Wealth assessment is relevant in health surveys, particularly in low- and middle-income countries, as it enables the analysis of health inequalities across groups with different wealth levels and helps to understand the access to and utilization of health services. In these contexts, socioeconomic position is usually estimated through an asset index, which relies on information about household assets and characteristics as a proxy for economic status.
Definition and classification:
The index is generated using principal component analysis (PCA), based on data on household assets and characteristics, such as type of drinking water source, sanitation facilities, flooring material, walls, roof, etc.
The index generates a continuous score, which is then used to categorize households into equal-sized groups, typically quintiles or deciles. In the case of quintiles, Q1 represents the poorest 20% of households and Q5 represents the wealthiest 20%, as illustrated below:
Under-five mortality rates in Bangladesh, by wealth quintiles. (Data source: Bangladesh DHS 1993-2011)
Example and interpretation:
“Coverage gaps in skilled birth attendance increased significantly across five different groups of wealth. In the poorest quintile, coverage was at 40%, while it rose to 55% in the second quintile, 70% in the third, 85% in the fourth, and reached 95% in the richest quintile. These results emphasize the substantial disparity between the wealthiest and poorest segments of society, illustrating a clear gradient in coverage based on wealth levels.”
Urban/rural residence is a stratification variable that highlights geographic inequalities in access to services, environmental exposures, and health outcomes. It is crucial for understanding the distribution of living conditions and is widely used in studies to compare outcomes, such as infant mortality, maternal care, and vaccination coverage, between urban and rural areas.
Definition and classification:
Most countries classify geographic areas as urban or rural based on administrative, demographic, or infrastructure criteria, although definitions vary widely across countries and datasets. Urban areas generally have higher population density, while agricultural activities and scattered housing characterize rural areas. In surveys, urban or rural households are classified according to the sampling cluster to which they belong, based on each country’s classification.
Example and interpretation:
“According to the 2018 Nigeria DHS data, under-five mortality was 155 per 1,000 live births in rural areas compared with 86 in urban areas, indicating an 80% higher rate among rural residents.”
“In urban areas, the coverage of skilled birth attendance was 69%, while in rural areas it was only 30%, resulting in a difference of 39 percentage points.”
Combining wealth and residence dimensions (urban/rural) is an important strategy to present health inequalities that would not be detected through isolated stratifications. This double stratification allows identification of intersectional patterns and groups at higher vulnerability, such as the poorest in rural areas, thereby supporting targeted policy formulation. However, its use requires caution, particularly in surveys with small samples.
Definition and classification:
Double stratification involves analyzing results by both stratifications simultaneously: wealth quintiles (economic stratification) and urban/rural residence (geographic stratification).
Example and interpretation:
“In the 2018 Guinea DHS, coverage of skilled birth attendance ranged from 96% among the richest in urban areas to 28% among the poorest in rural areas. Even within rural areas, a gradient was observed: 95% in the richest quintile versus 28% in the poorest.”
Education directly affects women’s ability to access information, understand medical advice, and make informed decisions. In this context, analyzing inequalities related to this stratifier is essential for identifying barriers to healthcare and informing equity-focused policies, particularly in maternal and child health, where education is strongly associated with service utilization, care practices, and health outcomes.
Definition and classification:
This stratifier can be categorized in two ways:
Three education levels:
Four education levels:
Note: Attendance at Koranic or other traditional schools is not considered.
Example and interpretation:
“In low- and middle-income countries, lower levels of formal maternal education are positively associated with breastfeeding, while the opposite is observed among women with higher levels of formal education.”
“A positive gradient was observed in length at birth across maternal education categories, with the shortest lengths recorded among boys and girls whose mothers belonged to the lowest education category.”
Inequalities related to age highlight specific biological, social, and behavioral risks, especially among adolescents. This group generally experiences greater vulnerability and faces challenges in accessing high-quality care, which can lead to worse health outcomes. Therefore, analyzing this stratifier is crucial for understanding disparities in maternal and child health, as age significantly affects risks and access to care throughout the prenatal, delivery, and post-natal periods.
Definition and classification:
Women’s or mothers’ age is expressed in completed years, using age at child’s birth for prenatal, delivery, and postnatal care indicators, and the age at the time of the survey for other indicators. Categories include:
Current age: 15-17, 18-19, 20-34, 35-49 years; pooled results for 15-19 and 20-49 years are also available.
Mother’s age at birth: 15-17, 18-19, 20-49 years; or 15–19, 20–34, 35–49 years.
Example and interpretation:
“In the 2022 Bangladesh DHS and the 2021 Cambodia DHS, using a combined maternal and neonatal health indicator, the prevalence of receiving zero interventions varied by women’s age. In Bangladesh, the proportion was highest among women aged 35-49 years (18.9%) and lowest among those aged 15-17 years (9.9%), representing a difference of 9.0 percentage points (pp). Conversely, in Cambodia, the highest prevalence occurred among women aged 15-17 years (3.7%), while the lowest was observed in the 35-49 age group (0.6%), leading to a difference of 3.1 pp. These findings suggest that the age groups most vulnerable to receiving no interventions differ between countries.”
Geographic stratification by subnational units is essential for enabling countries to identify and address their specific health needs. By analyzing data at the level of regions, states, provinces, districts, or other administrative divisions, it becomes possible to uncover inequalities that are often masked by national averages.
This approach makes it possible to identify geographic disparities in access to services and health outcomes, monitor progress at the local level, highlight underperforming areas, and support a more efficient allocation of resources.
Definition and classification:
Estimates for key indicators can be produced at the level of administrative units, usually the first or second level of decentralization (regions or provinces). In many countries, subnational stratification aligns with boundaries used for policy planning and service delivery.
Example and interpretation:
“In the 2021 Nigeria MICS survey, Composite Coverage Index (CCI) estimates ranged from 26.3% in Sokoto to 76.9% in Ebonyi, highlighting substantial geographic disparities in the coverage of essential maternal and child health services.”
Disaggregating estimates by child’s sex is crucial for uncovering gender inequalities that may begin as early as pregnancy and intensify during childhood. Although “sex” generally reflects a set of biological characteristics, in analytical contexts it helps to demonstrate how gender norms, based on socially assigned roles and values for boys and girls, differently affect access to health, nutrition, care, and development. In many contexts, girls may face disadvantages such as neglect or violence, while boys may also be affected by norms that limit their development or expose them to risks. Male child preference, still present in some societies, further exacerbates these inequalities.
Definition and classification:
The child’s sex is reported directly in the household survey questionnaire and classified as male or female.
Example and interpretation:
“Unexpected differences in mortality sex ratios may reflect community sex preference and gender discrimination in child health care. (…) Countries where girls were more likely to die than boys, accounting for overall mortality levels, were also countries where boys were more likely to receive health care than girls.”
Ethnic and racial minorities, Indigenous peoples, and other historically marginalized groups often face discrimination and forms of vulnerability resulting from historical, social, and political processes that shape access to rights, opportunities, and resources, thereby exacerbating health inequities. Analyses of health inequalities must take ethnicity and Indigenous identity into account within the specific context of each country to ensure that inequities are accurately identified and effectively addressed through interventions.
Definition and classification:
Categories related to ethnicity correspond to those recorded in national surveys and may vary by country. Depending on the context and the method of data collection, this variable can be defined based on:
Example and interpretation:
“Significant ethnic disparities in child survival were identified in more than two-thirds of the countries studied. Differences between ethnic groups tended to increase with child age at death, with greater disparities observed in deaths of children aged 1–4 years than for younger children.”
“Systematic disadvantages for Afrodescendants were found for demand for family planning, early marriage, household handwashing, and sanitation facilities. Wealth gaps among Afrodescendants were wider than those observed for non-Afrodescendants for most indicators and across all countries.”
Religion is an important social and cultural determinant of health, influencing health-seeking behavior, disease perception, fertility patterns, and decisions related to reproductive, maternal, neonatal, and child health (RMNCH).
In many contexts, religion is not merely a matter of personal belief but also shapes social norms and gender roles, affects health service utilization, determines dietary practices, and influences trust in public institutions. For instance, religious restrictions may lead families to prefer traditional birth attendants over health professionals, to hesitate in vaccinating children, or adopt specific contraceptive methods, directly affecting access to and quality of RMNCH care.
Definition and classification:
Religion-related categories are those recorded in national surveys and may vary from country to country.
Disaggregating health indicators by religion allows us to:
Example and interpretation:
“In the 2018 Guinea survey, disaggregation of RMNCH indicators by religion revealed substantial disparities: the prevalence of ‘no maternal and neonatal care’ — mothers and newborns who did not receive any of the essential interventions — was 28.0% among Muslims, but much lower among Christians (13.8%) and those reporting no religion (3.9%). For zero-dose DPT vaccination, a similar pattern was observed: 39.1% among Muslims versus 31.0% among Christians.
These findings indicate that religion correlates with significant differences in coverage. While religion itself is not a direct cause of poor health outcomes, it frequently intersects with other factors such as ethnicity, wealth, education, and geographic location.”
Empowerment is the process by which women strengthen their autonomy, expand their active participation in strategic decisions affecting their lives, and gain greater access to rights and resources in contexts historically marked by gender inequalities.
Definition and classification:
In surveys, women’s empowerment is measured through the SWPER (Survey-based Women’s emPowERment) Global, an index based on 14 items from the individual questionnaire and composed of three domains:
I) attitude toward violence, II) social independence, and III) decision-making.
For each domain, a continuous score is generated, where zero represents the average level of empowerment in the countries on which SWPER is based. Positive values indicate an empowerment level above the average of low- and middle-income countries, whereas negative values indicate a level below that average. From this continuous score, women can be classified into low, medium, or high empowerment levels. Further details on the construction of the SWPER can be found here.
Note: Data available exclusively from DHS surveys.
Example and interpretation:
“Children of more empowered women are less likely to be left without vaccination in low- and middle-income countries: children with highly-empowered mothers presented lower prevalence of zero-dose than those with less-empowered mothers.”
“The prevalence of women who had at least one daughter who had undergone female genital mutilation/cutting practices was consistently higher among low empowered women.”