《妇幼保健学 Maternal and Child Health》课程教学资源（文献资料）2005世界卫生组织报告珍爱每个母亲和儿童（英文）The World Health Report 2005 - Make every mother and child count

mothers and children matter-so does their health 9 saved in 2003 alone.Towards the turn of the millennium,however,the overall down- ward trend was showing signs of slowing.Between 1970 and 1990,the under-five mortality rate dropped by 20%every decade:between 1990 and 2000 it dropped by only 12%(see Figure 1.1). The global averages also hide important regional differences.The slowing down of progress started in the 1980s in the WHO African and Western Pacific Regions,and during the 1990s in the Eastern Mediterranean Region.The African Region started out at the highest levels,saw the smallest reductions(around 5%by decade between 1980 and 2000)and the most marked slowing down.In contrast,progress continued or accelerated in the WHO Region of the Americas,and the South-East Asia and European Regions. The result is that the differences between regions are growing.The under-five mor- tality rate is now seven times higher in the African Region than in the European Region; the rate was "only"4.3 times higher in 1980 and 5.4 times higher in 1990.Child deaths are increasingly concentrated in the African Region(43%of the global total in 2003,up from 30%in 1990).As 28%of child deaths still occur in South-East Asia. two of the six WHO regions-Africa and South-East Asia-account for more than 70%of all child deaths.Looking at it another way,more than 50%of all child deaths are concentrated in just six countries:China,the Democratic Republic of the Congo, Ethiopia,India,Nigeria and Pakistan. The fortunes of the world's children have also been mixed in terms of their nutritional status.Overall,children today are better nourished:between 1990 and 2000 the global prevalence of stunting and underweight declined by 20%and 18%,respec- tively.Nevertheless,children across southern and central Asia continue to suffer very high levels of malnutrition,and throughout sub-Saharan Africa the numbers of children who are stunted and underweight increased in this period (52). THE NEWBORN DEATHS THAT WENT UNNOTICED If further progress is to be made in reducing child mortality,increased efforts are needed to bring about a substantial reduction in deaths among newborns.The first global estimates of neonatal mortality,dating from 1983(53),were derived using historical data and are generally considered to give only a rough indication of the magnitude of the problem.More rigorous estimates became available for 1995 and for 2000.These are based on national demographic surveys as well as on statistical models.The new estimates show that the burden of newborn mortality is considerably higher than many people realize Each year,about four million newborns die before they are four weeks old:98%of these deaths occur in developing countries.Newborn deaths now contribute to about 40%of all deaths in children under five years of age globally,and more than half of infant mortality (54,55).Rates are highest in sub-Saharan Africa and Asia.Two thirds of newborn deaths occur in the WHO Regions of Africa(28%)and South-East Asia (36%)(56).The gap between rich and poor countries is widening:neonatal mortal- ity is now 6.5 times lower in the high-income countries than in other countries.The lifetime risk for a woman to lose a newborn baby is now 1 in 5 in Africa,compared with 1 in 125 in more developed countries(57). The above figures do not include the 3.3 million stillbirths per year.Data on stillbirths are even more scarce than those on newborn deaths.This is not surprising,as only 14%of births in the world are registered.Both live births and deaths of newborns go underreported;fetal deaths are even more likely to go unreported,particularly early fetal deaths

mothers and children matter – so does their health 9 saved in 2003 alone. Towards the turn of the millennium, however, the overall down￾ward trend was showing signs of slowing. Between 1970 and 1990, the under-five mortality rate dropped by 20% every decade; between 1990 and 2000 it dropped by only 12% (see Figure 1.1). The global averages also hide important regional differences. The slowing down of progress started in the 1980s in the WHO African and Western Pacific Regions, and during the 1990s in the Eastern Mediterranean Region. The African Region started out at the highest levels, saw the smallest reductions (around 5% by decade between 1980 and 2000) and the most marked slowing down. In contrast, progress continued or accelerated in the WHO Region of the Americas, and the South-East Asia and European Regions. The result is that the differences between regions are growing. The under-five mor￾tality rate is now seven times higher in the African Region than in the European Region; the rate was “only” 4.3 times higher in 1980 and 5.4 times higher in 1990. Child deaths are increasingly concentrated in the African Region (43% of the global total in 2003, up from 30% in 1990). As 28% of child deaths still occur in South-East Asia, two of the six WHO regions – Africa and South-East Asia – account for more than 70% of all child deaths. Looking at it another way, more than 50% of all child deaths are concentrated in just six countries: China, the Democratic Republic of the Congo, Ethiopia, India, Nigeria and Pakistan. The fortunes of the world’s children have also been mixed in terms of their nutritional status. Overall, children today are better nourished: between 1990 and 2000 the global prevalence of stunting and underweight declined by 20% and 18%, respec￾tively. Nevertheless, children across southern and central Asia continue to suffer very high levels of malnutrition, and throughout sub-Saharan Africa the numbers of children who are stunted and underweight increased in this period (52). THE NEWBORN DEATHS THAT WENT UNNOTICED If further progress is to be made in reducing child mortality, increased efforts are needed to bring about a substantial reduction in deaths among newborns. The first global estimates of neonatal mortality, dating from 1983 (53), were derived using historical data and are generally considered to give only a rough indication of the magnitude of the problem. More rigorous estimates became available for 1995 and for 2000. These are based on national demographic surveys as well as on statistical models. The new estimates show that the burden of newborn mortality is considerably higher than many people realize. Each year, about four million newborns die before they are four weeks old: 98% of these deaths occur in developing countries. Newborn deaths now contribute to about 40% of all deaths in children under five years of age globally, and more than half of infant mortality (54, 55). Rates are highest in sub-Saharan Africa and Asia. Two thirds of newborn deaths occur in the WHO Regions of Africa (28%) and South-East Asia (36%) (56). The gap between rich and poor countries is widening: neonatal mortal￾ity is now 6.5 times lower in the high-income countries than in other countries. The lifetime risk for a woman to lose a newborn baby is now 1 in 5 in Africa, compared with 1 in 125 in more developed countries (57). The above figures do not include the 3.3 million stillbirths per year. Data on stillbirths are even more scarce than those on newborn deaths. This is not surprising, as only 14% of births in the world are registered. Both live births and deaths of newborns go underreported; fetal deaths are even more likely to go unreported, particularly early fetal deaths

10 The World Health Report 2005 While the burden of neonatal deaths and stillbirths is very substantial,it is in many ways only part of the problem,as the same conditions that contribute to it also cause severe and often lifelong disability.For example,over a million children who survive birth asphyxia each year develop problems such as cerebral palsy,learning difficulties and other disabilities(58).For every newborn baby who dies,at least another 20 suf- fer birth injury,infection,complications of preterm birth and other neonatal conditions. Their families are usually unprepared for such tragedies and are profoundly affected. The health and survival of newborn children is closely linked to that of their moth- ers.First,because healthier mothers have healthier babies;second,because where a mother gets no or inadequate care during pregnancy.childbirth and the postpartum period,this is usually also the case for her newborn baby.Figure 1.2 shows that both mothers and newborns have a better chance of survival if they have skilled help at birth. FEW SIGNS OF IMPROVEMENT IN MATERNAL HEALTH Pregnancy and childbirth and their consequences are still the leading causes of death. disease and disability among women of reproductive age in developing countries -more than any other single health problem.Over 300 million women in the devel- oping world currently suffer from short-term or long-term illness brought about by pregnancy and childbirth;529 000 die each year(including 68 000 as a result of an unsafe abortion),leaving behind children who are more likely to die because they are motherless(59). There have been few signs of global improvement in this situation.However,during the 1960s and 1970s,some countries did reduce their maternal mortality by half over Figure 1.2 Neonatal and maternal mortality are related to the absence of a skilled birth attendant 100 of births without skilled attendant Maternal mortality ratio per 10 000 live births Neonatal mortality rate per 1000 live births 75 50 25 Europe Americas Western Westem Eastem South-East South-East Africa Pacific Pacific without Mediterranean Asia without Asia China India

10 The World Health Report 2005 While the burden of neonatal deaths and stillbirths is very substantial, it is in many ways only part of the problem, as the same conditions that contribute to it also cause severe and often lifelong disability. For example, over a million children who survive birth asphyxia each year develop problems such as cerebral palsy, learning difficulties and other disabilities (58). For every newborn baby who dies, at least another 20 suf￾fer birth injury, infection, complications of preterm birth and other neonatal conditions. Their families are usually unprepared for such tragedies and are profoundly affected. The health and survival of newborn children is closely linked to that of their moth￾ers. First, because healthier mothers have healthier babies; second, because where a mother gets no or inadequate care during pregnancy, childbirth and the postpartum period, this is usually also the case for her newborn baby. Figure 1.2 shows that both mothers and newborns have a better chance of survival if they have skilled help at birth. FEW SIGNS OF IMPROVEMENT IN MATERNAL HEALTH Pregnancy and childbirth and their consequences are still the leading causes of death, disease and disability among women of reproductive age in developing countries – more than any other single health problem. Over 300 million women in the devel￾oping world currently suffer from short-term or long-term illness brought about by pregnancy and childbirth; 529 000 die each year (including 68 000 as a result of an unsafe abortion), leaving behind children who are more likely to die because they are motherless (59). There have been few signs of global improvement in this situation. However, during the 1960s and 1970s, some countries did reduce their maternal mortality by half over Figure 1.2 Neonatal and maternal mortality are related to the absence of a skilled birth attendant Europe Americas Western Pacific Western Pacific without China Eastern Mediterranean South-East Asia without India South-East Asia Africa 0 25 50 75 100 % of births without skilled attendant Maternal mortality ratio per 10 000 live births Neonatal mortality rate per 1000 live births

mothers and children matter-so does their health 11 a period of 10 years or less.A few countries such as Bolivia and Egypt have managed this in more recent years.Other countries appear to have suffered reversals(see Box 1.3).Recent success stories in maternal health are less often heard than those for child health.This is partly because it takes longer to show results,partly because changes in maternal mortality are much more difficult to measure with the sources of information available at present. Today,predictably.most maternal deaths occur in the poorest countries.These deaths are most numerous in Africa and Asia.Less than 1%of deaths occur in high-in- come countries.Maternal mortality is highest by far in sub-Saharan Africa,where the lifetime risk of maternal death is 1 in 16,compared with 1 in 2800 in rich countries. Information on maternal mortality remains a serious problem.In the late 1970s, less than one developing country in three was able to provide data-and these were usually only partial hospital statistics.The situation has now improved but births and deaths in developing countries are often only registered for small portions of the popu- lation except in some Asian and Latin American countries.Cause of death is routinely reported for only 100 countries of the world,covering one third of the world's popula- tion.It is even difficult to obtain reliable survey data that are nationally representative. For 62 developing countries,including most of those with very high levels of mortal- ity,the only existing estimates are based on statistical modelling.These are even more hazardous to interpret than those from surveys or partial death registration.The countries that rely on these modelled estimates represent 27%of the world's births. Effectively,this leaves no record of the fate of 36 million-about 1 out of 4-of the women who give birth every year. Gradual improvements in data availability,however,mean that a growing database now exists of maternal mortality by country.Since 1990,a joint working group of WHO.the United Nations Children's Fund(UNICEF)and the United Nations Population Fund(UNFPA)has been regularly assessing and synthesizing the available information (60).It has not been possible.though,to assess changes over time with any confi- dence:the uncertainty associated with maternal mortality estimates makes it difficult to say whether that mortality has gone up or down,so no global downturn in maternal mortality ratios can yet be asserted. Nevertheless,there is a sense of progress,backed by the tracking of indicators that point to significant increases in the uptake of care during pregnancy and childbirth Box 1.3 A reversal of maternal mortality in Malawi Malawi is one country that experienced a sig- 2001.Third,the quality of care within health (itself not independent from the HIV/AIDS nificant reversal in maternal mortality:from facilities deteriorated.Between 1989 and 2001 epidemic).In remote areas one midwife often 752 maternal deaths per 100 000 live births in the proportion of deaths associated with defi- has to run the entire rural health centre and 1992 to 1120 in 2000,according to the Malawi cient health care increased from 31%to 43% is expected to be available for work day and Demographic and Health Surveys.According to In 2001 only one mother out of four who died in night,seven days a week.One maternity unit confidential enquiries into materal deaths in the hospital had received standard care.Wrong out of 10 is closed for lack of staff.Hospitals health facilities in 1989 and 2001,three fac- diagnosis(11%of deaths),delays in starting also experience severe shortages of midwives, tors apparently contributed to this increase. treatment(19%),wrong treatment (16%),or and unskilled cleaners often conduct deliveries. First,there was a sharp proportional increase lack of blood for transfusion(18%):deficient The shortage of staff in maternity units is in deaths from AlDS.This is not surprising hospital care was the leading principal avoid- catastrophic and rapidly getting worse;the since Malawi's national HIV prevalence has able factor in 38%of deaths. chances of Malawi women giving birth in a safe now reached 8.4%.Second,fewer mothers The diminishing coverage and the worsening environment diminish accordingly. gave birth in health facilities:the proportion of the quality of care are related to the dropped from 55%to 43%between 2000 and deteriorating situation of the health workforce

mothers and children matter – so does their health 11 a period of 10 years or less. A few countries such as Bolivia and Egypt have managed this in more recent years. Other countries appear to have suffered reversals (see Box 1.3). Recent success stories in maternal health are less often heard than those for child health. This is partly because it takes longer to show results, partly because changes in maternal mortality are much more difficult to measure with the sources of information available at present. Today, predictably, most maternal deaths occur in the poorest countries. These deaths are most numerous in Africa and Asia. Less than 1% of deaths occur in high-in￾come countries. Maternal mortality is highest by far in sub-Saharan Africa, where the lifetime risk of maternal death is 1 in 16, compared with 1 in 2800 in rich countries. Information on maternal mortality remains a serious problem. In the late 1970s, less than one developing country in three was able to provide data – and these were usually only partial hospital statistics. The situation has now improved but births and deaths in developing countries are often only registered for small portions of the popu￾lation except in some Asian and Latin American countries. Cause of death is routinely reported for only 100 countries of the world, covering one third of the world’s popula￾tion. It is even difficult to obtain reliable survey data that are nationally representative. For 62 developing countries, including most of those with very high levels of mortal￾ity, the only existing estimates are based on statistical modelling. These are even more hazardous to interpret than those from surveys or partial death registration. The countries that rely on these modelled estimates represent 27% of the world’s births. Effectively, this leaves no record of the fate of 36 million – about 1 out of 4 – of the women who give birth every year. Gradual improvements in data availability, however, mean that a growing database now exists of maternal mortality by country. Since 1990, a joint working group of WHO, the United Nations Children’s Fund (UNICEF) and the United Nations Population Fund (UNFPA) has been regularly assessing and synthesizing the available information (60). It has not been possible, though, to assess changes over time with any confi- dence: the uncertainty associated with maternal mortality estimates makes it difficult to say whether that mortality has gone up or down, so no global downturn in maternal mortality ratios can yet be asserted. Nevertheless, there is a sense of progress, backed by the tracking of indicators that point to significant increases in the uptake of care during pregnancy and childbirth 2001. Third, the quality of care within health facilities deteriorated. Between 1989 and 2001 the proportion of deaths associated with defi- cient health care increased from 31% to 43%. In 2001 only one mother out of four who died in the hospital had received standard care. Wrong diagnosis (11% of deaths), delays in starting treatment (19%), wrong treatment (16%), or lack of blood for transfusion (18%): deficient hospital care was the leading principal avoid￾able factor in 38% of deaths. The diminishing coverage and the worsening of the quality of care are related to the deteriorating situation of the health workforce Malawi is one country that experienced a sig￾nificant reversal in maternal mortality: from 752 maternal deaths per 100 000 live births in 1992 to 1120 in 2000, according to the Malawi Demographic and Health Surveys. According to confidential enquiries into maternal deaths in health facilities in 1989 and 2001, three fac￾tors apparently contributed to this increase. First, there was a sharp proportional increase in deaths from AIDS. This is not surprising since Malawi’s national HIV prevalence has now reached 8.4%. Second, fewer mothers gave birth in health facilities: the proportion dropped from 55% to 43% between 2000 and (itself not independent from the HIV/AIDS epidemic). In remote areas one midwife often has to run the entire rural health centre and is expected to be available for work day and night, seven days a week. One maternity unit out of 10 is closed for lack of staff. Hospitals also experience severe shortages of midwives, and unskilled cleaners often conduct deliveries. The shortage of staff in maternity units is catastrophic and rapidly getting worse; the chances of Malawi women giving birth in a safe environment diminish accordingly. Box 1.3 A reversal of maternal mortality in Malawi

12 The World Health Report 2005 in all regions except sub-Saharan Africa during the 1990s.The proportion of births assisted by a skilled attendant rose by 24%during the 1990s,caesarean sections tripled and antenatal care use rose by 21%.Since professional care is known to be crucial in averting maternal deaths as well as in improving maternal health,maternal mortality ratios are likely to be declining everywhere except for those countries which started the 1990s at high levels.For these,which are mainly in sub-Saharan Africa, there has been no sign of progress. A PATCHWORK OF PROGRESS,STAGNATION AND REVERSAL The slowing down of improvement of global indicators that so worries policy-mak- ers (67)hides a patchwork of countries that are on track,show slow progress,are stagnating or are going into reverse.As most progress is being made in countries that already have relatively low levels of maternal and child deaths.while the worst-off stagnate,the gaps between countries are inevitably widening. A total of 93 countries,including most of those in the high income bracket,are "on track"to reduce their 1990 under-five mortality rates by two thirds by 2015 or sooner. The on-track countries are those that already had the lowest rates in 1990(taken together they had a rate of 59 in 1990) Box 1.4 Counting births and deaths If nobody keeps track of their births and deaths. ing maternal mortality.Ascertaining cause of in childbirth had a sister,that her sister is alive women and children simply do not count(67). death and relating it to pregnancy is difficult, to tell the tale,that she knows of her sister's Mortality rates are frequently only rough particularly where most deaths occur at home. death,and knows her sister's age and preg- estimates,of varying reliability.This is because Misclassified or undercounting is frequent in nancy status at death.As maternal deaths are the ways of estimating mortality are far from countries with fully functioning vital registra- statistically rare,it is difficult to obtain reports perfect and,in many cases,insufficient priority tion systems-between 17%and 63%(65) on enough deaths to estimate the maternal is given to obtaining such vital information. -let alone in those where such systems cover mortality ratio with sufficient precision and It is often assumed that the quoted numbers only part of the population. reliability without undertaking more expensive of matemal and child deaths rely on hospital Many developing countries where births studies such as a reproductive age mortality statistics.But apart from the problems of and deaths are not routinely counted conduct survey(60).The result is that levels and trends maintaining reporting systems,only a frac- sample surveys asking women for their "birth are often very difficult to interpret. tion of events takes place in facilities.Hospital histories"and how many of their children have In countries where registration is incomplete information is currently the most flawed source died,when and at what age.These surveys and where no survey has been conducted,the of data on births and deaths. yield estimates of child mortality.Often quite only remaining option for assessing mortality The best approach to estimating maternal robust,they can be biased or inaccurate when is to construct a modelled estimate.This is and child mortality is to count births and deaths the surveys are badly sampled and not repre- effectively an educated guess based on infor- through vital registration systems.In many sentative of the population at large.Information mation from similar or neighbouring countries. developing countries,however,such systems on a deceased child whose mother has died A total of 28 countries rely only on such esti- are still incomplete.The births and deaths that herself will simply not be gathered.Mothers mates for neonatal mortality,62 for maternal are registered under-represent the rural popula- often do not know exact dates of birth or may mortality.These modelled estimates should be tion and the socioeconomically disadvantaged be unwilling or unable to recall at what age a treated with great caution,but may be the only In 47 countries of the world,less than 50%of child has died.Completeness and accuracy information available. the population registers their deaths.A reliable very much depend on the skills and the cultural For the first time,this World Health Report neonatal mortality rate,for example,can there- sensitivity of the interviewer.Unfortunately. presents,separately,tables with country esti- fore be calculated for only 72 countries-less finding out about the quality of survey data in mates of mortality derived from surveys or vital than 14%of births in the world.Internationally the public domain is often not possible. registration,where these are available,and recommended definitions of what constitutes Maternal mortality is even more difficult to tables for all countries with country estimates a neonatal death are not always used(62,63). estimate from sample surveys.Information that have been modelled and adjusted.These The calculated rates,especially in central Asia, must be gleaned from relatives.Generally, estimates can be found in Annex Tables 2a. are therefore not always comparable across women are asked whether their sisters died 2b and 8. countries(64).Vital registration systems are during pregnancy or shortly afterwards(66) currently even less satisfactory for estimat- This presupposes that each woman who dies

12 The World Health Report 2005 in all regions except sub-Saharan Africa during the 1990s. The proportion of births assisted by a skilled attendant rose by 24% during the 1990s, caesarean sections tripled and antenatal care use rose by 21%. Since professional care is known to be crucial in averting maternal deaths as well as in improving maternal health, maternal mortality ratios are likely to be declining everywhere except for those countries which started the 1990s at high levels. For these, which are mainly in sub-Saharan Africa, there has been no sign of progress. A PATCHWORK OF PROGRESS, STAGNATION AND REVERSAL The slowing down of improvement of global indicators that so worries policy-mak￾ers (67) hides a patchwork of countries that are on track, show slow progress, are stagnating or are going into reverse. As most progress is being made in countries that already have relatively low levels of maternal and child deaths, while the worst-off stagnate, the gaps between countries are inevitably widening. A total of 93 countries, including most of those in the high income bracket, are “on track” to reduce their 1990 under-five mortality rates by two thirds by 2015 or sooner. The on-track countries are those that already had the lowest rates in 1990 (taken together they had a rate of 59 in 1990). ing maternal mortality. Ascertaining cause of death and relating it to pregnancy is difficult, particularly where most deaths occur at home. Misclassified or undercounting is frequent in countries with fully functioning vital registra￾tion systems – between 17% and 63% (65) – let alone in those where such systems cover only part of the population. Many developing countries where births and deaths are not routinely counted conduct sample surveys asking women for their “birth histories” and how many of their children have died, when and at what age. These surveys yield estimates of child mortality. Often quite robust, they can be biased or inaccurate when the surveys are badly sampled and not repre￾sentative of the population at large. Information on a deceased child whose mother has died herself will simply not be gathered. Mothers often do not know exact dates of birth or may be unwilling or unable to recall at what age a child has died. Completeness and accuracy very much depend on the skills and the cultural sensitivity of the interviewer. Unfortunately, finding out about the quality of survey data in the public domain is often not possible. Maternal mortality is even more difficult to estimate from sample surveys. Information must be gleaned from relatives. Generally, women are asked whether their sisters died during pregnancy or shortly afterwards (66). This presupposes that each woman who dies If nobody keeps track of their births and deaths, women and children simply do not count (61). Mortality rates are frequently only rough estimates, of varying reliability. This is because the ways of estimating mortality are far from perfect and, in many cases, insufficient priority is given to obtaining such vital information. It is often assumed that the quoted numbers of maternal and child deaths rely on hospital statistics. But apart from the problems of maintaining reporting systems, only a frac￾tion of events takes place in facilities. Hospital information is currently the most flawed source of data on births and deaths. The best approach to estimating maternal and child mortality is to count births and deaths through vital registration systems. In many developing countries, however, such systems are still incomplete. The births and deaths that are registered under-represent the rural popula￾tion and the socioeconomically disadvantaged. In 47 countries of the world, less than 50% of the population registers their deaths. A reliable neonatal mortality rate, for example, can there￾fore be calculated for only 72 countries – less than 14% of births in the world. Internationally recommended definitions of what constitutes a neonatal death are not always used (62, 63). The calculated rates, especially in central Asia, are therefore not always comparable across countries (64). Vital registration systems are currently even less satisfactory for estimat￾in childbirth had a sister, that her sister is alive to tell the tale, that she knows of her sister’s death, and knows her sister’s age and preg￾nancy status at death. As maternal deaths are statistically rare, it is difficult to obtain reports on enough deaths to estimate the maternal mortality ratio with sufficient precision and reliability without undertaking more expensive studies such as a reproductive age mortality survey (60). The result is that levels and trends are often very difficult to interpret. In countries where registration is incomplete and where no survey has been conducted, the only remaining option for assessing mortality is to construct a modelled estimate. This is effectively an educated guess based on infor￾mation from similar or neighbouring countries. A total of 28 countries rely only on such esti￾mates for neonatal mortality, 62 for maternal mortality. These modelled estimates should be treated with great caution, but may be the only information available. For the first time, this World Health Report presents, separately, tables with country esti￾mates of mortality derived from surveys or vital registration, where these are available, and tables for all countries with country estimates that have been modelled and adjusted. These estimates can be found in Annex Tables 2a, 2b and 8. Box 1.4 Counting births and deaths

mothers and children matter-so does their health 13 A total of 51 other countries are showing slower progress:the number of deaths among children under five years of age is going down and the mortality rates are drop- ping,but not fast enough to reach one third of their 1990 level by 2015 unless they significantly accelerate progress during the coming 10 years.These countries started from a somewhat higher level than those that are on track:an average under-five mortality rate of 92 per 1000. More problematic are the 29 countries where mortality rates are "stagnating"- where the number of deaths continues to grow,because modest reductions of mortal- ity rates are too small to keep up with the increasing numbers of births.These are the countries that had the highest levels(207 on average)in 1990.Finally,there are 14 "reversal"countries,where under-five mortality rates went down to an average of 111 in 1990 but have increased since.During the 1990s there were more such countries than during the two previous decades combined.These reversals were also more pronounced than before.Countries that show reversal or stagnation are overwhelm- ingly in the African Region. This grouping of countries,categorized according to progress in under-five mortality during the 1990s,roughly corresponds to what happened in terms of neonatal and maternal health in these same countries.Although trend data are not available,neo- natal and maternal mortality is highest in the countries with reversal and stagnation in under-five mortality (see Table 1.1 and Figures 1.3-1.6). THE NUMBERS REMAIN HIGH As the situation improves at a slower pace than expected-and hoped for-the gains in avoided deaths are partially offset by the demographic momentum.The numbers of untimely deaths of mothers and children could well be on the increase,because while rates are dropping.the numbers of mothers,births and children continue to grow. Worldwide.the number of live births will peak at 137 million per year towards 2015 (68):3.5 million more than at present.Most of the increase will be in sub-Saharan Africa and in parts of Asia-Pakistan and northern India-where the number of births will continue to grow well into the 2020s,even if fertility continues to drop.These are areas where the protection of adolescents and young women against early or unwant- ed pregnancy is most inadequate,mortality from unsafe abortion most pronounced, giving birth most hazardous and childhood most difficult to survive. Why is it still necessary for this report to emphasize the importance of focusing on the health of mothers and children,after decades of priority status.and more than 10 years after the United Nations International Conference on Population and Develop- ment?Progress has slowed down and is increasingly uneven,with a widening gap be- tween rich and poor countries as well as,often,between the poor and the rich within countries.The reasons for this patchy progress are examined in the next chapter. No data available for five countries

mothers and children matter – so does their health 13 A total of 51 other countries are showing slower progress: the number of deaths among children under five years of age is going down and the mortality rates are drop￾ping, but not fast enough to reach one third of their 1990 level by 2015 unless they significantly accelerate progress during the coming 10 years. These countries started from a somewhat higher level than those that are on track: an average under-five mortality rate of 92 per 1000. More problematic are the 29 countries where mortality rates are “stagnating” – where the number of deaths continues to grow, because modest reductions of mortal￾ity rates are too small to keep up with the increasing numbers of births. These are the countries that had the highest levels (207 on average) in 1990. Finally, there are 14 “reversal” countries, where under-five mortality rates went down to an average of 111 in 1990 but have increased since. During the 1990s there were more such countries than during the two previous decades combined. These reversals were also more pronounced than before. Countries that show reversal or stagnation are overwhelm￾ingly in the African Region. This grouping of countries,1 categorized according to progress in under-five mortality during the 1990s, roughly corresponds to what happened in terms of neonatal and maternal health in these same countries. Although trend data are not available, neo￾natal and maternal mortality is highest in the countries with reversal and stagnation in under-five mortality (see Table 1.1 and Figures 1.3–1.6). THE NUMBERS REMAIN HIGH As the situation improves at a slower pace than expected – and hoped for – the gains in avoided deaths are partially offset by the demographic momentum. The numbers of untimely deaths of mothers and children could well be on the increase, because while rates are dropping, the numbers of mothers, births and children continue to grow. Worldwide, the number of live births will peak at 137 million per year towards 2015 (68): 3.5 million more than at present. Most of the increase will be in sub-Saharan Africa and in parts of Asia – Pakistan and northern India – where the number of births will continue to grow well into the 2020s, even if fertility continues to drop. These are areas where the protection of adolescents and young women against early or unwant￾ed pregnancy is most inadequate, mortality from unsafe abortion most pronounced, giving birth most hazardous and childhood most difficult to survive. Why is it still necessary for this report to emphasize the importance of focusing on the health of mothers and children, after decades of priority status, and more than 10 years after the United Nations International Conference on Population and Develop￾ment? Progress has slowed down and is increasingly uneven, with a widening gap be￾tween rich and poor countries as well as, often, between the poor and the rich within countries. The reasons for this patchy progress are examined in the next chapter. ing maternal mortality. Ascertaining cause of death and relating it to pregnancy is difficult, particularly where most deaths occur at home. Misclassified or undercounting is frequent in countries with fully functioning vital registra￾tion systems – between 17% and 63% (65) – let alone in those where such systems cover only part of the population. Many developing countries where births and deaths are not routinely counted conduct sample surveys asking women for their “birth histories” and how many of their children have died, when and at what age. These surveys yield estimates of child mortality. Often quite robust, they can be biased or inaccurate when the surveys are badly sampled and not repre￾sentative of the population at large. Information on a deceased child whose mother has died herself will simply not be gathered. Mothers often do not know exact dates of birth or may be unwilling or unable to recall at what age a child has died. Completeness and accuracy very much depend on the skills and the cultural sensitivity of the interviewer. Unfortunately, finding out about the quality of survey data in the public domain is often not possible. Maternal mortality is even more difficult to estimate from sample surveys. Information must be gleaned from relatives. Generally, women are asked whether their sisters died during pregnancy or shortly afterwards (66). This presupposes that each woman who dies If nobody keeps track of their births and deaths, women and children simply do not count (61). Mortality rates are frequently only rough estimates, of varying reliability. This is because the ways of estimating mortality are far from perfect and, in many cases, insufficient priority is given to obtaining such vital information. It is often assumed that the quoted numbers of maternal and child deaths rely on hospital statistics. But apart from the problems of maintaining reporting systems, only a frac￾tion of events takes place in facilities. Hospital information is currently the most flawed source of data on births and deaths. The best approach to estimating maternal and child mortality is to count births and deaths through vital registration systems. In many developing countries, however, such systems are still incomplete. The births and deaths that are registered under-represent the rural popula￾tion and the socioeconomically disadvantaged. In 47 countries of the world, less than 50% of the population registers their deaths. A reliable neonatal mortality rate, for example, can there￾fore be calculated for only 72 countries – less than 14% of births in the world. Internationally recommended definitions of what constitutes a neonatal death are not always used (62, 63). The calculated rates, especially in central Asia, are therefore not always comparable across countries (64). Vital registration systems are currently even less satisfactory for estimat￾in childbirth had a sister, that her sister is alive to tell the tale, that she knows of her sister’s death, and knows her sister’s age and preg￾nancy status at death. As maternal deaths are statistically rare, it is difficult to obtain reports on enough deaths to estimate the maternal mortality ratio with sufficient precision and reliability without undertaking more expensive studies such as a reproductive age mortality survey (60). The result is that levels and trends are often very difficult to interpret. In countries where registration is incomplete and where no survey has been conducted, the only remaining option for assessing mortality is to construct a modelled estimate. This is effectively an educated guess based on infor￾mation from similar or neighbouring countries. A total of 28 countries rely only on such esti￾mates for neonatal mortality, 62 for maternal mortality. These modelled estimates should be treated with great caution, but may be the only information available. For the first time, this World Health Report presents, separately, tables with country esti￾mates of mortality derived from surveys or vital registration, where these are available, and tables for all countries with country estimates that have been modelled and adjusted. These estimates can be found in Annex Tables 2a, 2b and 8. Box 1.4 Counting births and deaths 1 No data available for five countries