Senior Technical Advisor USAID Advancing Nutrition Arlington, Virginia
The Demographic and Health Surveys (DHS) and micronutrient status surveys (MNS) measure hemoglobin concentration to determine anemia prevalence. Our study compared results of hemoglobin concentration in children ages 6–59 months from these surveys, conducted within a year of each other, in three countries—Guatemala (2015/2015), Jordan (2009/2010), and Malawi (2015/2016)— to identify reasons explaining discrepancy in results of anemia prevalence.
We compared measures of central tendency and shape of the distribution (skewness, kurtosis, histograms) for hemoglobin concentrations.
DHS estimated lower mean values and higher standard deviations of hemoglobin versus MNS (11.4 ± 1.3 versus 12.0 ± 1.0 g/dL in Guatemala; 11.5 ± 1.4 versus 11.9 ± 1.2 g/dL in Jordan; and, 10.4 ± 1.5 versus 11.4 ± 1.4 g/dL in Malawi). DHS estimated a higher anemia prevalence in all countries (32 versus 12 percent in Guatemala; 34 versus 17 percent in Jordan; and, 63 versus 28 percent in Malawi). The data from the MNS showed high kurtosis and a negative skew values than DHS, implying more data at higher hemoglobin concentrations. DHS in all cases used a drop of capillary blood analyzed in HemoCue 201+; MNS of Guatemala and Malawi used pooled capillary sample and venous blood, respectively, in HemoCue 301; and, MNS of Jordan used venous blood in a hemo-counter.
Future research thus needs to examine prospectively two aspects of the survey design—use of different method of blood extraction and distinct equipment to assess hemoglobin— as likely contributing to the differences in the results.