Nutrition and Health Survey

Badghis Province, Afghanistan

 

February – March 2002

 

 

 

A collaborative survey by:

 

UNICEF

and

U.S. Centers for Disease Control and Prevention (CDC)

 

 

The following persons authored this report and are responsible for its content:

 

Bradley A. Woodruff

Meredith Reynolds

U.S. Centers for Disease Control and Prevention

Atlanta, Georgia USA

 

Felicite Tchibindat

Cyridion Ahimana

UNICEF – Afghanistan Country Office

ACKNOWLEDGMENTS

 

Although this survey was funded by UNICEF, it was the work of many people from many organizations.  The authors thank those attended the workshop in Mazar-i-Sharif where many valuable suggestions were discussed (see appendix 1 for list of participants). 

            Many other UNICEF staff contributed to various aspects of this survey, including Peter Salama, Head of the Health and Nutrition Section of the UNICEF – Afghanistan Country Office. Dr. Simon Azariah of UNICEF and Dr. Mesfin Teklu of World Vision assisted with much of the preparatory work.  In addition, many UNICEF national staff did translations and backtranslations, provided the information necessary to construct the local calendar, and offered invaluable advice on all aspects of the survey. 

            The Ministry of Public Health of Herat Province provided the personnel who acted as survey team supervisors.  The Office of the Governor of Badghis Province, especially Mr. Zaudin, provided advice on logistic matters, including identification of the location of selected villages. 

            The Norwegian Project Office/ Rural Rehabilitation Association for Afghanistan (NPO/RRAA), the Danish Committee for Aid to Afghan Refugees (DACAAR), Ockenden International (OI), and Oxfam-UK provided lists of villages in Badghis Province.  These lists were essential to carrying out the survey sampling.

            Major partners of UNICEF provided general advice and overall coordination: the World Food Programme (WFP), the World Health Organization (WHO), and the United Nations Office for Coordination of Humanitarian Affairs (UNOCHA). 

The authors offer special thanks to the survey workers (listed in appendix 1) whose intelligence, dedication, and overall competence in the face of extreme logistical difficulties made this survey possible.

TABLES OF CONTENTS

                                                                                                                                                  Page

List of tables, figures, and appendices .....................................................................................     4

Executive summary...................................................................................................................     5

Introduction ................................................................................................................................     6

Goals and Objectives ................................................................................................................     6

Methods

      I.      Sample size and sampling ..........................................................................................     7

      II.     Data collection .............................................................................................................     8

      III..   Definitions  ...................................................................................................................     9

      IV.    Data analysis  ..............................................................................................................   10

Results

      I.      Description of survey population .................................................................................   10

      II.     Child nutrition ...............................................................................................................   11

      III.     Child health and vaccination ........................................................................................   12

      IV.    Women of reproductive age ........................................................................................   13

      V.    Morality .........................................................................................................................   13

Conclusions and Discussion

      I.      Household characteristics ...........................................................................................   14

      II.     Child nutrition ...............................................................................................................   14

      III.    Child health and vaccination ........................................................................................   16 

      IV.    Women of reproductive age ........................................................................................   16

      V.    Mortality ........................................................................................................................   17

Recommendations

      I.      Children ........................................................................................................................   17

      II.     Women ........................................................................................................................   18

      III.    General ........................................................................................................................   18

References ................................................................................................................................   19

Tables                                                                                                                                          ......................................................................................................................................   20

Figures                                                                                                                                          29

Appendix 1

Appendix 2

Appendix 3

Appendix 4

LIST OF TABLES, FIGURES, AND APPENDICES

                                                                                                                                                  Page

 

Table 1.     Description of households included in survey ........................................................   20

Table 2.     Description of children and women of reproductive age included in survey ..........   21

Table 3.     Prevalence of acute and chronic malnutrition and underweight

                  (as defined by z-score) in children less than 5 years of age, by age .....................   22

Table 4.     Prevalence of acute and chronic malnutrition and underweight

                  (as defined by % of median) in children less than 5 years of age, by age .............   22

Table 5.     Potential risk factors for acute and chronic malnutrition in children < 5 years

                  of age .......................................................................................................................   23

Table 6.     Possible signs of micronutrient deficiencies in children less than 5 years of age,

                  by age ......................................................................................................................   24

Table 7.     WHO breastfeeding indicators ................................................................................   25

Table 8.     Cumulative prevalence of diarrhea and pneumonia in previous 2 weeks in

                  children less than 5 years of age, by age ...............................................................   25

Table 9.     Reproductive outcomes among women of reproductive age .................................   25

Table 10.   Prevalence of various categories of malnutrition (as defined by BMI) in

                  non-pregnant women of reproductive age ..............................................................   26

Table 11.   Distribution of MUAC measurements in all women of reproductive age ................   26

Table 12.   Prevalence of various categories of malnutrition using the combination of

                  BMI and MUAC, non-pregnant women of reproductive age ....................................   27

Table 13.   Age- and sex-specific death rates since Eid Qurban 1379 (13-16 February 2001)   27

Table 14.   Distribution of causes of death, for deaths with cause determined,

                  by age group ............................................................................................................   28

 

Figure 1.    Population pyramid of survey sample

Figure 2.    Distribution of weight-for-height z-scores, children < 5 years of age

Figure 3.    Distribution of height-for-age z-scores, children < 5 years of age

Figure 4.    Distribution of weight-for-age z-scores, children < 5 years of age

Figure 5.    Percent (3-month moving average) of children eating solid food, by age

Figure 6.    Percent (3-month moving average) of children breastfeeding, by age

Figure 7.    Distribution of BMI values in non-pregnant women of reproductive age

Figure 8.    Number of deaths, by month

 

Appendix 1.     Persons who attended the Mazar workshop and members of survey teams

Appendix 2.     List of villages included in survey

Appendix 3.     Data collection form in English

Appendix 4.     Questions used for verbal autopsy

 

EXECUTIVE SUMMARY

 

            This 30 cluster survey of Badghis province, the first province-level assessment of nutrition and health since the change of government in Afghanistan, was designed to establish baseline data on the nutritional status of children and women and the level and causes of mortality.  Data were collected 17 - 31 March 2002 on 507 households with 545 children less than 5 years of age and 555 women 15 – 49 years of age.  Major findings are summarized below:

 

INTRODUCTION

 

            For decades, Afghanistan has been one of the world’s poorest countries with the worst health statistics.  Health problems in the country’s population, already significantly exacerbated by more than 20 years of civil conflict, have recently been further worsened by 3 years of severe drought and the military conflict following the events of September 11, 2001.  An estimated six million people have little or no access to health care [1], while an estimated seven million rely on food aid to survive.  Many available health statistics, such as the estimate of infant mortality rate of 220 per 1000 live births and maternal mortality rate of 1700 per 100,000 live births [2], are outdated and were originally based on local data collections or estimates.  Several local surveys, conducted in various parts of Afghanistan, have demonstrated acute malnutrition and high crude mortality rates [3, 4]. 

Based on these surveys, it is estimated that an average of 10% of children less than five years of age are acutely malnourished, of which 2% are severely malnourished.  Reports of declining nutritional status of children and women are increasingly common, particularly from the western and northern regions of the country, and many humanitarian aid organizations have implemented feeding programs and intensified food distributions in these regions. UNICEF, along with participating partners, including WFP and national and international non-governmental organizations, have developed a Concept Paper proposing a framework for nutritional surveillance composed of 1) monthly nutritional status information from health centers, 2) cross-sectional quantitative surveys, and 3) community-based sentinel site nutritional surveillance.  This survey of Badghis province represents one of the first population-based surveys and can serve as a model for subsequent province-level surveys.

 

 

GOALS AND OBJECTIVES

 

            The overall goal of this survey is to assess the health and nutritional status of children less than 5 years of age and women of reproductive age (15 – 49 years of age).  This assessment will be used to establish baseline data for a nutritional surveillance system and to provide recommendations to national and international organizations providing health and nutrition services.

 

The specific objectives of this survey are to estimate:

·         The prevalence of salt iodation

·         The prevalence of acute and chronic malnutrition in children less than 5 years of age

·         The prevalence of clinically apparent anemia and vitamin deficiencies (riboflavin and vitamins A, C, and D) in children less than 5 years of age

·         WHO breastfeeding indicators

·         The two-week cumulative prevalence of diarrhea and acute respiratory infection in children less than 5 years of age

·         The prevalence of malnutrition in women of reproductive age (15-49 years of age)

·         The prevalence of iodine vitamin A deficiency in women of reproductive age

·         The reproductive history of women of reproductive age

·         Crude mortality rate and causes of death

·         Age-specific mortality rates, including the mortality rate among children less than 5 years of age

·         The coverage of recent measles vaccination campaigns among children 9 months to 5 years of age

 

METHODS

 

I.  Sample size and sampling

 

Sample size calculations used the following assumptions: 1) the limit of statistical significance (alpha) is 0.05 (that is, 95% confidence interval will be used), and 2) the power (beta) equals 0.8.  The UNICEF Multi-Indicator Cluster Survey (MICS) [5] completed in 2000 in six provinces in Eastern Afghanistan provided estimates of prevalence and demographic data which we used to formulate assumptions for sample size calculations.  We assumed an average of 1.3 children less than 5 years of age and one woman 15-49 years of age per household.  Because the sample size required to achieve a given statistical precision increases as the estimated prevalence approaches 50%, the assumed prevalence rates for this survey were assumed to be closer to 50% than those found in the MICS. 

 

Assumptions and estimated sample size for nutrition outcomes, Badghis Nutrition and Health Survey, March 2001. 

 

    *    DEFF = Design effect

  

Because the primary objective of this survey was to measure nutritional status of young children, the final sample size used was based on nutrition outcomes in young children.  The largest sample size (n=534) is needed to obtain an estimate of the prevalence of anemia.  Because emigration from Badghis Province has been extensive in recent years, we assumed that 25% of households would be unreachable.  With an average of 1.3 children less than 5 years of age per household and household absenteeism or non-response of 25%, a total of 534 households are needed to find 534 children.  These households were to be grouped into 30 clusters of 18 households each.   This would result in a precision of ±4 percentage points for acute malnutrition in children assuming an estimated prevalence of 20%.  Approaching 18 households per cluster should result in a total of 684 women of reproductive age included in the survey.  This would result in a precision of ±5 percentage points for anemia, assuming a prevalence of 50%.  Assuming a cumulative incidence of death of 5.7% (crude mortality rate equivalent to 1.5 deaths per 10,000 population per day), a sample of 534 households with an average of 6 persons per household would yield 1,217,520 person-days of experience and 183 deaths.  If the design effect for mortality were 2, the precision around a point estimate of 1.5 deaths per 10,000 per day would be ±0.3.  Although the sample size was based on anemia prevalence, because the equipment and supplies for measuring hemoglobin concentration were unavailable at the time data collection began, this survey did not collect these data.

Lists of all villages in each district in Badghis Province were obtained from organizations responsible for relief food distribution.  These organizations had recently registered every village in Badghis Province and the number of households contained therein.  In the first sampling stage, 30 clusters were selected after constructing a single list of all villages in Badghis Province, the number of households in each, and a column of cumulative sums.   Sampling probability proportional to size was done by calculating a sampling fraction (the total number of households in Badghis Province divided by the number of clusters [30]), and adding this sampling fraction repeatedly to an initial random number.  Each village in which the resulting number fell became the site for one cluster.  No village contained more than one cluster.  Because population estimates were not available for Qala-e-Nau town at the time of sampling, it was not included in the sampling frame.  However, its 2000 households do not constitute a large proportion of the estimated 114,000 households in Badghis Province.  Appendix 2 shows a list of the villages selected for clusters.  Two villages in Ghormach District were felt to be inaccessible due to security risk.  Two other nearby villages were substituted for them.  

For the second sampling stage, a mullah, abob, or other village leader was asked to provide a preexisting list of all households in each village or to help create one.  The households on the list were then numbered, and 18 households were selected using a random number table.  In each village, it was confirmed that all households currently residing in the village were included on the list of households.  Any missing households, such as recent returnees, were added to a preexisting list.  In villages containing more than 200-300 households, village leaders helped create a list of mosques with the number of subscribing households in each.  Similar to the first stage sampling, one mosque was then chosen probability proportional to size.  All households subscribing to this mosque were then listed, and 18 selected as described above.

A household was defined as any group of persons occupying the same structure and sharing household resources, such as food and bedding.  Members of a household were not necessarily relatives by blood or marriage.  If no one was at home at a selected house, a neighbor was consulted concerning the whereabouts of members of the household.  If the members had departed permanently or were not expected to return before the survey team had to leave the village, the household was skipped and not replaced.  If household members were expected to return, the survey team revisited the house at least twice more before declaring the household missing.

 

II.  Data collection

 

            A data collection tool was created after consultation with national and international organizations providing nutrition and health services to the population of western Afghanistan.  The types of data collected conform to the recommendations of the February meeting in Kabul of organizations working in nutrition and food.  The entire form was translated from English into Dari, the major language of Badghis province, and then back-translated into English by a second translator (see appendix 3 for the English form).  The survey instrument was pretested in 9 households in a village near Herat City which were not included in the survey, and revisions were made based on this experience. 

Data were collected by 5 teams of 4 persons each.  Each team included a supervisor, a logistics coordinator, and two interviewers (one female and one male).  All supervisors were medical doctors.  All survey workers received 4 days of classroom training and 1 day of field practice training under close supervision.  During data collection for the first cluster done by each team, an expatriate supervisor or consultant from UNICEF monitored each team.

At each household, interviewers asked questions about displacement and water source for the household.  In addition, a household census was taken as of Eid Qurban of the previous year (1379 in the Afghan calendar; 13-16 February 2001 in the Gregorian calendar). Births and deaths occurring in each household between Eid Qurban and the date of the survey were recorded along with month of occurrence. A local calendar of events was used to determine ages of household members and dates of death.  The cause of each death was classified into 1 of 11 categories (war related injury, non-war related injury, measles, tetanus, watery diarrhea, dysentery, meningitis, pneumonia, malaria, scurvy, or other) using a hierarchical series of questions adapted from a WHO recommended protocol for verbal autopsies (see appendix 4 for the questions used). [6]  A sample of salt from each selected household was tested for iodine content.

            Survey workers asked questions of each woman of reproductive age in each household regarding nightblindness, number of pregnancies, number of births, date of last delivery, and tetanus vaccination history.  In addition, an examination for goiter was performed and each woman had mid-upper arm circumference (MUAC), height, and weight measurements taken.

Information was gathered from an adult household member, preferably the mother, on each child less than 5 years of age regarding nightblindness, breastfeeding history, vitamin A supplementation and vaccination history, and recent diarrhea and acute respiratory infection.  Because vaccination cards have not been issued during recent mass immunization campaigns, mothers' reports were taken as evidence of vaccination against measles and receipt of vitamin A supplementation.  BCG vaccination was confirmed by examination for a characteristic scar on the child’s left arm. 

The physician supervisor performed a physical examination targeted to signs of micronutrient deficiency.  Survey workers then measured the child’s weight and height.   Children less than 5 years of age and women of reproductive age were weighed to the nearest 100 grams with UNICEF Uniscale.  For children less 24 months of age, length was measured to the nearest millimeter in the recumbent position using a standard height board.  Children 24 months of age or older were measured in a standing position.  Women’s height was measured using a portable stadiometer while the woman was standing against a vertical surface or support.  MUAC in women was measured using a standard measuring tape. 

 

III.  Definitions

 

Z-scores were used in most analyses of anthropometric data on children in this survey. However, percent of median is used in many situations where a simpler calculation is needed, such as screening for admission to feeding programs.  Therefore, for purposes of comparing the results of this survey to other data, the prevalence rates of acute and chronic malnutrition and underweight are also presented percent of median.  The relevant definitions are as follows:

 

Type of malnutrition	Anthropometric index	Degree of malnutrition	Definition using z-score	Definition using percent of median
		None	> -2.0	> 80%
Acute	Weight-for-height	Moderate	> -3.0 but < -2.0	> 70% but < 80%
		Severe	< -3.0 or edema	< 70% or edema
		None	> -2.0	> 90%
Chronic	Height-for-age	Moderate	> -3.0 but < -2.0	> 80% but < 90%
		Severe	< -3.0	< 80%
		None	> -2.0	> 80%
Underweight	Weight-for-age	Moderate	> -3.0 but < -2.0	> 70% but < 80%
		Severe	< -3.0	< 70%

 

Z-scores and percent of median were derived from comparison of children in the survey sample to the NCHS/CDC/WHO reference population.[7] 

            Malnutrition among non-pregnant women of reproductive age was assessed using BMI WHO-recommendationed categories.[7]  Body mass index (BMI) was calculated as weight in kilograms divided by the square of the height in meters.

 

BMI                              Category of malnutrition

               < 16.0                                   Severe thinness

               16.0 – 16.9                           Moderate thinness

17.0 - 18.4                            Mild thinness

                                    18.5 - 24.9                            Normal

                      25.0 - 29.9                            Overweight

                      > 30                                      Obese

 

            This report presents the proportion of non-pregnant and pregnant women who have MUAC measurements less than 22.0 cms; however, because no consensus exists regarding the cut-off points for MUAC for adults, the distribution of MUAC results is also presented.

Malnutrition in non-pregnant women of reproductive age was also defined using a combination of BMI and MUAC [8], as follows :

 

Categories of nutritional status using a combination of BMI and MUAC

 

 

Although not considered normal nutritional status, the health consequences among persons in the category “Normal?” have not been determined. 

 

V.  Data analysis

 

            Data were keypunched and analyzed using EpiInfo version 6.04b. The calculation and analysis of anthropometric indices was carried out using the Epinut module.  Indicators of the precision of prevalence estimates, such as confidence intervals, for major health outcomes accounted for the cluster sampling and implicit stratification used in selecting the sample for this survey.  Stratification was achieved because the list of villages was first ordered by district before selecting the first stage sample.  A p value <0.05 was considered to be statistically significant.  The strength of association between potential risk factors and various outcomes was estimated using relative risks, and 95% confidence intervals (CI) were used to judge the statistical precision of point estimates for relative risks.

 

 

RESULTS

 

I. Description of survey sample

 

Households

            The survey sample included 507 households.  Interviews were done by female survey workers in 359 (70.8%) households and by male survey workers in the remaining households.  Few households were displaced from their homes at the time of the survey (table 1).  The most common primary source of water for the majority of households was rivers and streams.  Overall, only 26 (5.1%) of households obtained water from a safe source (borehole, truck, or piped system).  The current source of water was the same as in the previous years for most households.  The mean average time required to bring water to the household was 94 minutes.  Although the majority of households had water within a 1 hour round trip, more than one-quarter required more 2 hours. About one-half of all households had received some form of food aid since the change in government that occurred in October 2001.  Iodine was present in the salt of very few households.

 

Individuals

Overall, the survey collected data on 3075 members of selected households; these household members included 1529 (49.8%) males, 1542 (50.2%) females, and 4 for whom sex was not recorded.  A population pyramid for these household members is shown in figure 1.  

The survey collected more detailed data on 545 children less than 5 years of age.  For 528 (97.1%) of these children, interviews were conducted with the mother.  For 10 (1.8%) children, the mother was not alive at the time of the survey.  The age distribution of children is shown in table 2.  Among these children, 240 (44.1%) were boys, 304 (55.9%) were girls, and 1 did not have sex recorded; 59 (10.8%) children were 0 – 6 months of age.   The survey sample also included 565 women of reproductive age.  The distributions of age and literacy of these women are shown in table 2.  The large majority of women were not literate. 

 

II.  Child nutrition

 

Acute malnutrition

Acute malnutrition was not common in children less than 5 years of age (table 3).   Of those children with acute malnutrition, 6 (17%) had edema.  Among children without edema, the mean weight-for-height z-score is -0.28 and the standard deviation is 1.1.  The prevalence of acute malnutrition is greatest among children 12 – 23 months of age.  In other age groups, the prevalence of acute malnutrition is not statistically significantly greater than that found in the reference population (2.3%).  Among all children less than 5 years of age, the prevalence is not statistically higher among the 301 girls (7.0% with acute malnutrition) than among the 238 boys (5.9%).  The distribution of weight-for-height z-scores is shown in figure 2.  The entire curve for the survey sample is slightly shifted to the left when compared to that of the reference population. 

 

Chronic malnutrition

Chronic malnutrition was much more common that acute malnutrition (table 3).   A greater proportion of children had severe chronic malnutrition than moderate chronic malnutrition.  The mean height-for-age z-score in the survey sample is –2.19 and the standard deviation is 1.59.  This elevated standard deviation demonstrates that either height or age may not be determined with precision.  The prevalence of chronic malnutrition increases with age.  Chronic malnutrition was somewhat more common in boys that in girls (147 [61.8%] vs. 162 [54.2%], respectively (relative risk = 1.1, 95% confidence interval = 0.99, 1.3).  The distribution of height-for-age z-scores is shown in figure 3.  The entire curve for the survey sample is substantially shifted to the left when compared to that of the reference population, demonstrating that virtually all the children in the survey sample have some degree of chronic malnutrition. 

 

Underweight

Underweight is also very common in the survey sample (table 3).   The mean average weight-for-age z-score in the survey sample is –1.57 and the standard deviation is 1.34.  The distribution of weight-for-age z-scores is shown in figure 4.  The entire curve for the survey sample is substantially shifted to the left when compared to that of the NCHS/CDC/WHO reference population, demonstrating that virtually all the children in the survey sample have some degree of low weight. 

 

Percent of median

            The prevalence rates of acute and chronic malnutrition and underweight as expressed in percent of median are presented in table 4. 

 

Risk factors for acute and chronic malnutrition

            The association between potential risk factors and acute and chronic malnutrition is shown in table 5.  Few of the potential risk factors analyzed demonstrate an association with either acute or chronic malnutrition.  Nevertheless, none of the eight children of literate mothers were acutely malnourished, and a smaller proportion of children of literate mothers than children of illiterate mothers had chronic malnutrition, albeit without statistical significance.  Mother's who were malnourished, as measured by BMI, were more likely to have acutely malnourished children.  Receipt of relief food was inversely associated with acute malnutrition and not at all associated with chronic malnutrition. 

 

Micronutrient deficiencies and supplementation

      Vitamin A.  The prevalence of night blindness and Bitots spots is shown in table 6.  Mothers of 145 (26.6%) children could not answer the question regarding night blindness, usually because these children were too young to determine if night blindness existed.  Among all children less than 5 years of age, 295 (55.6%) had received vitamin A supplementation at any time in their lives. 

      Vitamin C.  The prevalence of signs of vitamin C deficiency is shown in table 6.  Signs of scurvy were more common in children 12 months of age or older than in infants less than 12 months of age.

      Vitamin D.  The prevalence of major signs compatible with vitamin D deficiency is shown in table 6.  As with vitamin C, signs of vitamin D deficiency were more common in children 12 months of age and older than in infants. 

      Riboflavin.  The prevalence of angular stomatitis, one sign of riboflavin deficiency, is shown in table 6. 

      Anemia.  The prevalence of pallor is shown in table 6.  Despite the fact that this sign is only consistently found in children with more severe anemia, it was found relatively commonly in the survey sample. 

 

Infant feeding 

The breastfeeding indicators recommended by WHO are shown in table 7.  Exclusive and predominant breastfeeding are very common.  In contrast, timely complementary feeding is not common.  Breastfeeding at 1 year of age is nearly universal, and at 2 years of age, about one-half of children are still breastfeeding.  Few children less than 12 months of age have received feeding from a bottle with a nipple.  The mothers of only 3 (0.6%) children report giving formula or dried milk to their children. 

The mean and median time betweeen birth and first breastfeeding was 13.1 and 3.0 hours, respectively.  Overall, 175 (79%) of 222 breastfeeding children had initiated breastfeeding by 24 hours after birth.  Figures 5 shows the proportion of children, by age, who eat solid or semi-solid food.  Figure 6 shows the proportion who are still breastfeeding.  From these figures, the median age of introduction of complementary food is 10-11 months, and the median age of weaning is 20-23 months.

 

III.  Child health and vaccination

 

Childhood morbidity  

Overall, a large proportion of children less than 5 years of age had had diarrheal disease or acute respiratory infection in the 2 weeks prior to the survey (table 8).  However, relatively few children with these illnesses were taken to a health facility.   

 

Vaccination 

Measles.  The mothers of only 277 (59.4%) children 9-59 months of age reported that the children had been vaccinated since October 2001.  The mass vaccination campaign recently completed in Badghis Province targeted children 9 months to 14 years of age. 

BCG.  Only 73 (13.5%) children had a scar indicating BCG vaccination at some time in their lives.

 

IV. Women of Reproductive Age

 

Reproductive history

            Overall, 78 (14.0%) women of reproductive age were pregnant at the time of the survey (table 9).  The highest prevalence of pregnancy is found in the age groups 25-29 and 30-34 years.  Pregnancy is less common in women 40 years of age and older.  As expected, the average number of pregnancies and births increases with age; it reaches a plateau in women 40-44 years of age.  For women less than 35 years of age, the prior delivery was an average of less than 2 years prior to the survey. 

 

Nutritional status

            BMI.  The prevalence of various degrees of malnutrition, as defined by BMI, among non-pregnant women of reproductive age is shown in table 10.  The overall prevalence of undernutrition in non-pregnant women of reproductive age was not greatly elevated; however, a much larger proportion are at nutritional risk.  The distribution of BMI values in this group is shown in figure 7.  Among non-pregnant women of reproductive age, the mean and median BMI were 21.1 and 20.6, respectively, and the range of BMI values was 14.0 - 33.4. 

MUAC.  The distribution of MUAC in all women of reproductive age is shown in table 11.  Overall, 95 (17.1%) women had a MUAC < 22.0 cms.  Among all women of reproductive age, the mean and median MUAC values were 23.9 cms and 23.8 cms, respectively, and the range was 15.0 - 34.6 cms. 

BMI/MUAC.  The proportion of non-pregnant women of reproductive ages who fall into various categories using the combined BMI and MUAC criteria is shown in table 12. 

Micronutrient deficiencies.  Among women of reproductive ages, 62 (11.5%) had detectable goiter, indicating substantial iodine deficiency and a risk of having iodine-deficient children at birth.  Additionally, 26 (4.7%) women reported having nightblindness, a symptom of vitamin A deficiency. 

 

V. Mortality

 

            The crude mortality rate in the survey sample between Eid Qurban 1379 (13-16 February 2001 in the Gregorian calendar) and the time of the survey was 0.72 deaths per 10,000 population per day.  This crude mortality rate corresponds to rates of 2.2 deaths per 1000 per month and 26.3 deaths per 1000 per year.  Age- and sex-specific mortality rates are shown in table 13.  The age-specific mortality rate for children less than 5 years of age (2.51 per 10,000 per day) is substantially higher than the rates in other age groups.  The mortality rate for children < 5 years of age corresponds to rates of 7.6 deaths per 1000 children < 5 years of age per month or 91.6 deaths per 1000 children < 5 years of age per year.  The distribution of deaths over time is shown in figure 8.  With the exception of March 2001, deaths are not markedly clustered over time. 

For all deaths, including those for which cause was not determined, 55 (65%) occurred in children less than 5 years of age, and 8 (9%) occurred in person 50 years of age and older.  Shown in table 14 is the distribution of causes of those deaths for which a cause was determined by verbal autopsy.  For all ages combined and for children less than 5 years of age, bloody diarrhea and pneumonia were the most common causes of death identified by verbal autopsy. 

Information on nutritional status was collected for 67 (79%) of the 85 deaths identified in the survey.  Malnutrition was reported to be an aggravating factor among 38 (57%) of these 67 deaths and in 25 (58%) deaths in children less than 5 years of age.  Among the seven deaths of women of reproductive age for whom pregnancy or delivery status is known, 2 (22%) were, at the time of death, pregnant or within 40 days of giving birth. 

 

 

CONCLUSIONS AND DISCUSSION

 

I.            Household characteristics

 

Water supply

        Overall, the accessibility and quality of water is poor for households in Badghis Province.  Most households depend on unsafe sources for their water supply.  Moreover, the members of many households spend substantial time obtaining this water. 

 

Food

        Food aid has not been universally received by households in Badghis Province.  Iodine is virtually absent from salt consumed in Badghis households. 

 

II.     Child nutrition

 

Acute malnutrition

            Acute protein-energy malnutrition, although not uncommon, is not currently an overwhelming public health problem among children less than 5 years of age in Badghis Province.  Among these children, only children 12-23 months of age appear to be at elevated risk of acute malnutrition.  Infants less than 12 months are not at disproportionate risk. 

            The shape of the distribution curve of weight-for-height z-scores and the relatively low standard deviation imply that the anthropometric measurements taken during the survey are relatively precise.  Moreover, these measurements are probably accurate given that Uniscales were used to measure weight and survey supervisors carefully checked the accuracy of the height boards used to measure height and length.

            In some severe emergency situations elsewhere, high mortality among children less than 5 years of age has decreased the apparent prevalence of acute malnutrition.  Malnourished children had died, leaving only those with better nutritional status to be measured in a survey.  However, this is probably not the situation in Badghis Province given the mortality rate measured in children less than 5 years of age.  Although this rate is elevated above baseline, it is not nearly as high as the age-specific mortality rates measured in the extreme situations mentioned above.

            The level of acute malnutrition among children in Badghis Province may not, in and of itself, justify extensive feeding programs.  Nonetheless, communicable diseases, especially dysentery, can, even in the absence of severe food shortage, have an important affect on nutrition status.  The level of morbidity and mortality due to pneumonia and diarrhea in Badghis Province, in combination with the somewhat elevated prevalence of acute malnutrition, may justify enrollment of malnourished children in targeted supplementary feeding programs.  Moreover, a small hospital-based therapeutic feeding program may be necessary for the relatively few children with severe acute malnutrition.  Nonetheless, there is little justification for blanket supplementary feeding or implementation of specialized therapeutic feeding centers.

 

Chronic malnutrition

            In contrast to acute malnutrition, chronic malnutrition is a very common problem affecting all age groups of children less than 5 years of age.  The prevalence of severe chronic malnutrition among children 12 months of age and older is of special concern.  Nonetheless, the wide standard deviation of height-for-age z-scores and their skewed distribution indicate that this index was not measured with adequate precision.  Most mothers could not produce documentation of their children's dates of birth nor state a specific date of birth. Moreover, reported ages showed a strong preference for certain ages.  More than 3% of the 545 children were 14, 24, 26, 35, or 48 months of age.  If the number of births were constant over time for this age cohort, only 1.7% of the sample should be of any specific age.  As a result, the results of this survey should not be used as a specific baseline against which to evaluate current or future interventions targeted toward chronic malnutrition.          Regardless of these limitations, the very low height-for-age z-scores found in this survey, as well as data from other sources, indicate a long-standing chronic nutrition problem among children less than 5 years of age in Afghanistan in general. 

 

Risk factors for malnutrition

            The lack of association between receipt of food aid and either acute or chronic malnutrition in children may indicate that one-time or sporadic food distributions, as carried out in Badghis Province this past winter, have had no effect on children's nutritional status. 

            There is a clear association between mothers' nutritional status and the presence of acute malnutrition in their children less than 5 years of age.  Unfortunately, a cross-sectional survey cannot determine if this is a cause-effect relationship.  It may be true that in households with insufficient food, all members become malnourished.  On the other hand, there may be mechanisms by which maternal nutritional status may directly influence child nutrition status.  Regardless, provision of food to households with malnourished members may assist in the recovery of all malnourished household members.

 

Micronutrient deficiencies

            Signs and symptoms of several micronutrient deficiencies are not uncommon among children less than 5 years of age in Badghis Province.  The prevalence of Bitots spots exceeds the widely accepted threshold of 0.5% which defines vitamin A deficiency as a public health problem.  In addition, only about one-half of children have received vitamin A supplementation at any time in their lives. 

            The gum signs of scurvy are present in children less than 5 years of age, especially in children older than 12 months.  However, these signs may be due to many conditions, such as poor oral hygiene.  The more specific signs, including widespread bruising and perifollicular hemorrhage, are absent or much less common.  Although scurvy has been widely reported in western and central Afghanistan, the results of this survey do not clearly indicate that scurvy is a widespread, serious problem in Badghis Province in children less than 5 years of age at this time. 

            On the other hand, relatively specific signs of rickets, including rachitic rosary at the rib-cartilage junction and bowed legs, are more common.  This may indicate that vitamin D deficiency is a problem among children less than 5 years of age, especially those 24 months of age and older.  In other countries, vitamin D deficiency is a seasonal condition which largely resolves in the summer months when children spend more time outdoors in sunlight. 

            Although angular stomatitis is present in children in Badghis Province, it can result from may causes, including fungal infection, iron deficiency, and others.  In populations with widespread riboflavin deficiency, angular stomatitis can often be identified in a much larger proportion of the population.  For example, in Bhutanese refugee adolescents, angular stomatitis was found in 29% of those included in a population-based random sample.[9]   For these reasons, we cannot definitively co