Appendix A
Case Studies of Child Mortality

Data from national systems of vital statistics are incomplete and usually unpublished. Therefore, we have to rely on specific population-based surveys, a special analysis of vital registration data for small areas, or hospital records to estimate the structure of the causes of death among infants and children in sub-Saharan Africa.

This appendix includes summaries of 9 case studies that provide some data on the causes of death among infants and children. These studies are of very uneven quality and differ in the amount of details they provide. The data are drawn from three types of sources: vital statistics, hospital records, and population surveillance systems. Each of these sources has its limitations. Vital registration systems may record a large number of deaths, but if deaths are unattended or not certified by a physician, the reported cause of death may well be incorrect. Hospital data may be more accurate in determining the cause, but they are subject to a selection bias. Population surveillance systems generally have excellent coverage and consistent diagnosis of cause of death, but they are based on small populations and generally limited to a few years. Their results may therefore produce a good picture of a fairly small area and population, but it may be difficult to make generalizations based upon their findings. Where it is possible, these results also include the percentages of death attributed to other and undetermined causes.



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 155
Appendix A Case Studies of Child Mortality Data from national systems of vital statistics are incomplete and usually unpublished. Therefore, we have to rely on specific population-based surveys, a special analysis of vital registration data for small areas, or hospital records to estimate the structure of the causes of death among infants and children in sub-Saharan Africa. This appendix includes summaries of 9 case studies that provide some data on the causes of death among infants and children. These studies are of very uneven quality and differ in the amount of details they provide. The data are drawn from three types of sources: vital statistics, hospital records, and population surveillance systems. Each of these sources has its limitations. Vital registration systems may record a large number of deaths, but if deaths are unattended or not certified by a physician, the reported cause of death may well be incorrect. Hospital data may be more accurate in determining the cause, but they are subject to a selection bias. Population surveillance systems generally have excellent coverage and consistent diagnosis of cause of death, but they are based on small populations and generally limited to a few years. Their results may therefore produce a good picture of a fairly small area and population, but it may be difficult to make generalizations based upon their findings. Where it is possible, these results also include the percentages of death attributed to other and undetermined causes.

OCR for page 155
SIERRA LEONE (WESTERN AREA) Infant and child mortality in Sierra Leone is among the highest in the world, with the infant mortality rate (IMR) between 1980 and 1985 estimated at 166 per 1,000 live births (United Nations, 1991). Kandeh (1986) conducted an analysis of the causes of infant and childhood deaths in the Western Area of the country where the main hospitals for mothers and children are located. Of the 3,783 infant deaths recorded in the vital registration system of the Western Area between 1969 and 1971, only 1,772 (47 percent) were medically certified. Of these deaths, 33 percent (582) occurred during the first week of life, 27 percent (483) between the second and fourth weeks, and the remaining 40 percent (707) during the second through eleventh months. As shown in Table A-1, neonatal mortality was dominated by tetanus (41 percent), which was also the leading cause of death in infancy. This cause was followed by ''anoxic and hypoxic conditions," which accounted for 18 percent of neonatal deaths. Both conditions are closely associated with poor care of the delivery and of the newborn child. Other important causes of neonatal death were pneumonia and other acute respiratory infections (ARIs) (6 percent), birth infection (5 percent), congenital anomaly (3 percent), and septicemia (1 percent). Postneonatal mortality was due primarily to pneumonia (30 percent) and other ARIs (5 percent), diarrhea (21 percent), measles (9 percent), and malaria (6 percent). Malnutrition per se was not coded, but may be included among avitaminosis (3 percent) and anemia (4 percent), although this last category probably reflects malaria mortality as well. Medically certified deaths of children between ages 1 and 4 years were attributed to measles (22 percent), pneumonia (22 percent), diarrhea (13 percent), avitaminosis (10 percent), anemia (10 percent), malaria (4 percent), tuberculosis (3 percent), meningitis (2 percent), accidents (1 percent), dysentery (1 percent), and other causes (10 percent). This study covered a large population and included both an urban and a rural area. The results are similar to other studies (see Table 2-3 for comparison). However, a few features limit the comparability with other studies: Severe malnutrition was not coded as a cause of death; pertussis was either neglected or included in the "other ARI" category; malaria played a smaller role than elsewhere, which may be due to reporting biases or to the local environment; diarrhea seemed to have been underestimated; and some categories were unconventional (such as anoxic and hypoxic conditions).

OCR for page 155
TABLE A-1 Ranking of Major Underlying Causes of Death in Sierra Leone (western area) by Age Group, 1969-1979   Neonatal (N = 1,065), 1969-1971 Postneonatal (N = 732), 1969-1971 Children, Aged 1-4 (N = 924), 1974-1976 Rank Cause of Death Percentage Cause of Death Percentage Cause of Death Percentage 1 Tetanus 41.1 Pneumonia 30.4 Measles 22.3 2 Anoxic/hypoxic 18.3 Diarrhea 21.2 Pneumonia 22.2 3 Pneumonia 5.1 Measles 9.1 Diarrhea 13.1 4 Birth infection 4.6 Malaria 6.0 Avitaminosis 10.3 5 Congenital anomalies 3.3 Other ARI 5.0 Anemia 9.7 6 Diarrhea 1.6 Anemia 4.2 Malaria 4.6 7 Septicemia 1.4 Avitaminosis 2.9 Tuberculosis 3.1 8     Tetanus 2.4 Meningitis 2.2 9     Septicemia 1.6 Accidents 1.3 10     Congenital anomalies 0.9 Dysentery 0.8   SOURCE: Kandeh (1986).

OCR for page 155
MACHAKOS, KENYA The Machakos area of Kenya (1975-1978) had a much lower mortality rate than other areas of Africa, which may imply a different cause-of-death structure. Data were collected in this rural area between 1975 and 1978. Omondi-Odhiambo et al. (1990) report that in the neonatal period, after asphyxia and prematurity, which accounted for almost half of the deaths, the leading cause of death was ARI, which was associated with 18 percent. The cause of 15 percent of the neonatal deaths was unknown. During the postneonatal period, the leading causes were intestinal infections and pneumonia (40 and 28 percent, respectively), followed by measles (14 percent). Only 7 percent were reported as unknown cause. Among children ages 1 to 4 years, measles was responsible for 32 percent of all deaths. Nutritional deficiencies were the cause of 17 percent of all child deaths. Pneumonia and diarrhea were listed as causing 13 and 14 percent of deaths, respectively. Pneumonia and diarrhea were more important among infants than among children aged 1-4 years, whereas measles and malnutrition were more often the causes of death of children; 4 percent were of unknown cause. NATIONAL ESTIMATES FOR KENYA Ewbank et al. (1986) used data from 1976 to 1987 on hospital inpatient deaths, registered deaths by district, and results of available epidemiologic data (including the Machakos study) to estimate the cause-of-death structure for Kenya. They attempted to adjust for the selectivity of hospital deaths and for the fact that both the coverage of registered deaths and the distribution of deaths by cause differ among districts. Their estimates of the number of deaths by cause for children under age five are given in Table A-2. In addition, they estimated that there were about 10,900 deaths at all ages due to malaria. Because a large proportion TABLE A-2 Under-5 Cause-of-Death Structure of Kenya Cause of Death Estimated Number of Deaths Percent Respiratory 26,600 50.0 Measles 14,600 27.5 Diarrhea 9,300 17.5 Neonatal tetanus 2,200 4.2 Pertussis 400 0.8 Total 53,100 100.0   SOURCE: Ewbank et al. (1986).

OCR for page 155
of malaria deaths occur to children under age 5, malaria would rank about fourth on this list. DAKAR AND SAINT-LOUIS, SENEGAL Cantrelle et al. (1986) studied vital registration data from the cities of Dakar and Saint-Louis, between 1973 and 1980. The quality of the data on causes of death was to a large extent determined by the place of death. When the death occurred in a health unit, as was the case for 68 percent of the deaths in Dakar and 58 percent in Saint-Louis, the cause of death was more precise because it was generally established by a physician. When the death occurred at home, a public official was left to determine the cause. In Dakar, congenital disorders and perinatal diseases were the cause of 25 percent of all infant deaths; in Saint-Louis, these causes were responsible for 33 percent of all infant deaths. The percentages varied over the eight-year period, with a range of 19 to 30 percent observed in Dakar and 23 to 48 percent in Saint-Louis. In Dakar, the second leading cause of infant mortality was diarrhea, accounting for 11.6 percent of deaths. In Saint-Louis, malnutrition and dehydration (combined into one category) were listed as the second leading cause of infant death, with 17 percent. Measles was the third leading cause in both cities, accounting for 10 percent in Dakar and 7 percent in Saint-Louis. Of infant deaths in Dakar, 6 percent were attributed to malnutrition and dehydration, followed by bronchopulmonary disorders (6 percent). Diarrhea, the fourth leading cause of infant mortality in Saint-Louis, was responsible for 5 percent of the deaths, followed by bronchopulmonary disorders (5 percent). The remaining deaths were attributable to other causes or could not be determined. For children, the leading cause of deaths in both cities was measles. In Dakar, measles was listed as the cause of 28 percent of the deaths, whereas in Saint-Louis, it was responsible for 21 percent. Diarrheal and intestinal diseases were the second leading cause of death in Dakar (12 percent), and the third major cause in Saint-Louis (10 percent). Malnutrition and dehydration were the second major cause of child death in Saint-Louis, accounting for 18 percent. Malaria was also a major cause of death in Saint-Louis, with 9 percent, whereas in Dakar, it was much less important, with only 2 percent. On the other hand, bronchopulmonary disorders accounted for 9 percent of the child deaths in Dakar, but only 4 percent in Saint-Louis. Cantrelle et al. (1986) pointed out that ecological conditions should be considered when examining trends. A major drought in Saint-Louis in 1972 and 1973 probably reduced food production, thus contributing to the high percentage of deaths from malnutrition and dehydration. Moreover, heavy rainfall in Saint-Louis during 1975 corresponded to an increase in deaths from malaria, a cause that was relatively unimportant in Dakar. With respect

OCR for page 155
to infant mortality, they indicated that with the exception of measles, other major causes of death were often associated with the consequences of low rainfall and drought, such as low crop production and reduced purchasing power. NIAKHAR, SENEGAL Garenne and Fontaine (1990) analyzed 808 verbal autopsies recorded by a demographic surveillance system in Niakhar, a rural area of Senegal. All deaths of all ages occurring in 30 villages were studied by using a comprehensive questionnaire. Results for children below age 5 for the years 1983-1989 are shown in Table A-3 (M. Garenne, personal communication, 1992). Two causes accounted for more than half of all neonatal deaths: neonatal tetanus and low birthweight. The few cases of malnutrition involved newborns who were improperly nourished because their mothers died shortly after delivery. For postneonatal mortality, diarrhea and acute respiratory infections accounted for half of all deaths. Pertussis accounted for more deaths than measles, because this disease strikes soon after birth, whereas cases of measles are extremely rare before 4 months of age. For children aged 1-4 years, diarrhea was still the most common cause of death. Malaria and acute malnutrition were more important than ARI. Measles preceded pertussis and cholera, which was very rare among infants. Hepatitis and tuberculosis were probably underestimated, because of the less typical symptoms that are harder to identify by using verbal autopsies. Deaths due to typhoid and congenital syphilis were never diagnosed, but do exist in the study area. In this study, malnutrition included acute malnutrition (kwashiorkor and marasmus) and three cases of anemia, which may have been caused by malaria. An independent study of risk factors showed that two-thirds of the deaths outside the neonatal period were attributable to poor nutritional status. The diarrhea category for postneonates included acute watery diarrhea (16 percent), persistent diarrhea (12 percent), and dysentery (2 percent); and for children 1-4, acute watery diarrhea (9 percent), persistent diarrhea (16 percent), and dysentery (3 percent). BAMAKO, MALI A study of the vital statistics collected between 1974 and 1985 in Bamako, Mali, provides information on cause-of-death structure (Fargues and Nassour, 1988). Registration of deaths is virtually complete in Bamako because a death certificate from the government is needed for burial. If death occurs in a hospital, the cause of death is noted by the attending physician or

OCR for page 155
TABLE A-3 Ranking of Major Underlying Causes of Death in Niakhar, Senegal, 1983-1989, by Age Group   Neonatal (N = 407) Postneonatal (N = 461) Children Aged 1-4 (N = 1,015) Rank Cause of Death Percentage Cause of Death Percentage Cause of Death Percentage 1 Tetanus 39.1 Diarrhea 29.9 Diarrhea 27.4 2 Low birth weight 28.0 ARI 19.5 Malaria 12.0 3 Pneumonia 4.9 Malaria 6.9 Malnutrition 8.5 4 Birth trauma 2.2 Pertussis 6.1 ARI 8.3 5 Birth defect 1.7 Malnutrition 5.4 Measles 8.0 6 Diarrhea 1.0 Measles 3.7 Pertussis 4.2 7 Malnutrition 0.5 Meningitis 2.7 Cholera 3.7 8     Septicemia 1.3 Meningitis 2.0 9     Varicella 1.0 Varicella 0.5 10         Septicemia 0.4 11         Epilepsy 0.4 12         Hepatitis 0.3 13         Tuberculosis 0.3   SOURCE: M. Garenne, personal communication (1992).

OCR for page 155
nurse. However, for a death that occurs at home, the certificate may not accurately reflect the true cause of death. Some causes of death such as measles are easily recognizable to individuals reporting the death, but other causes such as dehydration may be reported incorrectly as diarrhea. Similarly, many people are aware of fever as a symptom of malaria and report deaths as being caused by malaria when a high fever is present. The leading causes of infant mortality were conditions related to the perinatal period (prematurity), which accounted for 24 percent of the deaths. Following that, 12 percent were attributed to malaria, whereas measles was responsible for 10 percent. The fourth leading cause was intestinal infection, accounting for 9 percent. The other major causes of death were meningitis (4 percent), malnutrition (4 percent), dehydration (4 percent), and pneumonia (3 percent). The remaining causes of death either accounted for less than 1 percent of the total, and approximately 28 percent were classified as "other." Among children of ages 1 to 5, measles was reported as the cause of 34 percent of deaths, followed by malaria with 16 percent. In this group, 12 percent was attributed to malnutrition, whereas intestinal infection accounted for 8 percent. Pneumonia and dehydration each accounted for 3 percent of the deaths; anemia and meningitis, 1 percent each. Approximately 20 percent of deaths to children ages 1 to 5 years were attributed to other causes. MALUMFASHI, NIGERIA Data collected by Tomkins et al. (1991) between 1977 and 1978 in the Malumfashi area of Nigeria yielded basic cause of death information. During this period, 111 infants died, resulting in an infant mortality rate of 88 deaths per 1,000 live births. Among these infants, 12 percent were reported to have died of measles, 23 percent of pyrexia, and 13 percent of diarrhea. Among children ages 1 to 4, a similar pattern was observed. With a child mortality rate of 34 deaths per 1,000 (137 deaths), 12 percent of the deaths were ascribed to measles, 24 percent to pyrexia, and 17 percent to diarrhea. FOUR PROVINCES OF SUDAN Verbal autopsies provide information on the deaths of infants and children in a study in Greater Khartoum and the Blue Nile, Kassala, and Kordofan provinces of the Sudan between 1974 and 1976 (Sudanese Ministry of Health and World Health Organization, 1981). It appears that infant mortality was probably underreported. Among infants, 48 percent of the deaths were attributed to diarrhea and 12 percent to fevers. Measles was responsible for 11 percent of infant deaths, followed by pneumonia (9 percent), malaria (5 percent), and asphyxia (4 percent). The remaining 11 percent of deaths

OCR for page 155
were attributed to other and undetermined causes. The leading causes of death among children were diarrhea (46 percent), measles (18 percent), malaria (7 percent), pneumonia (6 percent), and fever (6 percent); 18 percent were attributed to other and undetermined causes. KASONGO, ZAIRE The Kasongo study was conducted in Zaire between 1974 and 1977 (van Lerberghe, 1989). The survey gathered information on 229 deaths of children under age 5, and deaths were classified into five groups: rash, respiratory disease, diarrheal disease, other diseases, and unknown. Those classified as rashes were attributed to measles if the child had a rash that started less than 30 days before the death. Respiratory deaths included those in which the mother attributed the death to a respiratory illness (cough, dyspnea, or whooping cough), dyspnea or cough were reported at death, or respiratory illness was reported between the preceding visit and death. Deaths were ascribed to diarrhea when they were attributed to diarrhea by the mother or when there was mention of diarrhea at death if there was no association with rash, respiratory illness, or any other specific cause of death. Among the deaths of infants under 5 months of age, 40 percent were of unknown origin and 30 percent were classified as "other" causes. Respiratory illnesses were the leading identifiable cause, representing about 20 percent of all deaths. Diarrhea was reported as the cause of about 7 percent, and rash about 3 percent. Among older children, rash (measles) was the leading cause of death: approximately 45 percent of the deaths among children 6 to 11 months, and 75 percent among children 12 to 17 months. In the groups aged 18 to 59 months, rashes were the largest single cause, ranging from about 60 to 35 percent, respectively. Respiratory illnesses and diarrhea each caused about 20 percent of the deaths among infants between 6 and 11 months, and only about 5 percent each in the age group 12 to 17 months. Diarrheal diseases accounted for about 10 percent of the deaths between ages 18 and 47 months, but were not cited as the cause of death in the group aged 48 to 59 months. Respiratory illnesses were reported to be the cause of about 20 percent of the deaths among children 24-35 months and 48-59 months. However, respiratory illnesses were not listed as the cause of any deaths among children 36-47 months. Among children 6 to 59 months, unknown causes of death ranged from approximately 2 to 15 percent of the total of registered deaths.