1998년 9월 23일부터 10월 26일에 걸쳐 수행된 영양조사로 세계식량계획, 유엔아동기금, 유럽연합 등 3개 단체와 북한 당국이 공동으로 참여하였다. 이 글은 당시 조사 결과보고서 원문이다. (www.wfp.org)
Nutrition Survey of The Democratic People’s Republic of Korea
Report by the EU, UNICEF and WFP of a study undertaken in collaboration with
the Government to DPRK (November 1998)
Introduction
The Nutrition Survey aimed to provide the Government and its international
partners with an assessment of the current nutritional situation of young
children in the DPRK. The assessment was designed to serve as a reference
for future evaluations of the combined impact of the range of programmes
implemented by the Government and its partners.
The specific objective of the Nutrition Survey was to assess the nutritional
status of a representative sample of children from six months to seven years
of age.
Survey Methods
Study population
The Democratic People’s Republic of Korea (DPRK) consists of 9 provinces and
3 major municipalities. There are 212 counties in these areas. The counties
consist of smaller administrative entities; these are the Ris (in rural
areas), or the Dongs (in urban areas). The total population of the country
is over 22 Million.
The Central Bureau of Statistics of the DPRK is the body responsible for
providing information related to the population census and the population
size of different counties in the country. This information was made
available at the planning stage for the purpose of sample selection.
The universe of the sampling excluded 82 counties to which access is
currently not granted to international assistance personnel or to which
access could not be guaranteed prior to the survey. Thus the sample selection
was based on 130 counties to which access could be guaranteed, representing
71 percent of the country’s population and 61 percent of all counties, with
a range from 22 percent to 91 percent for various provinces.
Sample Design
The required sample size was estimated for different indicators, using the
following basic assumptions: the number of persons per household = 4.5; and
the proportion of population under 5 years = 11%; design effect = 2 to 10.
A standard margin of error of 5 percentage points was used, and the
prevalence of malnutrition was estimated as a rough guess of what the true
values might be. It was agreed that a sample size of 3600 households
would be drawn for the study.
Multi-stage sampling was used to identify clusters at the level of the
Ris/Dongs. At the first stage 30 counties were selected with probability
proportional to size (PPS). Within each of the selected counties, 4
Ris/Dongs were then selected using PPS, based on the information that all
Ris consist of about 1000 households, and all Dongs have about 1500
households.
In the third stage, 30 households were selected from each of the 120
sampled Ris/Dongs. The selection was made by systematically sampling at
the Ri/Dong level prior to data collection. Annex 1 lists the provinces,
counties and clusters included in the study. They are listed in order of
regions: .i.e., East Coast (North Hamgyong, South Hamgyong, Kangwon); West
Coast (North Pyongan, South Pyongan, Nampo city, North Hwanghae, South
Hwanghae) and the Central Region (Ryanggang, Chagang, Pyongyang city, Kaesong
city). Annex 2 contains a map of the counties in DPRK.
The household selection took place at the head office of the RI/Dong, where
a complete and up-to-date list of all households was made available to the
teams. Systematic sampling was used to select the 30 households from each
cluster (Ri/Dong). At the request of the government, a copy of the list of
selected households was given to the Ri/Dong chairman on the day before the
start of the field work: families were informed of the intended visit and
thus a call-back visit was rarely needed. The Ri/Dong officials guided the
team to the selected households where the target group was surveyed, i.e.
children aged 6 months to 84 months.
Data Collection
The field work was carried out by fourteen teams, each consisting of one
international and four nationals - two health officers, one interpreter and
one driver. Each team had two coordinators, one from the Government and the
other representing one of the international partners in the survey - EU,
UNICEF or WFP. The counties to be surveyed were divided into three groups,
each being assigned a different coordinating international partner. EU
contributed personnel to three teams, UNICEF contributed to eight teams and
WFP contributed to three teams. Following the provision of the agreed
protocol for the survey, field work started on the 23rd September, 1998 and
ended on the 16th October, 1998. Appendix 5 lists the names of the members
of the field teams as well as the international coordinators and technical
advisers.
The coordinator in each team next recorded the age, sex, height and weight
of the children, and examined them for signs of oedema.
Weights and heights were measured using the standard methods recommended by
the WHO. UNICEF electronic scales (Seca, USA) were used for weight
measurement, and Shorr infant/child height measuring boards (Shorr, Maryland)
were used for height measurement.
All field team members, international and national, were intensively trained.
This was done at the Institute of Child Nutrition (ICN) in Pyongyang. The
training covered the methods of sampling, anthropometric measurement and
identification of oedema.
The international members of the field teams received an extra training
session (reinforcement training) in the UNICEF office on the 21st of
September, two days before the start of the field work proper. It aimed
to strengthen their knowledge of the survey methods. A full review of the
sampling methods at the Ri/Dong level was done and a learning-by-doing
exercise was employed to ensure subsequent accuracy in the measurement
of heights and weights. Volunteer children were used for this session.
International co-supervisors were instructed to help maintain high standards
of data collection, especially by checking each questionnaire for
completeness, accuracy and legibility. In addition, stress was laid on
checking the accuracy of the equipment during every field visit.
A number of documents were distributed to the teams, in both Korean and
English languages, to facilitate training and the subsequent reliability
of data collection. These included guidelines for use of the scales, a
summary procedure for nutritional status measurement, a field manual, and
a list of the responsibilities of the team leaders. The field manual is
included in Appendix 3.
Data processing and analysis
Raw data from the completed English language questionnaires were compiled
by designated individuals from each of the international partners involved
in the survey. A Data entry file was created using EPI-INFO (version 6).
All data entered by the three agencies were then merged to create a complete
data file. Double data entry was made in the UNICEF office to help reduce
transcript★ion errors. All data entry operators were trained by the UNICEF
consultants. They were also introduced to the process of data editing, to
data entry programmes , and to file handling methods. Consistency checks
and editing were done to produce a clean set of data for analysis.
The raw data on weights and heights were converted into indices using
EpiInfo EPINUT, based on the growth reference curves developed by the
National Center for Health Statistics (NCHS). This procedure is recommended
by the World Health Organization. Information on age, sex, weight and height
were used to calculate the values of various anthropometric indices - height
-for-age (HA), weight-for-age (WA), and weight-for-height (WH). These indices
were expressed in terms of Z-scores, relative to the international growth
reference values, as recommended by WHO.
The cutoff points recommended by WHO, CDC, and other authorities, to
classify low anthropometric levels were used in the analysis. Children
whose anthropometric indices fell below a Z-score of - 2 were considered
moderately or severely malnourished.
Preliminary data analysis was done at the UNICEF office using EPI INFO
and the Statistical Package for Social Sciences (SPSS). A second round of
analysis was subsequently done under the coordination of representatives of
the three international partners cooperating in the survey. SAS was used for
this analysis in which adjustment was made for the sample design. Since the
sample selection was based on assumptions about the numbers of households
in the Ri/Dong, an adjustment was made in keeping with the actual numbers
of households found in the Ri/Dong. Also, all cases with values less than
-6Z score were excluded from the height for age computation. These cases
were considered to contain errors in the recorded ages.
Main findings
Table 1 shows the age and sex distribution of the children studied. Females
and males were roughly equally represented in the sample.
Overall malnutrition prevalence is shown in Table 2, for each of the three
indices computed. Moderate and severe wasting, or acute malnutrition,
affected approximately 16 percent of the children, including about three
percent with oedema. Moderate and severe stunting, or chronic malnutrition,
affected about 62 percent of all children surveyed, while the prevalence of
moderate and severe underweight, or low weight for age, was approximately 61
percent.
The age and sex distribution of malnutrition is presented in Table 3,
from which it can be seen that the prevalence of wasting peaks in the age
range 12 to 35 months, before and after which it is less. On the other hand,
stunting and underweight continue to rise through the fourth year and tend
not to decline thereafter. The data also show that the prevalence of
malnutrition, measured by all three indicators, tended to be higher among
boys than girls. While the distributions of weight for age and height for
age were normally shaped, their mean Z scores were very negative, indicting
that the whole population of children seems to have been affected by the
crisis.
Table 1. Age and Sex Structure of the Sample |
Age group |
Males |
N% |
Females |
N% |
Total |
N% |
6-<12months |
61 |
|
|
