Analysis of the nutritional status of children under age 5 years:a case study of Namibia

概要

［課程－２］
審査の結果の要旨
氏名 藤村 真耶
This study aimed to analyze the nutritional situation of children under age 5 years in
Namibia. Part 1 identified the prevalence of child undernutrition in stunting, wasting,
and underweight of the 2016 NHIES, and analyzed the national trends of
undernutrition from 1992 to 2016 from the NDHS and 2016 NHIES. Part 2 conducted
a comparative analysis of three prevalence estimates methods of undernutrition (WHO
flags, SMART flags, and PROBIT method) in Namibia from the 2016 national data.
In Part 1, the 2016 NHIES data showed a national undernutrition proportion of 30.3%
stunted children, 11.2% wasted children, and 19.6% underweight children under age 5.
In addition, the country progress in child undernutrition from 1992 to 2016 across the
different surveys was significant for stunting (p= 0.01) and wasting (p <0.01), but not
underweight (p = 0.12). Finally, the 2016 NHIES found a high percentage (9%) of
biologically implausible child measurements.
Part 2 showed that surveys with poor data quality of child anthropometry can lead to
various interpretations of trends and targeting based on the analysis method. After
comparing three methods of analysis to calculate the undernutrition prevalence, the
estimates using the WHO flags consistently resulted in the largest prevalence, followed
by estimates with the SMART flags, and the estimates using the PROBIT method. The
differences in the estimates occur from the different criteria of biological implausibility
of each method. Based on the method, extreme measurements from measurement
errors are excluded from the analysis and it influences the prevalence of
undernourished children. Thus, in addition to analyzing estimates with the WHO
method, the data should be adjusted to exclude the biologically implausible
measurements by using methods such as SMART and PROBIT.
The prevalence estimates derived from poor quality data should be adjusted with
multiple analysis methods for stunted and wasted children in the Namibian context
and similar low- and middle- income countries with poor data quality in child
anthropometry. If the national survey has a percentage of the biologically implausible

measurements of over 1%, then undernutrition prevalence should be compared with
other methods before planning public health policies and programs. After comparing
the WHO flags, SMART flags, and PROBIT methods, different trends, prevalence,
target areas and subgroups of the child population are identified as increased risk of
undernutrition.
The public health importance of high quality anthropometric data was emphasized, and
it can be applied to other low- and middle- income countries with similar data quality
issues in child anthropometry. This study has contributed valuable new evidence on ways
to analyze national surveys with poor child anthropometry data.
よって本論文は博士(保健学)の学位請求論文として合格と認められる。