The following is an archived copy of a message sent to a Discussion List run by the Campaign Against Sanctions on Iraq.
Views expressed in this archived message are those of the author, not of the Campaign Against Sanctions on Iraq.
[Main archive index/search] [List information] [Campaign Against Sanctions on Iraq Homepage]
Just a reminder -- The observer likes you to put 'Letter to the Editor' in the subject field of emailed letters. Also, a quote from the Lancet that might be useful (Sweeney implies that the Lancet rubbished UNICEF's mortality survey). ------------------------------------------------------------------- Lancet Volume 355, Number 9218 27 May 2000 "Assessment of data quality... The quality of data was undertaken by an independent panel which reviewed both the procedures used and the quality of various aspects of data. No problems were detected. " --------------------------------------------------------------------- Full text of EXACTLY what they said about it is pasted below for those who are interested. (the tables don't come out well in plain text, but there is plenty of useful information.) Cheers, Glenn. ---------------------------------------------------------------------- Articles Volume 355, Number 9218 27 May 2000 Sanctions and childhood mortality in Iraq Lancet 2000; 355: 1851 - 1857 Download PDF (120 Kb) Mohamed M Ali, Iqbal H Shah Department of Epidemiology and Population Health, London School of Hygiene and Tropical Medicine, University of London, UK (M M Ali MSc); and World Bank Special Programme of Research, Development and Research Training in Human Reproduction, Department of Reproductive Health and Research, World Health Organization, Geneva, Switzerland (I H Shah PhD) Correspondence to: Mohamed M Ali, Centre for Population Studies, London School of Hygiene and Tropical Medicine, 49-51 Bedford Square, London WC1B 3DP, UK (e-mail: Mohamed.Ali@lshtm.ac.uk) Summary Introduction Methods Assessment of data quality Results Discussion References Summary Background In 1999 UNICEF, in cooperation with the government of Iraq and the local authorities in the "autonomous" (northern Kurdish) region, conducted two similar surveys to provide regionally representative and reliable estimates of child mortality (the subject of this paper) and maternal mortality. Methods In a cross-sectional household survey in the south/centre of Iraq in February and March, 1999, 23 105 ever-married women aged 15-49 years living in sampled households were interviewed by trained interviewers with a structured questionnaire that was developed using the Demographic and Health Surveys questionnaire and following a pre-test. In a similar survey in the autonomous region in April and May 14 035 ever-married women age 15-49 were interviewed. Findings In the south/centre, infant and under-5 mortality increased during the 10 years before the survey, which roughly corresponds to the period following the Gulf conflict and the start of the United Nations sanctions. Infant mortality rose from 47 per 1000 live births during 1984-89 to 108 per 1000 in 1994-99, and under-5 mortality rose from 56 to 131 per 1000 live births. In the autonomous region during the same period, infant mortality declined from 64 to 59 per 1000 and under-5 mortality fell from 80 to 72 per 1000. Childhood mortality was higher among children born in rural areas, children born to women with no education, and in boys, and these differentials were broadly similar in the two regions. Interpretation Childhood mortality clearly increased after the Gulf conflict and under UN sanctions in the south/centre of Iraq, but in the autonomous region since the start of the Oil-for-Food Programme childhood mortality has begun to decline. Better food and resource allocation to the autonomous region contributed to the continued gains in lower mortality, whereas the situation in the south/centre deteriorated despite the high level of literacy in that region. Lancet 2000; 355: 1851-57 See Editorial Introduction During the past 20 years Iraq has witnessed spectacular social and economic development, followed by a dramatic decline. The per capita gross domestic product (GDP), for example, was estimated at US$3510 in 1989, but only US$450 in 1996.1 This was primarily due to an 85% reduction in oil production and the devastation of industrial and service sectors of the economy.2 Before 1991, much progress has been made in building roads and infrastructure as well as improving human skills by expansion of education and advanced training. During the same period, healthcare reached about 97% of the urban and 79% of rural population.3 The healthcare system was based on an extensive and expanding network of health facilities linked up by reliable modes of transport and communications. The country had a well-developed water and sanitation system and 90% of the population was estimated to have access to safe drinking water.4 Infant mortality fell from 71 per 1000 live births in 1965 to 29 per 1000 in 1989 while under-5 mortality declined from 111 to 44 per 1000 live births.5,6 These gains were stopped with the start of the Gulf conflict. Trade sanctions were imposed against Iraq in August, 1990. The war started on Jan 16, 1991, and a ceasefire was declared after 6 weeks. After the Gulf War, the autonomous region of Iraq, inhabited mainly by Kurds, was placed under United Nations administration and sanctions were strictly enforced against Iraq. Child mortality is a critical measure of the wellbeing of children and a good proxy indicator of the overall level of development. Immediately after the Gulf conflict an international study team did an extensive Iraq-wide mortality and nutrition survey7 and found a three-fold increase in under-5 mortality from 1985 to the first half of 1991.8 However, since 1991 there has been no countrywide child-mortality survey, and the mortality levels have been the source of considerable speculation and debate. Recent malnutrition surveys in Iraq have, however, shown that the prevalence of underweight under-5 children has increased two-fold since 1991--ie, from 12% to 23%.9 Since an increase in malnutrition is usually associated with increased childhood mortality, it is likely that mortality has also increased. Figure 1: Map of Iraq The United Nation Children's Fund (UNICEF), in partnership with the Government of Iraq and the local authorities in Al-Suleimaniyah, Dohouk, and Erbil (figure 1), did two parallel regionally representative household surveys in the south/centre and autonomous governorates of Iraq. The main purpose was to measure the levels, trends, and differentials in childhood and maternal mortality. Given the controversy surrounding previous studies,10,11 considerable care was taken in the design and implementation of these two surveys so that they could yield regionally representative and reliable estimates. Although data were not collected on the direct causes of childhood mortality, trends should throw light on the impact of a deteriorating economic and health situation in Iraq, following the Gulf War. The maternal mortality survey will be reported elsewhere. Methods Survey design The two surveys were designed to provide estimates of overall childhood mortality rates by key socioeconomic and demographic characteristics of respondents. Samples of 24000 households for the south/centre and of 16000 households for the Autonomous region were deemed necessary to estimate reliably the child mortality indices. The 1997 census, which covered only the south/centre, and the 1997 Directorate of Reconstruction and Development (DRD)/UNICEF survey, which covered the three autonomous governorates, were used as the sampling frame. A three-stage, stratified, self-weighting sampling design was used over all 79 districts in the south/centre and all 23 districts in the autonomous region. The sampling frame provided a complete listing of enumeration areas together with estimates of population size needed for the selection of the first two sampling stages. The sample design called for the selection of all districts, and then the use of a three-stage stratified design in each district (Qada'a). In the first stage, two subdistricts (Nahiya) from each district were selected by using probability proportion-to-size (PPS). Subdistricts were stratified into urban and rural parts and divided into clusters (Majal) of households. At the second stage, clusters were selected systematically with PPS from each urban and rural stratum separately. The PPS procedure was based on the numbers of households in each cluster recorded in 1997. Within each selected cluster, 15 households were selected from the address list created by the sampling frame by systematic random sampling. All ever-married women aged 15-49 years, identified during the household interview, were selected for individual interview, using a structured questionnaire. Questionnaire design and pre-test The questionnaire was designed to collect data on child and maternal mortality, and little additional information was sought from respondents beyond that needed to derive the mortality estimates. The questionnaire drew upon the worldwide experience of measuring childhood and maternal mortality from household surveys, by using relevant parts of the Demographic and Health Surveys (DHS) core questionnaire12 and was pre-tested. Note that the DHS programme has been doing household cross-sectional surveys since 1985, providing information from more than 50 developing countries, on family planning, maternal and child health, child survival, AIDS and sexually-transmitted diseases, and reproductive health. Training and fieldwork Both supervisors and interviewers were trained, using material from the DHS and the Gulf Child Health Survey of Iraq, 1989.6 Survey staff were recruited from the health sector. Supervisors were senior medical officers, and interviewers were recent graduates from medical schools or junior doctors. Fieldwork was done between February and March, 1999, in the south/centre and between April and May, 1999, in the autonomous region. In the south/centre 186 staff, including 16 governorate supervisors and 170 interviewers were responsible for data collection. More than 50% of the supervisors and team leaders, and all interviewers were women. In the autonomous region there were 37 supervisors and team leaders; of the 168 interviewers, 80% were women. Considerable care was taken to achieve a good coverage and high quality data. Team leaders were responsible for ensuring that all sample households were contacted, all eligible women interviewed, and responses completed for all relevant mortality questions. If inconsistencies were found the interviewer was asked to resolve them, going back to the household, if necessary. The supervisor again checked sample household coverage and also did checks on the internal consistency of responses. To monitor any changes to the responses in the original questionnaire, the interviewers had to enter responses in ballpoint pen, so that any subsequent change would be obvious. Any such change had to be initialled by the field staff person making the change. Editing and checking was done at the central level (Baghdad) for completeness and consistency. Data were available at the UNICEF office where most of the work on data editing, checking, and analysis was done. Analysis Similar to the approach used in DHS,13 childhood mortality estimates presented in this paper are based on the information provided in the birth history section of the questionnaire. Questions on the number of sons and daughters living with the mother, the number who live elsewhere, and the number who have died were asked first. These questions were then followed by a retrospective birth history in which the respondent was asked to list each of her live births, beginning with the first one. Information was collected in the birth history on the sex, month, and year of birth, survivorship status, and current age or age at death for each live birth. Synthetic cohorts were created from the information in the birth history on the child's date of birth, survivorship status and the reported age at death or current age. Life-table method was then used to estimate the mortality rates. Standard errors were estimated using Greenwood's formula14 and were adjusted for the design effect using jack-knife repeated replications.15 The following rates were derived: infant mortality rate (IMR)--the cumulative probability of dying between birth and exact age 1 year per 1000 live births; under-5 mortality rate (U5MR)--the cumulative probability of dying between birth and exact age 5 years per 1000 live births. The relative risk of dying by age 1 year and by age 5 years were estimated by logistic regression models for survey data that take into account the survey design (stratification and clustering) in the estimation of the relative risks and their standard errors. Assessment of data quality Achieving high quality data was a primary goal and several steps were taken to ensure that the data collected would yield reliable estimates of childhood mortality. Age misreporting, selective omission of births and/or deaths by time period, age or sex and errors in timing of births or deaths are especially important. The quality of data was undertaken by an independent panel which reviewed both the procedures used and the quality of various aspects of data. No problems were detected. Age reporting Many surveys show that respondents tend to report ages, which are multiple of five or sometimes of two. However, in these surveys in Iraq the distribution of respondent's current age showed no clustering of multiples of five or two; nor did the distribution of all live births, last birth, and first birth (aged 0-119 months). Omission of live births A commonly observed error in survey data, is the omission of children ever born or those who died, resulting in decreases in parity or decreases in the proportion dead in the oldest age-group. Because the births to older women involve an extended recall period and that their infant or children might have had excessive mortality making it difficult to remember the number of all births. However, no such pattern in average parity by age was noted. The selective omission of births and deaths, by age of women, was evaluated for omission of live births or by deaths. The distribution of births by calendar years for living, dead, and all children, by sex ratio at birth, and ratio of births by calendar year were examined. The result showed no systematic omission of children of a particular sex or of deaths of boys or girls by calendar period. Displacement of live births More serious errors in timing lie in the systematic displacement of birth dates by the mother nearer to, or further from, the time of the survey. Systematic errors in dating births have a characteristics pattern, concentrating births in the period 5-14 years before the survey, and perhaps even from the most recent 5-year period. To search for a distortion of this sort, consistency checks were made, including a simple assessment of births and deaths by period. Birth history A retrospective birth history is susceptible to several possible data collection errors. First, only surviving ever-married women of ages 15-49 were interviewed; therefore, no data were available for children of women who died. The resulting mortality estimates will be biased if the fertility of surviving women and non-surviving women differs substantially. The extent of this bias is not known but is not considered to be a serious source of bias. Another problem concerns the fact that the mortality estimates are based only on those births reported by ever-married women of reproductive age at a given point in time (ie, the time of the survey) and these are truncated because women older than 49 years are not interviewed. As this time period extends, the resulting censoring of information becomes progressively more severe, and the high rates of childhood mortality are usually associated with more advanced maternal age. Therefore, results presented for more than 15 years before the survey should be interpreted with caution. However, child mortality rates more than 15 years before the survey were compared with published results from surveys done in the 70s and the 80s,5,6,16 and the two rates were very similar, reassuring the quality of survey data. The Brass method of indirectly estimating child mortality18 is commonly used to check the quality of data and for comparison with estimates derived from life-tables. This approach, which is based on the number of children ever born and children died by duration of marriage, was also applied. The indirect estimates of infant, and under-5 mortality rates were identical to the direct estimates derived from the birth history. Results The intensive field supervision, household call-backs, short questionnaire, and well-organised training and field operations all contributed to a successful implementation and high coverage of the two surveys. In the south/centre, of 23 978 sampled households, 23920 (99·8%) were successfully covered, 21 048 had at least one ever-married woman age 15-49 years, and all these eligible respondents were interviewed, resulting in 23 105 completed questionnaires. In the autonomous region 15942 (96·6%) households were successfully covered of 16499 sampled. 13109 had at least one eligible respondent yielding 14 035 completed questionnaires. Background characteristics of respondents 14943 (64·7%) of the 23105 women interviewed in the south/centre and 11807 (84·1%) of the 14035 women in the autonomous region were living in urban areas at the time of the survey (table 1). Most (93·0% in the south/centre and 94·0% in the autonomous region) of the ever-married women were currently married. As is typical in the middle east region, a high proportion (61·0%) of women were married to a cousin or other relative in the south/centre, but less than half in the autonomous region. In the south/centre, about one-third of the respondents never attended school, less than half (48%) had some primary education, and 22% had more than primary education. In the autonomous region, more than half of the respondents never attended school, around one-third had some primary education, and only 1656 (12%) had more than primary education. The mean numbers of children ever born was 4·3 children per ever-married women in the south/centre and 4·6 in the autonomous region. South/centre Autonomous region n (%) n (%) Place of residence Urban 14943 (64·7) 11807 (84·1) Rural 8162 (35·3) 2228 (15·9) Regions of South/centre North 3401 (14·7) Middle 10947 (47·4) Middle Euphrates 4541 (19·7) South 4216 (18·2) Autonomous governorates Erbil 4839 (34·5) Dohouk 3500 (24·9) Al-Suleimaniyah 5696 (40·6) Current age (years) 15-19 1379 (6·0) 909 (6·5) 20-24 3582 (15·5) 2386 (17·0) 25-29 4798 (20·8) 2966 (21·1) 30-34 4323 (18·7) 2338 (16·7) 35-39 3598 (15·6) 1944 (13·9) 40-44 2895 (12·5) 1919 (13·7) 45-49 2530 (11·0) 1573 (11·2) Relation to last/current husband Cousin 8666 (37·5) 3524 (25·1) Other relative 5351 (23·2) 3083 (22·0) Not related 9088 (39·3) 7428 (52·9) Education No education 6811 (29·5) 8026 (57·2) Primary* 11140 (48·2) 4353 (31·0) Intermediate or higher 5154 (22·3) 1656 (11·8) Total 23105 (100·0) 14035 (100·0) *Includes respondents with no certificates. Table 1: Distribution of respondents by background characteristics and region Childhood mortality rates Table 2 shows the number of children born and the number who died at age 1 and age 1-5 years for different 5-year periods before the survey, by region. The table shows that most deaths happen in the first year of life (85% in south/centre and 80% in autonomous region). South/centre Autonomous region Number born Number died Number born Number died <1 year 1-5 years <1 year 1-5 years 1994-99 27194 2852 366 15061 798 90 1989-94 25750 2135 443 14992 1061 240 1984-89 19003 869 193 13559 861 223 1979-84 14416 742 163 10496 771 235 1974-79 8665 469 108 6506 686 271 Table 2: Birth and deaths before age 1 year and between 1 and 5 years The infant mortality rate (IMR) and the under-5 mortality rate were estimated for 5-year periods before the survey (figure 2). In the south-centre, both IMR and under-5 mortality rate consistently show a major increase in mortality over the 10 years preceding the survey (which roughly corresponds to the period following the Gulf conflict). The IMR more than doubled from 47 (95% CI 44-50) for the period roughly equal to 1984-89 to 108 (104-112) deaths per 1000 live births for the period roughly equal to 1994-99. Note that by 1984-89, the IMR in the south/centre had declined from 57 (52-62) deaths per 1000 live births during the period roughly corresponding to 1974-79. The IMR was, however, never as high in the previous years as in the 5 years just before the survey. Under-5 mortality rate also increased during the same time from 56 (52-58) to 131 (127-135) deaths per 1000 livebirths. In a short span of 10 years following the Gulf conflict, childhood mortality in the south/centre thus increased more than two-fold and to levels above those of 20-24 years previously. FIgure 2: Infant mortality rates (top) and under-5 mortality rates (bottom) by 5-year period preceding the survey A different picture, however, emerges for the autonomous region. During the same period, both the infant and the under-5 mortality rate fell, although a rise was witnessed for 1989-94 (5-9 years preceding the survey). More specifically, IMR declined from 64 (60-68) in 1984-89 to 59 (55-63) deaths per 1000 live births in 1994-99 and the under-5 mortality rate dropped during the same period from 80 (75-85) deaths to 72 (68-76) deaths per 1000 live births. Unlike the pattern in the south/centre, childhood mortality rates declined in the autonomous region to rates nearly half those that existed 20-24 years ago. Differentials in childhood mortality Differentials in childhood mortality are commonly used in documenting socioeconomic, demographic and environmental disparities. Table 3 shows IMR and under-5 mortality with their 95% CIs, for the most recent 10-year period (1989-99), by place and region of residence, mother's education, mother's age at birth, relation to husband, and sex of child. In the last 10 years before the survey, the overall IMR and under-5 mortality are significantly higher in the south/centre than in the autonomous region, with 47% difference in IMR and about 42% in under-5 mortality rate. IMR Under-5 mortality rate South/centre rate (95% CI) Autonomous region rate (95% CI) South/centre rate (95% CI) Autonomous region rate (95% CI) Overall 94·1 (91-96·5) 63·6 (60·9-66·4) 112·0 (109·3-114·7) 78·9 (75·8-81·9) Place of residence Urban 88·0 (84·9-91·2) 61·5 (58·5-64·6) 103·0 (99·6-106·4) 75·6 (72·2-78·9) Rural 103·2 (99·2-197·3) 72·7 (65·9-79·4) 125·8 (121·4-130·3) 92·9 (85·3-100·4) South/centre regions North 96·0 (89·7-102·3) ·· 113·0 (103·2-119·8) ·· Middle 90·9 (87·2-94·5) ·· 106·1 (102·2-110·1) ·· Middle Euphrates 94·4 (88·8-100·0) ·· 113·0 (106·9-119·0) ·· South 100·5 (94·7-106·2) ·· 125·1 (118·7-131·5) ·· Autonomous givernorates Erbil ·· 65·9 (61·0-70·7) ·· 79·7 (74·5-85·0) Dohouk ·· 68·6 (63·3-73·9) ·· 87·6 (81·7-93·5) Al-Suleimaniyah ·· 57·5 (53·1-61·9) ·· 71·0 (66·2-75·9) Schooling Never attended 100·5 (96·0-105·0) 69·2 (65·4-72·9) 121·4 (116·5-126·3) 86·6 (82·5-90·7) Primary 97·0 (93·4-100·6) 59·4 (54·6-64·3) 115·3 (111·4-119·2) 72·1 (66·8-77·5) Intermediate or higher 78·1 (69·8-86·3) 37·1 (26·0-48·1) 92·0 (83·0-101·0) 49·6 (36·6-62·7) Age of mother at birth (years) <20 105·9 (97·9-113·8) 78·9 (70·5-87·3) 125·4 (116·9-134·0) 92·9 (83·9-101·9) 20-29 89·7 (86·4-93·0) 59·7 (56·0-63·4) 106·9 (103·3-110·5) 73·8 (69·7-77·9) 30+ 85·1 (81·3-89·0) 63·9 (59·2-68·7) 101·0 (96·8-105·3) 81·5 (76·1-86·8) Relation to husband Cousin 98·4 (94·3-102·4) 74·0 (68·1-79·9) 118·6 (114·1-123·0) 91·1 (84·6-97·7) Other relative 101·8 (96·6-107·0) 62·0 (56·2-67·8) 119·6 (114·0-125·2) 78·7 (72·2-85·3) Not related 85·1 (81·3-89·0) 59·4 (55·7-63·1) 101·0 (96·8-105·3) 73·1 (69·0-77·2) Sex of child M 99·0 (95·5-102·5) 68·5 (64·5-72·5) 117·8 (114·0-121·7) 83·5 (79·2-87·9) F 88·8 (85·4-92·3) 58·4 (54·6-62·3) 105·9 (102·2-109·6) 73·9 (69·7-78·2) Table 3: Childhood mortality rates and socioeconomic and demographic characteristics and sex of the child Children in rural settings had a higher risk of dying before age 5 years. One in ten children in urban south/centre died compared with one in eight in the rural south/centre. In the autonomous region 1 in 13 children in urban areas died and 1 in 11 from rural areas died. Within the south/centre, the mortality rates vary between regions. The middle region (which includes the capital, Baghdad) has lower mortality rates than the other three regions. The only appreciable difference is between the middle and the south regions. Within the autonomous region, the Al-Suleimaniyah governorate had the lowest childhood mortality compared with the other two governorates, primarily because it is mostly urban, though the difference was not significant. The results also show that within each region, children born to mothers with at most primary education had significantly higher mortality compared with children born to mothers with at least intermediate education (table 3). Children born to young or old mothers had higher rates than children born to mothers aged between 20-29 years. Children born to related parents had higher risk of dying than children born to unrelated parents. Mortality is higher for boys than for girls in both regions. Because of the potential confounding effect of the various background characteristics on the mortality trend exhibited above, logistic regressions of the risk of dying by age 1 year and by age 5 years were fitted to ascertain whether the trends in mortality (in terms or relative risks) remained unchanged after controlling for the selected background characteristics. Interactions between time periods and characteristics were examined and none was significant. Table 4 presents the result of the regression analysis. The results showed that the adjusted magnitude and the direction of the association between childhood mortality, time periods, and women's characteristics are very similar to the unadjusted results presented in figure 2 and table 3. Infant mortality Under-5 mortality South/centre relative risk (95% CI) Autonomous region relative risk (95% CI) South/centre relative risk (95% CI) Autonomous region relative risk (95% CI) Period (1984-89) 1994-99 2·32 (2·15-2·50) 0·87 (0·80-0·96) 2·14 (2·00-2·28) 0·78 (0·71-0·85) 1989-94 1·83 (1·69-1·97) 1·14 (1·05-1·25) 1·80 (1·68-1·93) 1·12 (1·04-1·21) 1979-84 >1·12 (1·02-1·23) 1·14 (1·04-1·25) 1·11 (1·02-1·22) 1·18 (1·09-1·28) /FONT> 1974-79 1·15 (1·02-1·28) 1·63 (1·48-1·80) 1·16 (1·05-1·28) 1·81 (1·66-1·97) Place of residence (urban) Rural 1·10 (1·05-1·16) 1·21 (1·13-1·30) 1·13 (1·08-1·18) 1·26 (1·19-1·35) South/centre regions (middle) North 1·08 (1·01-1·15) ·· 1·09 (1·03-1·16) .. Middle Euphrates 1·01 (0·95-1·08) ·· 1·03 (0·97-1·09) ·· South 1·08 (1·02-1·15) ·· 1·13 (1·07-1·20) ·· Autonomous governorates (Al-Suleimaniyah) Erbil ·· 1·12 (1·04-1·20) ·· 1·08 (1·02-1·15) Dohouk ·· 1·13 (1·05-1·22) ·· 1·12 (1·05-1·20) Schooling (intermediate) Never attended 1·26 (1·16-1·36) 1·52 (1·30-1·79) 1·30 (1·21-1·40) 1·64 (1·42-1·90) Primary 1·23 (1·14-1·32) 1·30 (1·10-1·53) 1·25 (1·17-1·34) 1·32 (1·13-1·54) Age of mother at birth (20-29 years) <20 1·15 (1·07-1·23) 1·26 (1·17-1·36) 1·13 (1·06-1·20) 1·20 (1·12-1·28) 30+ 1·07 (1·01-1·13) 1·03 (0·95-1·11) 1·06 (1·02-1·12) 1·02 (0·95-1·09) Relation to husband (not related) Cousin 1·12 (1·06-1·18) 1·19 (1·11-1·28) 1·13 (1·08-1·19) 1·22 (1·14-1·29) Other relative 1·13 (1·07-1·20) 1·03 (0·95-1·10) 1·12 (1·06-1·18) 1·05 (0·99-1·12) Sex of the child (female) Male 1·17 (1·12-1·23) 1·14 (1·08-1·21) 1·16 (1·11-1·20) 1·09 (1·03-1·14) Table 4: Adjusted relative risks of dying by age 1 year and age 5 years by socioeconomic and demographic characteristics Overall, the results show high mortality among male children born to mothers with primary or no education and those living in rural areas. South/centre seemed to have been affected more from the consequences of the UN sanctions than the autonomous region. Although the differentials in the two regions are broadly similar, mortality rates in the autonomous region are much lower than in the south/centre. Discussion Since 1991, most of the hospitals and health centres in the south/centre of Iraq have had little or no repair or maintenance work carried out. The functional capacity of the healthcare system has been further degraded by shortages of water and power supply. Communicable diseases, such as water-borne diseases and malaria, which had previously been under control, returned as epidemics in 1993 and have now become part of the endemic pattern of the precarious health situations in the south/centre of Iraq.18,19 Acute malnutrition in the south/centre region of Iraq rose from 3% to 11% and the percentage of low-birth-weight babies (less than 2·5 kg) rose from 4% in 1994 to about 25% of registered births in 1997, mainly due to maternal malnutrition which was estimated to be up to 70% in 1997.20 The most vulnerable groups have been the hardest hit, especially children under 5 years of age who are being exposed to unhygienic conditions.18 Drinking water is currently accessible only to 50% of urban and 33% of rural populations in south/centre, which is far below the proportion before sanctions (90% in urban and 70% in rural areas21) In April, 1995, the UN Security Council adopted a resolution (SCR 986), which was intended to provide for the humanitarian needs of the Iraqi people. This initiative came to be known as the "Oil-for-Food" programme. The adoption of the "Oil-for-Food" programme has played an important part in averting major food shortages in Iraq and has helped to alleviate the health situation, especially in the autonomous region. Since the beginning of the programme, the extent of malnutrition seems to have stabilised in the more populous south/centre of Iraq, albeit at an insufficient caloric level, whereas in the autonomous governorates the situation has improved somewhat. In the autonomous governorates the availability of equipment and trained staff, as well as drugs and supplies, have contributed to substantially increased patient attendance and better care. There is a decline in some infectious diseases such as measles, and better control of poliomyelitis. Water and sanitation have also improved in the autonomous region, because of the "Oil-for-Food" programme.21 Results from the two surveys on childhood and maternal mortality in Iraq clearly show that childhood mortality in the south/centre increased during the period of the UN sanctions that followed the Gulf conflict. Information from several other studies and surveys shows an increase in the rates of malnutrition and in babies born with low birthweight. In the 10 years since the Gulf conflict, infant and under-5 mortality has more than doubled in the south/centre. Childhood mortality rates in the south/centre are now much higher than those 20-24 years ago. Childhood mortality in the autonomous region, however, presents a different picture. Historically, the three autonomous governorates were less developed than the 15 governorates of the south/centre. Since 1991, however, relief has been coupled with development activities. The region also has a less severe embargo as it contains most of the country's rain-fed agriculture and has borders with neighbouring countries that are more open than those in south/centre. The "Oil-for-Food" goods have been distributed more rapidly, the per capita allocations from the programme were set 22% higher than in south/centre and the cash component of the programme assistance in the autonomous region equivalent in value to 10% of all goods. In addition, many more non-governmental organisations (34, compared with 11 in south/centre) are present.23 All these advantages in the autonomous region compared with the south/centre have contributed to a decline in infant and under-5 mortality. Although rates of mortality in the two regions differ, the differentials by place and region of residence, education and age of the mother, blood relation and sex of the child are similar. The patterns are plausible and reinforce the credibility of the survey findings. As found in other countries, children born to women with no education, or in rural areas, have higher mortality than those born to women with education, and in urban areas. Despite a high amount of literacy in the south/centre compared with the autonomous region, childhood mortality rates are higher in the south/centre. Clearly, education of mother, suggested to be a determinant of infant-child mortality, has a limited effect in deteriorating socioeconomic and health conditions, as seen in the south/centre of Iraq. Contributors Mohamed Ali designed the study, did the data management, and analysis, and wrote the paper. Iqbal Shah did the analysis and wrote the paper. Acknowledgments We thank UNICEF for permission to use the data from the two surveys on infant, child, and maternal mortality; the Government of Iraq; the local authorities in the autonomous region for their support and major contributions; the staff and consultants who implemented the surveys in difficult circumstances; the respondents who graciously gave their time for the interviews; and Gareth Jones, Paul Van Look, John Blacker, John Cleland, Carine Ronsmans and Evelyn Dodd. UNICEF provided the financial and technical support for the surveys. References 1 UNICEF. The state of the world's children 1997. New York: UNICEF, 1997. 2 The Economist. Iraq country profile. London: The Economist Intelligence Unit, 1998. 3 UNICEF. Situation analysis of children and women in Iraq 1998. Baghdad: UNICEF, 1998. 4 Hoskins E. Public health and the Persian Gulf War. In: Levy B, Sidel V, eds. War and public health. New York: Oxford University Press, 1997. 5 UNICEF. Iraq immunization, diarrhoeal disease, maternal and childhood mortality survey: evaluation series No 9. Amman: UNICEF Regional Office for the Middle East and North Africa, 1990. 6 Ministry of Health. National child health survey: preliminary report 1989. Gulf Council of Ministers, 1990. 7 The Harvard Study team. The effect of the Gulf crisis on the children of Iraq. N Engl J Med 1991; 325: 977-80 [PubMed]. 8 Ascherio A, Chase R, Cote T, et al. Effect of the Gulf War on infant and child mortality in Iraq. N Engl J Med 1992; 327: 931-36 [PubMed]. 9 Central Statistical Organisations. The 1996 multiple indicators cluster survey: a survey to assess the situation of children and women in Iraq. Final report with result from south-centre governorates. Baghdad: UNICEF, 1997 (UNICEF IRQ/97/288). 10 Zaidi S. Child mortality in Iraq. Lancet 1997; 350: 1150. 11 Ronsmans C, Campbell O. Sanctions against Iraq. Lancet 1996; 347: 198 [PubMed]. 12 Institute for Resource Development. Model "B" questionnaire with additional health questions and commentary for low contraceptive prevalence countries(basic documentation Number 4) Colombia: Institute for Resource Development 1987. 13 Bicego G, Ahmad O. Infant and child mortality: DHS comparative studies. Number 20. Calverton: Marco International, 1996. 14 Greenwood M. Report on public health and medical subjects. Number 33, Appendix 1. London: H M Stationery Office, 1926. 15 Le Thana N, Verma VK. An analysis of sample designs and sampling errors of the demographic and health surveys. DHS Analytical Report Number 3. Calverton: MACRO International, 1997. 16 Central Statistical Office. Iraq fertility survey. Baghdad: Central Statistical Office, 1974. 17 United Nations Manual X. Indirect techniques for demographic estimation (ST/ESA/Series A/81). New York: UN 1983; 81-85. 18 United Nations. Report on the second panel established pursuant to the note by the President of the Security Council of 30/12/99 concerning the current humanitarian situation in Iraq (S/1999/356). New York: UN 1999. 19 United Nations. Review and assessment of the implementation of the humanitarian programme, established pursuant to security council resolution 986. S/19. New York: UN 1999. 20 Food and Agricultural Organization and the World Food Programme (FAO/WFP). Food supply and nutrition assessment mission to Iraq (TCP/IRQ/6713). Rome: FAO, 1997. 21 UN working group on repairing water delivery and treatment facilities. Assessment and review of the humanitarian programme (SCR 986). Baghdad: UNICEF, 1999. 22 Garfield R. Health and well-being in Iraq: sanctions and the impact of oil for food program (unpublished, New York: Columbia University). body.article1845.htmlbody.article1845.html body.article1858.htmlbody.article1858.html _______________________________________________ Sent via the discussion list of the Campaign Against Sanctions on Iraq. To unsubscribe, visit http://lists.casi.org.uk/mailman/listinfo/casi-discuss To contact the list manager, email casi-discuss-admin@lists.casi.org.uk All postings are archived on CASI's website: http://www.casi.org.uk