The prevalence and burden of anemia disproportionately affects children under 5 years of age, pregnant women, and non-pregnant women of reproductive age because of the increased nutrient needs and susceptibility to infections, as well as menstruation in non-pregnant women of reproductive age. Because of these biological factors, most data on anemia are collected for these three target groups. While men can suffer from anemia, women and children are most vulnerable and are the focus of most public health interventions.
How is anemia categorized?
According to the World Health Organization (WHO), anemia is a severe public health problem when the anemia burden is higher than 40 percent; moderate, if over 20 percent; and mild, if over 5 percent (WHO 2011c).
Table 1: Severity of Anemia as a Public Health Problem
|Public Health Problem
How is anemia measured?
Measuring hemoglobin is the primary method for assessing anemia. Anemia is diagnosed if the amount of hemoglobin present in the bloodstream is below the set thresholds, based on age, sex, and physiological status.
The thresholds of hemoglobin in Table 2 are the suggested cutoffs for anemia severity, with differences based on sex, age, and pregnancy status. Often, these different levels of anemia are presented as “any anemia” that combines those with mild, moderate, and severe anemia into one grouping.
The HemoCue system, commonly used in the field, includes a portable photometer, a microcuvette (for collecting blood), and dry hemoglobin conversion reagents. Measurements can also be done on blood samples in a lab.
Table 2: Anemia Cutoffs in Hemoglobin (grams/liter) at Sea Level
|Children 6–59 months of age
|Equal to or above 110
|Children 5–11 years of age
|Equal to or above 115
|Children 12–14 years of age
|Equal to or above 120
|Non-pregnant women of reproductive age
(15 years of age and above)
|Equal to or above 120
|Equal to or above 110
|Men (15 years of age and above)
|Equal to or above 130
Where can we get these data?
Surveys that collect hemoglobin measurements include—
- Demographic and Health Surveys
- Malaria Indicator Surveys
- National Micronutrient Surveys
- WHO Global Database on Anemia.
Many research and evaluation activities collect biomarker data related to anemia and its causes. Kassebaum et al. (2014) includes a list of available data from 150 countries in supplemental tables 1 and 2 of the online appendix. For more information: www.ncbi.nlm.nih.gov/pmc/articles/PMC3907750/bin/supp_123_5_615__index.html.
Information related to anemia prevalence is rarely collected through routine data sources, but it may be available through the country’s health monitoring information system. Consider the usage of health care services in your context when interpreting findings, because not all people suffering from anemia will seek services at a facility.
- Living above sea level and smoking increases hemoglobin concentrations, resulting in an underestimate of the prevalence of anemia. Applying adjustments to hemoglobin concentrations corrects this underestimation. Adjustments are applied by subtracting a set value from individuals’ hemoglobin concentrations, depending on how many meters above sea level an individual resides (Table 3) and/or how frequently he/she smokes (Table 4). Make adjustments before applying anemia cutoffs. If these factors are not properly adjusted, the results will underestimate anemia for populations at higher altitudes and for smokers. If you are using secondary data, many surveys may have made these adjustments. If they have not, and they include populations living 1,000 meters above sea level--or data are from a population of frequent smokers, include it as a weakness in your limitations.
- There is some indication that capillary blood has a slightly higher hemoglobin concentration than venous blood. Studies in the field in low- and middle-income countries report that hemoglobin measurement in capillary blood samples trend higher than from venous samples: 10 of 13 studies, with the difference ranging from 1 to 17 g/L. This trend is also seen in studies done in laboratory settings (Rappaport et al. 2017). Thus, when reviewing studies or reports, consider the blood collection methods when comparing results between surveys that used different techniques. An example of this is in Bangladesh in which the prevalence of anemia differed, despite being collected the same year; and it was hypothesized this was the result of using capillary blood in one survey and venous blood in the other survey (see Box 1). If you find that surveys used different collection methods, include it as a weakness in your limitations.
Box 1: Difference in Anemia Prevalence Between Two Studies in Bangladesh
Example: Bangladesh conducted surveys using different methods in 2011/12. As seen in Figure 11, the Demographic and Health Survey (DHS) data (capillary) showed higher levels of anemia than the national micronutrient survey results (venous) held in the same year.
Table 3: Hemoglobin Concentration Adjustments for Altitude
(meters above sea level)
|Measured Hemoglobin Adjustment (g/l)
Table 4: Hemoglobin Concentration Adjustments for Smoking Status
|Measured Hemoglobin Adjustment (g/l)
|½ -1 packet/day
|≥ 2 packets/day
Describe variations in anemia burden
While national prevalence rates can help you understand the overall burden of anemia in your country, variations at the subnational level are common. These subnational variations are important for programmers and policymakers interested in targeting their interventions to the most affected populations. Reviewing disaggregated national anemia data can help identify areas or groups with an anemia burden higher than the national average. Patterns of anemia may vary within countries because of many factors: the burden of anemia-related diseases and infections, functionality of supply chain and distribution networks, availability of micronutrient-rich foods for consumption, etc. Income inequality and women's empowerment are often reflected in anemia rates that vary with socioeconomic status and maternal education (Kassebaum et al. 2014).
Anemia prevalence varies over time and with populations. The anemia burden can shift from being more severe to less, or the opposite. Discuss with stakeholders the specific factors that could influence the anemia rates at the national and subnational levels. If data are available, review the anemia prevalence for your target groups by geographic area, income, education, or other similar factors to see if any populations are disproportionately affected by anemia. Disaggregation of data by additional indicators—such as sex, pregnancy status, age, education levels, and urban versus rural residence—may also reveal important information. You can prepare graphs of anemia prevalence–by target group or by various characteristics–to illustrate the variation in the anemia burden in your country. These types of basic data are often collected in surveys as part of a “Background” or “Household” characteristics section. For more details on these possible indicators, see Table 5.
Table 5: Possible Disaggregation Indicators for Anemia
|Many surveys report a wealth index or percentiles. An example (based on wealth quintile) is poorest, poorer, middle, richer, and richest.
|The prevalence of anemia often varies between females and males
|Nutrient requirements vary across age groups. Examples of these groupings are—
|Pregnant and lactating women have additional nutrient requirements; they can be reached through a different set of delivery platforms than the non-pregnant population.
|Often grouped by level of school completed. Examples include no formal schooling, some primary schooling, completed primary, completed some secondary schooling, completed secondary, and completed post-secondary education.
|Urban and rural populations have different risk factors for anemia; they often do not have access to the same delivery platforms for anemia prevention and control programs.
|In many countries, anemia can vary significantly across social groups that may face different risk factors and have different access to anemia prevention and control programs. These can include ethnicity, case, religion, indigenous groups, etc.