocols for the emergency management of heamorrhage are available in the hospitals and clinics, and that their use is monitored.

2. Midwives and doctors conducting deliveries at all levels of care must:

• Use partograms for every patient

• Be well versed in the correct use and interpretation of partograms.

• Commence an intravenous infusion before onset of the second stage of labour in all patients with a risk for postpartum haemorrhage, i.e. patients with a parity of 5 or more.

• The active management of the third stage of labour should be routinely practiced for all deliveries as it has been clearly shown to decrease the incidence of postpartum haemorrhage.

• Recognise excessive blood loss in the third stage of labour early and immediately proceed to an intravenous infusion and measures to contract the uterus well (including the use of oxytocin).

3. Women that will be allowed a trial of scar must be managed according to a strict protocol.

4. Women with postpartum haemorrhage transferred to secondary hospitals must:

• Be admitted immediately to the labour wards.

• Adequate resuscitation must commence immediately.

• Be managed according to an appropriate protocol.

5. All antenatal patients must be screened for anaemia during the antenatal period and corrective measures instituted if cases are identified.

Table 4.1. The distribution of deaths in the Provinces

Table 4.2. Primary Causes of death

Table 4.3. Final Causes

Table 4.4. Avoidable factors, missed opportunities and sub-standard care

Chapter 5 - Early Pregnancy Loss

Introduction

Early pregnancy loss contributed 16 (12%) of the maternal deaths in this interim report (Appendix - Table 11.2). The distribution is shown in Table 5.1. However, it is likely that this is a severe underestimation because KZN did not report any deaths due to septic abortions (Table 5.2). Also ectopic pregnancies are also being grossly under reported because in the "Near miss" study¹ they represented 10% of the cases. The near misses reflect the maternal death pattern fairly well.

This chapter will concentrate further solely on the septic abortions.

Demographic data

The average age of women dying from septic abortion was 28,2 (± 7,6) years and the median parity was 1 (range 0-4). One case was less than 20 years and 5 cases were older than 30 years. Four cases were pregnant for the first time and one case had a parity of 4. The average gestational age was 14,3 weeks at the time of the abortion, and 71% of the women were in their second trimester at the time of the abortion.

Final and contributory causes of death

These are shown in Table 5.3 and reflect all the complications of septic shock.

Avoidable factors, missed opportunities and sub-standard care

In 7 cases it was recorded that the women had induced the abortion themselves or with the help of people outside of the health services. It also was recorded that in these cases there was no functioning Termination of Pregnancy (TOP) services in the area from which these women originated (Table 5.4).

In the 9 cases where there was a problem with the emergency care of the patient, the diagnosis of septic abortion was not made in 8 cases. The diagnosis was missed at all levels of health care. The resuscitation was also inadequate in 5 of the cases. An emergency hysterectomy was only performed in 2 cases, indicating that the delay in making the diagnosis severely compromised the chances of the woman surviving.

Recommendations

1. Termination of Pregnancy services. The patchy implementation of the Choice of Termination of Pregnancy Act of 1996 needs to be rectified. The public needs to know that termination of pregnancy is available and specifically at which sites it is available.

2. The recognition of septic shock must be re-emphasized to all medical personnel involved in the care of pregnant women and all casualty officers.

3. A standard protocol for the management of abortions should be available in each hospital where abortions are managed.

Reference

Mantel GD, Buchmann E, Rees H, Pattinson RC. Severe acute maternal morbidity: A pilot study of a definition for a "near miss". Br J Obstet Gynaecol in press

Table 5.1. Primary cause of death in early pregnancy

Table 5.2. Distribution of septic abortion deaths within the Provinces

Table 5.3. Final and contributory causes of septic abortion deaths

Table 5.4. Avoidable factors, missed opportunities and sub-standard care

Chapter 6 - Pregnancy Related Sepsis

Introduction

There were 21 maternal deaths, directly due to pregnancy related sepsis. Of these 12 were as a result of septic abortion and have been discussed in Chapter 5. The remaining 9 are discussed in this Chapter. All pregnancy related sepsis was responsible for 16% of all maternal deaths.

Demographics

The average age of the women was 23,6 ( 5,7) years, and median parity was 1, (range 1-3). Seven of the patients had a parity of 1. The were 2 patients less than 20 years, 4 patients between 20-24 years, 2 patients 25-29 years and one was more than 34 years old.

Eight of the deaths occurred in a secondary hospital and the remaining one in a tertiary hospital.

Primary and final cause of deaths

Of the 9 cases of puerperal sepsis, 6 occurred after caesarean section and three after vaginal delivery. Two of the caesarean sections were performed for obstructed labour. See Table 6.1.

The final cause of death was multi-organ failure in 3 patients; septic shock in a further 3; and pulmonary oedema in 1. See Table 6.2.

Avoidable factors, missed opportunities and sub-standard care

Instances of sub-standard care associated with the 9 deaths have been divided into patient related; administrative; and management related problems.

• Patient-related problems. These were present in 5 cases. In four, there had been delay in seeking medical assistance. The reasons therefor (e.g. transport problems) were not apparent. One patient declined care at one hospital and discharged herself to go to another hospital.

• Administrative problems. These were identified in 4 cases. They included an insufficient number of intensive care beds; inadequate numbers of appropriately-trained medical staff; and poor communication between the referring and admitting clinics or hospitals.

• Management-related problems. Such problems were identified in a number of cases. In several, more than one instance of sub-standard care was evident. Prominent defects in management included poor note-keeping; poor use or absences of management protocols; delay in the diagnosis and management of recognized complications of pregnancy; and delay in the referral of patients, either to more senior staff or to a secondary or tertiary hospital when complications arose.

Recommendations

1. Ensure that the danger signs for problems in the postpartum period are incorporated into the health education programme for antenatal patients and patients awaiting discharge post delivery

2. Ensure that all clinics and hospitals are aware that all clinical notes must accompany a patient, when she is transferred to another institution, and to ensure this occurs.

3. Ensure protocols for the management of postpartum sepsis are available at all delivery sites, especially where casearean sections are being carried out.

4. Ensure that prophylactic antibiotics are used routinely for ceasarean sections as there is overwhelming evidence that this significantly decreases postpartum morbidity.

Table 6.1. Primary cause of death

Table 6.2. Final cause of death

Chapter 7 - Pre-existing Medical Diseases

Introduction

Pre-existing medical diseases contributed to 10,5% of all the maternal deaths. The most common of the conditions were those related to the heart, these comprising half of all deaths due to pre-existing medical conditions (Table 7.1).

Demographics

The demographics of women dying due to heart disease is noted here as it is the only significant group. The average of the women with pre-existing heart disease was 34,1 (± 8,1) years, and the median parity 3 (range 1-5). There were 5 cases 35 years and older and 4 cases with a parity of 3 or more.

Avoidable factors, missed opportunities and sub-standard care

Four of the women dying of pre-existing heart disease did not attend any antenatal clinics. Also medical practitioners were aware of the medical condition of four of the women to prior to pregnancy, but there appears to have been no counselling regarding contraception.

Recommendations

1. The importance of pre-conception counselling must be re-emphasized to all medical practitioners, especially those seeing women with heart disease.

2. Contraceptive advice must be given to women with heart disease, especially regarding family size and the dangers of pregnancy.

3. Women with pre-existing heart disease should be referred to tertiary level hospital for review during the pregnancy and for planning of the antenatal care and delivery. The delivery should at least take place at secondary level hospitals.

Table 7.1. Primary causes of maternal deaths in women with pre-existing medical diseases.

Chapter 8 - Obstetric Anaesthesia and Analgesia Deaths

Anaesthetic accidents were responsible for 6 maternal deaths - 4,5% of the total number (Table 8.1). The women were apparently healthy prior to undergoing the anaesthetic.

In these cases the note keeping was generally poor and it appeared that the medical staff did not have the required skills to manage the complications. There could also be a considerable under-reporting on anaesthetic deaths as this interim report is poorly represented by provinces with large rural areas.

The future reports will pay particular attention to anaesthetic deaths. More information must be obtained regarding circumstances surrounding the anaesthetic accidents before any substantive comment can be made.

Table 8.1. Maternal deaths due to anaesthetic accidents

Chapter 9. Methodology of data collection and data analysis

Study Administration

All information regarding maternal deaths are treated as privileged information and kept locked. All deaths are given a unique case number by the administrative clerks responsible for receiving and sorting forms. All identifiers are removed from the forms which are then entered on a stand alone computer in a room accessible only to designated staff members of the DOH and Professor Pattinson (a member of the NCCEMD).

Provincial coordinators identified in each province oversee and monitor activities related to the collection and collation of the information and are responsible ensuring for timeliness and completeness of data.

Study Design

A prospective method is used for the Enquiry. All maternal deaths in South Africa, which had a Maternal Death Notification Form filled-in and received by the NCCEMD by the 14^th February, since the onset of the pilot studies (1^st August, 1997) are the study population for this interim report.

Data Collection

A specific data collection tool was devised for the purpose of the enquiry and distributed in all nine provinces to be used as the notification form. It was designed to enable the collection of data related to the demographics of the patient as well as relevant information related to her medical history and current pregnancy. Clinic information recorded from the time of admission to the time the patient died is also reflected on the notification form.

Data processing

All reported cases were entered on Access 97 in a database designed specifically for the enquiry. The program uses Windows 95 and allows for analysis and yields measures of central tendency and measures of dispersion.

Quality Assurance

Supplementary information regarding the final cause of death, avoidable factors, missed opportunities and substandard care for each case was complied by provincial assessors of the NCCEMD on a separate sheet, the assessor form.

All information was cleaned before analysis by staff members of the Directorate Health Systems Research and Epidemiology, who also supervised the data entry process. Only cleaned information has been used in this report.

Data Analysis

Only descriptive data analysis was performed in this interim report.

Maternal Mortality Rates

No mortality rates have been calculated in this interim report because the data is incomplete. In future reports a mortality rate will be calculated using an estimated number of live births (as described below) as the denominator and the maternal deaths as a numerator.

In order to generate the estimated number of live births for 1997 and 1998, the 1991 census using age specific fertility rates by age will be used and the denominator derived. For example, according to figures derived from the 1991 census there were 11 072 397 women of child-bearing age in South Africa in 1997 and the corresponding estimated live births that occurred then are 1 369 634 (Table 9.1).

Once the complete report of the 1996 census is available this will form the basis for the number of live births calculation. The accuracy of the estimates will be checked using areas where the number of live births is known.

Table 9.1. Estimated number of women of childbearing age and live births in South Africa for 1997

Chapter 10. Conclusion

In this interim report, certain specific problem areas have been identified which have special reference to medical personnel and administrators. They are:

• Referral hospitals for hypertension. It was clear that in each health region a hospital needs to be identified that is equipped and trained to manage the complications of hypertension. A clear set of referral criteria and referral patterns needs to be put in place in each region.
• Pneumonia and AIDS. The presence of pneumonia in a women under 25 years of age should immediately alert the medical staff to the possibility of the women being HIV positive and she should be counselled and tested for HIV.
• Recognition of obstructed labour. Emphasis needs to be placed on the recognition of cephalopelvic disproportion. The use of the partogram, and the implicit management protocols which follow, must be re-emphasised at all levels of care.
• Management of labour in women with previous caesarean sections. Women with a previous caesarean section must undergo their labours in a secondary hospital and a clear protocol must be in place for the management of these women at the secondary level hospitals.
• Availability of blood. A look needs to be taken at the availability of emergency blood at institutions providing delivery services.
• Termination of Pregnancy services. The patchy implementation of the Choice of Termination of Pregnancy Act of 1996 needs to be rectified. The public needs to know that termination of pregnancy is available and specifically at which sites it is available.
• Multidisciplinary care. The value of multidisciplinary care needs to be re-emphasied in the health services and at the training institutions. This is especially important in relation to women with pre-existing heart disease.
• Anaesthetic services. More attention needs to be placed on the training of doctors to provide safe anaesthesia for pregnant women.
• Prophylactic antibiotics. It has been clearly shown that the use of prophylactic antibiotics prior to caesarean sections decreases the post-operative morbidity. This should be standard policy at all places performing caesarean sections.
• Family planning service. The women over 35 years of age are at special risk of dying during pregnancy. The attention of the family planning services needs to be focused on this issue and the clients counseled accordingly.

Table 11.1. Cases submitted per Province

Note: These are the cases reported by the 14/2/98, the maternal death notification forms of 133 were in possession of the NCCEMD as of 14/2/98.

Table 11.2. Summary of the demographic details of maternal deaths

Table 11.3. Primary Obstetric Cause of Death

Note: Not all the causes of death are noted in the sub-categories, only the major ones.

Table 11.4. Assessor-s view of avoidable factors, missed opportunities and sub-standard care

Appendix 2. Definitions and Method of Assessment

Assessing Avoidable Factors, Missed Opportunities and Sub-Standard Care in Confidential Enquiries into Maternal Deaths

As health care workers involved in the care of pregnant women, we are aware of our own and the health systems shortcomings. We mostly aspire to improving the standard and quality of health care but are often at a loss at what to do.

One effective method for achieving an improvement in care and pointing us in the right direction is to identify problems in the care of pregnant women and devise solutions for these problems. It is very difficult to analyse the standard of care of every pregnant woman, therefore a sample of women needs to be taken. The sample should be easily defined and representative of the whole population. There should not be too many cases to overwhelm the investigators, but enough to get meaningful results. The sample should also comprise of cases where the problems are most likely to be identified. Maternal deaths clearly meet all these criteria. Maternal deaths as the criteria for admission to the sample has the added advantage in that by monitoring the Maternal Mortality Rate, the success of the solutions implemented can be measured.

This audit method of improving care can be summarised by reference to the tip of the iceberg theory, viz. by concentrating on a solving factors related to a small but important area of a problem, the ripple effects will be felt by a large number. For the "small but important area" read maternal deaths, for "problem" read standard and quality of care of pregnant women, and for "large number" read all pregnant women. Hence, by concentrating on solving factors related to maternal death with respect to the quality and standard of care of pregnant women, the ripple effects of these solutions will be felt by all pregnant women.

The effectiveness of this method has been clearly demonstrated in relation to the care of the fetus and newborn^1,2.

The basic philosophy rests on two assumptions:

1. The study of a "few but important" cases will detect the problems in care of all cases.
2. The problems that lead to a maternal death are the same as those of other pregnant women, but are just more severe, or the woman had an unlucky break.

There is compelling evidence that, provided there is a large enough sample, a relatively few cases can represent the picture for the population. There is also strong evidence that the problems of care in maternal deaths are similar to those where pregnant women suffered acute severe morbidity but survived³.

Therefore, to achieve the goal of improving the care of all pregnant women, the National Department of Health has made it compulsory (from 1/12/97) to notify the Department of all maternal deaths. Prior to this the National Department of Health appointed a National Committee for the Confidential Enquiry into Maternal Deaths (NCCEMD) and tasked it with analysing the maternal deaths and proposing workable solutions for the problems detected.

Aims of assessors

The aims of each maternal death assessment are firstly to detect errors or omissions in the care of the woman such that these errors or omissions can be prevented in the future. It is not sufficient for an assessor to say the care was sub-standard, the assessor has to pinpoint what specifically in the care was sub-standard. The second objective for an assessor is that once a problem is identified in the care of a patient it must be described in such a way that it is logical, useable and readily understandable for the average health worker.

Definition of sub-standard care

The NCCEMD adopted the definition of sub-standard care that is used by the United Kingdom.

"Substandard care

The term substandard care has been used in this report to take into account not only failure in clinical care, but also some of the underlying factors which may have produced a low standard of care for the patient. This includes situations produced by the action of the woman herself, or her relatives, which may be outside the control of the clinicians. It also takes into account shortage of resources for staffing facilities; and administrative failure in the maternity services and the back-up facilities such as anaesthetic, radiological and pathology services. It is used in preference to the term "avoidable factors" which was used previously in the England and Wales Reports until 1979 and has also been used in the Scottish and Northern Ireland reports. This was sometimes misinterpreted in the past, and taken to mean that avoiding these factors would necessarily have prevented the death. "Substandard" in the context of the report means that the care that the patient received, or care that was made available to her, fell below the standard which the authors considered should have been offered to her in this triennium." ⁵

It is extremely important to understand that in assessing the case, the assessor is looking at the care in totality. This includes how were the basic care was performed, e.g. during antenatal care was the haemoglobin level recorded, as well as the event that led to the woman-s death.

When performing an assessment of a maternal death, the spirit that pervades the assessor-s enquiry is how could we have done better? The basic guidelines that the assessor employs are those commonly used in the country and those that are applicable to the level of care. The Maternal Manual of the Perinatal Education Programme is used as a basis for assessing care at a primary level, and local protocols are used for tertiary level care.

The assessment of care system used in the Confidential Enquiry

The basic structure of the assessment of care system devised for the confidential enquiry is to analyse 1) the woman and her environment; 2) the administrative circumstances surrounding the care and 3) the quality of health care, each individually. Analysis in these three categories has the obvious advantage that information gained in these areas can be fed back immediately to the relevant sectors for action.

1. Patient related problems. (The woman and her environment)

Assessing avoidable factors or missed opportunities related to the woman and her environment is divided into three sections:

• The first deals with issues relating to the woman herself. For example a woman may delay reporting to a clinic that her membranes have ruptured and this delay may result in chorioamnionitis ultimately leading to her death. This delay may have been due to lack of knowledge.

• The second relates to the influence the family has on the behaviour of the woman. For example, the woman may decline a hysterectomy for severe puerperal sepsis until she has consulted with her family, and this delay may contribute to her death.

• The third deals with the effect of the environment (community) where the woman lives on her behaviour. For example the lack of a telephone to call the ambulance service in a community, or the attitude of the community towards the health services may influence the outcome.

Clearly much of this information may not be currently available in patient-s case notes, but hopefully it will be discussed in the meeting where the Maternal Death Notification Form is filled in. One of the purposes of widely distributing the method of assessment of maternal deaths is to raise awareness in professionals and in this way begin getting more information regarding each case. However, when collecting and interpreting the information it is important not to victim blame. The full story tends to be complicated, and often circumstances dictate the behaviour and not neglect or desire to do damage on the side of the woman.

2. Administrative problems.

Administrative problems are easier to classify and the information easier to obtain. The problems are classified as:

• Transport, e.g. lack of transport from the clinic to hospital or hospital to hospital.
• Barriers to access to health care facility, e.g. hostile reception at clinics or hospitals
• Lack of accessibility e.g. lack of termination of pregnancy services
• Lack of health care facilities, e.g. intensive care beds
• Lack of personnel
• Lack of appropriately trained staff
• Communication e.g. break down in communication between patients and health care workers or between health care workers.

3. Standard of health care

The assessment of the standard of care given to the woman by the health workers is divided into distinct periods and the care is assessed in each period where applicable. The periods are:

1. Antenatal care }
2. Intrapartum care } these refer to the routine care of the patient.
3. Postpartum care }
4. Emergency event/admission }
5. Resuscitation } these refer to the event that lead to the death
6. Anaesthesia }

The assessment of each section is based on how the following were performed in each case:

• initial assessment i.e. history, examination and special investigations,
• problem identification or diagnosis,
• management plan or management protocol followed,
• follow-up monitoring performed

(Assessment of the resuscitation follows the ABCD protocol - airways, breathing, circulation, drugs).

This system is used because it is the way each patient should be managed and is consistent with the way that the medical staff has been taught.

Cases where a death might not be able to be assessed because of lack of notes or missing notes are also recorded. Valuable information is obtained here. The very fact there are no notes indicates sub-standard care, also cases of missing files indicates a major administrative problem. The assessors also look at the behaviour of the health personnel and record areas where they think there are problems. For example, was the patient seen regularly or not.

In the South African Confidential Enquiry into Maternal Deaths, the assessors have a checklist which covers all aspects of the case as described above. The checklist is used to ensure a uniform approach to case assessment. The areas where sub-standard care are thought to be present are recorded and subsequently all the information is computerised. The cause of each death is recorded and the assessment of care in each case is linked. In this way disease specific information is obtained and specifically where the problems lie in a particular disease can be identified. Consequently, it is easier to formulate solutions.

For example, this interim report shows that maternal deaths due to septic abortion are in the top five of the major causes of death. In three-quarters of the cases the diagnosis of septic shock was initially missed and only diagnosed too late. In less than one-fifth a hysterectomy was performed. It was also apparent that in two thirds of cases the women herself or people from outside the health service induced the abortion. In the areas where these women came from there were no Provincial termination of pregnancy services. This indicates a lack of accessibility of a health service for women and this has been brought to the attention of the authorities.

References

1. Pattinson RC, Makin JD, Shaw A, Delport SD. The value of incorporating avoidable factors into perinatal audits. S Afr Med J 1995;85:145-147

2. Pattinson RC, E de Jonge, Pistorius LR, et al. Practical application of data obtained from a Perinatal Problem Identification Programme. S Afr Med J 1995;85:131-132

3. Mantel GD, Buchmann E, Rees H, Pattinson RC. Severe acute maternal morbidity: A pilot study of a definition for a "near miss". Br J Obstet Gynaecol in press

4. Theron GB. National confidential enquiries into maternal deaths in South Africa. Specialist Medicine 1998;19:6-12

5. Report on Confidential Enquiries into Maternal Deaths in the United Kingdom 1985-87, page xiv.

Top