Child Survival, Maternal Health and Mortality: research findings by Medical Research Council

Meeting report

Opening Statement
Ms Debbie Bradshaw, Researcher: MRC Burden of Disease Research Unit, brought greetings from MRC Acting President, Dr Ali Dhansay. He apologised for absence, but he was delighted that the MRC had the opportunity to interact with the Committee. Her presentation would try to make sense of the data on maternal and child mortality and would include research conducted by the MRC. She hoped the briefing would address some of the challenges the country faced. She acknowledged the work done by her colleague, Nadine Nannan, a MRC Senior Scientist, and other employees at the MRC.

Medical Research Council Briefing
Ms Bradshaw said that the presentation looked at what the MRC wanted to measure, the data available in the country, what the data said, and the research being conducted by the MRC. With the release of the Millennium Development Goals Report, there was a lot of attention paid to child and maternal mortality.

Child Mortality
The “Under-5 Mortality Rate” looked at the chance of a newborn baby dying before the age of five and was represented out of every 1000 live births. Western Europe countries had a rate of 4.6, so out of 1000 babies born, 4.6 would die. In Central Africa, 219 babies out of 1000 would die. There was a wide range between countries and regions. The global average was 72. The rate at which newborn babies died in South Africa was 69 out of 1000 babies born; however, this figure was in contention. UNICEF reported that the global average decreased from 89 in 1990 to 60 in 2009.The report showed that rates for Sub-Saharan Africa declined from 180 to 129.

The ideal source of data for monitoring child mortality and other forms of mortality was to have complete vital registration of births and deaths together with regular censuses and mid-year population estimates. This meant that births and deaths had to be registered with the Department of Home Affairs (DHA) on time. An estimation of the country's population was also needed. Not all births and deaths were being registered on time. The country needed routine estimates of the Infant Mortality Rate (IMR) and Under-5 Mortality Rate (U5MR), as well as information about the causes of the deaths. In the absence of complete vital registration, the country used other methods. They used a Full Birth History (FBH) from a representative sample of women from information gathered from Statistics South Africa, the Department of Health (DoH), the MRC, or the Human Sciences Research Council (HSRC). Information was gathered from a sample of women from households across the country that was representative of all women in the country. The women were then asked about the dates of the births of their children, if each child was still alive, and how old the child was if he/she died. This was quite a detailed survey, so the Summary Birth History (SBH) was introduced. The survey asked the mothers how many births they had, and how many of the children survived. The MRC hoped this survey would be included in the 2011 census.

The data from Statistics SA was gathered from the Department of Home Affairs. Statistics SA's data for the U5MR suggested that the registered deaths increased from 30 000 in 1997 to approximately 60 000 in 2006. Data for 2007 showed a slight decrease for the registration of deaths to just under 60 000. Statistics SA's information seemed to be on the low side when compared with other estimates. It was difficult to interpret the trends for death registrations as the information was incomplete. The number of registrations improved in recent years making it difficult to distinguish increases in death rates from improvements in the registration of deaths.

The Demographic and Health Survey (DHS) from the DoH was a national household survey from a sample of households chosen to represent the country. There were three surveys. The first survey was conducted before 1994 and showed that child mortality was declining. The next survey in 1998 showed that the decline had been reversed and was starting to increase. The last survey conducted in 2003 did not concur with earlier studies. The MRC found that there were problems with the field work and the data quality of the survey. This meant the information taken from the last survey was invalid, as the information was also quite old. However, the surveys showed that child mortality declined until the early 1990s but that the current levels remained very uncertain.

The 2007 Community Survey conducted by Statistics SA was an enormous household survey that asked women how many children she had given birth to and how many were still alive. This was the SBH survey methodology. The results showed the level of child mortality for the U5MR has been stable since 1998. This was the best piece of empirical data that the country had currently. It showed that there were between 70 and 75 deaths for every 1000 babies born.

The MRC looked at how to assess the trends in child mortality. Every country had to identify all the nationally representative data and assess their quality. This required looking at sampling errors, omissions of deaths, misreporting of a child's age at death or their date of birth, selection bias, violation of assumptions, and bias from the impact of HIV/AIDS. A regression line was then fit into all data points that met data quality standards and the adjustment for the impact of HIV/AIDS. South Africa had very good data for the prevalence of HIV in pregnant women. The presentation showed that there was not a lot of data available in the country and all of the surveys showed different outcomes. This meant that different analysts were coming up with different information. There was a lot of uncertainty but it was clear that by 2005 rates were either stagnant or increasing, and the country was not making progress towards the MDGs. The MDG report for South Africa had an implausible level for child mortality. It said there were 104 child mortalities for every 1000 births in 2007. The MRC was not quite sure how this figure was derived. The MRC's assessment was that the child mortality rate was approximately 75 deaths per 1000 births. It was important for the country to look at what was actually causing babies to die. The MRC was unhappy with the quality of information that doctors were putting on the death certificates, as many did not say what the underlying cause of the death was. This was the information that the country needed in order to reduce mortality. There was quite a bit of work that needed to be done to train doctors to complete death certificates.

The MRC looked at causes of deaths in newborns and children under five years for 2000-2005. They estimated that 35% of the deaths were due to HIV/AIDS, 6% was due to pneumonia, 5% due to injuries, 11% due to diarrhoea, 11% due to other child illnesses, and 2% due to sepsis and meningitis. 30% of the mortalities were due to neonatal causes such as infections, pre-term births, birth asphyxia, and congenital conditions.

Maternal Mortality

The Maternal Mortality Rate (MMR) looked at the number of maternal deaths relative to 100 000 live births. Maternal deaths were often missed in routine statistics. It was very difficult to get the full picture. There was a system of confidential enquiry with hospitals for maternal deaths that happened in labour wards or obstetric units. This information was reported to the DoH. It was a challenge to get information about maternal deaths that happened in wards other than the obstetrics and labour wards. This usually happened when mothers became sick after leaving the hospital. Once they became sick and were admitted to the hospital again, they would not be admitted back into the labour unit. Information on maternal deaths in the private sector as well as maternal deaths in the community was also found to be lacking.

The ideal data systems for maternal mortality included the implementation of the vital registration system that would include a question on the pregnancy status of deceased females. A health facility audit was also needed so that it could review each death to assess whether it was preventable and contributed to statistics. There was a need for data linkage between the confidential enquiry system and the death registration system. In the absence of an ideal system, countries have tried to include in the surveys or census questions of whether there was a death in the household in the past year, how old the person was if there was a death and what the gender was. If the person was a woman of child bearing age, they would specifically ask if she was pregnant. A different approach was used in the DHS where if a woman was interviewed, she was also asked about her sisters. This was called the sibling method.

When the data was put together it was found that there was a whole range of estimates. The statistics were problematic. The confidential enquiry showed that there were approximately 120 maternal deaths per 100 000 births. Vital statistics showed there were approximately 210 maternal deaths.  The UN Model estimated that the rate was 400 per 100 000 births in 2005. The MRC's model estimated a maternal mortality rate of 300 for 2006. This showed that data systems were not followed and that there was uncertainty. However, it was clear that the maternal mortality rate has increased. The confidential enquiry showed that 38% of the deaths were due to infections that aggravated the pregnancy and caused death. The MRC thought that the main reason for the increase was due to HIV/AIDS, as it was the most common infection. There was an urgent need to strengthen the health information system to adequately monitor maternal and child mortality.

Ms Bradshaw informed the Committee that a presentation on the MRC's 2009/10 Annual Report had been made to the Portfolio Committee on Health on 27 October 2010. She wanted to use the presentation to pick out certain important information for the Members. The MRC took the DoH's strategic plan very seriously. The plan included of strengthening research and development. Four of the DoH's deliverables included increasing life expectancy, decreasing maternal and child mortality, combating HIV/AIDS and decreasing the burden of TB, and improving the effectiveness of the health system. With its current and improved portfolio of research, the MRC would support the DoH in its endeavours. 

The MRC was involved in a series of articles produced in the Lancet in 2009, where they looked at the health system from a number of different angles. The Lancet had an article called Saving the Lives of South Africa's Mothers, Babies and Children: Can the Health System Deliver. The paper identified that there were a lot of programmes that the health services in the country had already identified as part of their policy. However, these different health services were not communicating with one another. The continuity of care was a major issue. It was also found that 8% of mothers exclusively breast fed babies in their first three months of life. This was not high enough and had to be improved. Initially, this topic was confusing because of the HIV/AIDS infection, but there were now clear guidelines about breast feeding. The key messages of the article said that HIV/AIDS and poor implementation of existing packages of care were the main reasons for the lack of progress towards the MDGs, and full coverage of key packages of interventions such as the treatment and prevention of the HIV infection and provision of comprehensive maternal and neonatal care would put South Africa on track. It was financially feasible to achieve a high coverage of priority care for mothers, neonates and children; however, it required a 2.4% increase in expenditure and the money had to be spent strategically. The strengthening of leadership, accountability mechanisms, and a high quality of care interventions was also required.

Discussion
Ms D Duncan (DA) addressed the FBH and SBH surveys. She asked if the entities conducting the surveys ever considered asking what the women thought the causes of the deaths could be. 

Ms Bradshaw replied that the FBH survey included a question concerning common conditions that caused infants to die such as whether the infant had diarrhoea. However, this method did not work when it came to more serious illnesses that required a doctor's diagnosis. The questions had to be as short as possible, as it was impossible to spend long periods of time with each woman in the sample.

The Chairperson asked if the MRC ever interacted with the DHA regarding the timeous registration of births and deaths.

Ms Bradshaw answered that she had been working on this challenge for a number of years. Between 1996 and 1998 there were big efforts from the DHA, the DoH, Statistics SA and the MRC to work together to improve the registration process. There has been quite a marked improvement over the past few years, as registration coverage increased from 56% to approximately 90%. However, the collaborations had not been consistent. Currently, the Minister of Health was committed to address this issue. He set up an advisory committee to analyse the statistics and see what had to be done. They just had their first meeting. There were continuous attempts, in general, to improve the registration process. She did not have the statistics on hand, but the registration of births has increased remarkably over the last two years.

Ms P Maduna (ANC) commented that the DHA was dealing with a piece of legislation that would focus on the registration of births. Registration would have to done within 31 days of the birth of the baby. This would help to improve the registration process.

Ms P Lebenya (ANC) asked if the surveys took into consideration those that lived in poverty in rural areas of the country. Many of these women did not even go to hospitals and had to deliver their babies at home. These were the areas that were most affected by child mortality.

Ms Bradshaw replied that the Member was correct. That was why it was important to conduct the surveys on a sample of women that were selected from all living conditions. All women were included in the sample or census approach. The vital registration system would be able to increase the coverage of births and deaths; however, it might not reach those in deep rural areas.

Ms Duncan was concerned that the country had such a large number of research institutions that used their own research methodologies, which resulted in a number of different outcomes. She asked if it was fair to compare the decline of child mortality in 1990 with the results given for after 1990 if different samples were used for the different surveys. 

Ms Bradshaw explained that Demographic and Health Survey conducted around 1990 consisted of fourteen surveys conducted in different homelands and the data was collected separately. The information was pooled on the basis of the size of each homeland. The Member had a right to be concerned about whether the data was correct. The MRC was encouraged by the fact that the information from the survey was consistent with the data from the next DHS conducted in 1998.

Ms Nannan added that it was important that surveys were nationally representative like the DHS was.

Ms Duncan clarified that her main concern was that a comparison should be done using the same sort of data or the information would not be accurate.

Mr D Kekana (ANC) said that it would have been helpful for the Committee if the MRC showed the results for the mortality rates from the surveys and gave Members an idea of how the problems could be remedied.

Ms Bradshaw replied that the presentation showed the research conducted by the MRC and what they advised should be done to address the issues.

The Chairperson noted that the MRC said that the MDG report's child mortality rate of 104 was implausible. She asked whose responsibility it was to monitor and ensure that there were accurate statistics for child and maternal mortality.

Ms Bradshaw answered that she thought the Statistician-General (SG) was responsible for it, but he had a difficult job because there was so little data available to the country since the vital registration process was not adequate yet and there was no survey to look at a woman's full birth history. The SG had very little reliable data to work with, but she thought he could have done a better job with estimating from the data he already had.

Ms Nannan added that it was possible that the consultation process for producing the MDG report was not as wide as it should have been.

Ms Bradshaw added that she thought, ultimately, it was Statistics SA's responsibility.

The Chairperson noted that Ms Nannan and Ms Bradshaw's responses had been very helpful. It was a concern that the consultation processes leading to the MDG report might not have been broad enough. It was the Committee's duty to raise this concern so they could overcome these problems in the future.

Ms Maduna asked what the MRC thought the cause of deaths in labour wards was. Was it due to carelessness of doctors and nurses?

Ms Bradshaw explained that she was not the correct person to ask as she did not work that closely with health services. However, she was aware that people felt that doctors and nurses needed more training and supervision.

Ms Duncan asked if 8% of women that were breastfeeding were 8 out of every 100 women in the country.

Ms Bradshaw responded in the affirmative. She clarified that 8 out of 100 women breast fed exclusively. Research showed that the combination of breast feeding and formula feeding was a very bad combination. Mothers either had to breast feed or formula feed, but they could not combine the two.

The Chairperson noted that after listening to the presentation, it was obvious that there were certain questions that the Committee had to raise with the DoH and not with the MRC. The MRC would make time to invite the DoH to Parliament. One of the issues was that the country seemed to have inexperienced staff to at hospitals that could not deal with certain matters. Sometimes the services rendered by hospital staff were not good enough.

Ms Duncan noted that the presentation showed that the country found it difficult to obtain information from hospitals in the private sector. How did the MRC deal with this? The hospitals might be private, but the people were still South African. It was important to know this information. The MRC should tell the Committee if there was anything the Members, as Parliamentarians, could do to assist them. Private entities should not be withholding certain information that research institutions in the country needed. This matter was of great concern to the Committee. She also addressed causes of death regarding maternal mortality. She thought the MRC had to take domestic violence into account because the physical fighting could be a lot of damage to pregnant women. She thanked the MRC for the work they had done and for the helpful presentation.

Ms Bradshaw replied that she did not think the private sector was withholding information; her concern was that the DoH was not collecting the information or they had not created the correct channel to collect the information. The information regarding maternal and child mortality had to be collected from the private sector at provincial level. There were some provinces that were performing better than others in this respect. She thanked the Member for raising the issue of the impact of domestic violence on maternal mortality. The Gender and Health Unit of the MRC did some important work on gender violence and were in the process of looking at murders of women and what proportion of the murders were by intimate partners.

The Chairperson thanked the MRC for their presentation, saying that the information she had given the Committee was very valuable and insightful. She advised that they include more information on what was being done in rural areas in future presentations.

The meeting was adjourned.