Effective Coverage in Urban Poor Areas the dhs program

Why study urban poverty and effective coverage?

Two-thirds of the global population are forecasted to live in urban areas by 2050. Advantages of urban living are not experienced equally by everyone living in cities. Urban poor people have been found to experience poorer health outcomes than urban non-poor, due to lack of access and low quality of health care.

This study uses Demographic and Health Survey (DHS) and Service Provision Assessment (SPA) data to calculate effective coverage cascades for antenatal care (ANC) and sick child care for urban poor and urban non-poor populations in six countries. Effective coverage is a health system measure which incorporates multiple aspects of health system performance–need, use, readiness, and quality–into one measure shown as a cascade (see Figure 1 below). The first column represents the target population who has a need for the health service. This study concerns two target populations: women who gave birth at least once in the previous two years and children under age 5 with symptoms of of diarrhea or acute respiratory infection (ARI) in the two weeks before the survey. The cascade continues with the second column representing service contact coverage: those in the target population who seek care at a health facility.

Figure 1. Effective coverage cascade for antenatal care (ANC) and sick child care. Each step in the cascade is the product of the measure in the column and all measures in previous columns.

Input-adjusted coverage incorporates a measure of health facility readiness and represents those who sought care at a health facility that has all necessary inputs to deliver the health service. In this study, intervention-adjusted coverage includes women who received four or more ANC visits at a health facility that was “ready” and children under age 5 diagnosed with diarrhea or pneumonia who received appropriate treatment from a health facility that was “ready”. Finally, quality-adjusted coverage includes a measure of quality of care. Quality-adjusted care represents the proportion of the target population who received the complete intervention at a high level of quality from a health facility that was ready to provide the health service.

Effective coverage cascades were calculated for ANC for six countries and sick child care for five countries (excluding Ethiopia). Nonoverlapping confidence intervals were used to determine significant differences in effective coverage between urban non-poor and urban poor populations.

Read more:

Click a study country to see a map of household populations living in urban areas and households in urban poor clusters by region, ANC effective coverage cascades for urban poor and urban non-poor women for each of the six study countries, and sick child care effective coverage cascades for urban poor and urban non-poor children for all countries except Ethiopia.

Afghanistan

Figure 2. Map of Afghanistan showing percent of household population living in urban areas by province, and percent of urban household population living in urban poor clusters by province

For Afghanistan, only seven provinces were included in the SPA (Kabul, Nangarhar, Paktya, Kunduz, Balkh, Kandahar, and Hirat), so this analysis was restricted to these provinces.

Over half of the household population in Afghanistan live in urban areas. In nearly all provinces included in this analysis, most of the urban household population lives in urban poor clusters. Only in Hirat are fewer than half (42%) of urban households poor.

Figure 3. ANC effective coverage cascades for urban non-poor and urban poor women in Afghanistan

Nearly three-quarters of both urban non-poor and urban poor women in Afghanistan received at least one ANC visit from a skilled provider (gray bars in Figure 3), while 65% of urban non-poor and 68% of urban poor women received care from a facility ready to provide ANC (light purple bars in Figure 3). These first two measures, of service contact coverage and input-adjusted coverage of ANC services, do not vary much between urban poor and urban non-poor women. Only about one-quarter of urban women in Afghanistan received the complete intervention, with 28% of urban non-poor and 23% of urban poor women receiving at least four ANC visits at a health facility (intervention-adjusted coverage). Even fewer women in Afghanistan received quality ANC according to recommended standards.

Of the six study countries, Afghanistan has the highest overall ANC readiness score but the lowest overall ANC process quality score.

Figure 4. Sick child care effective coverage cascades for urban non-poor and urban poor children in Afghanistan

Urban poor children consistently have lower sick child care coverage than urban non-poor children in Afghanistan, though differences are not significant as confidence intervals overlap. Fifty-five percent of urban non-poor children under age 5 with diarrhea or acute respiratory infection (ARI) symptoms sought care at a health facility, compared to 44% of urban poor children (gray bars in Figure 4). Only 37% of urban non-poor and 29% of urban poor children received the complete intervention of appropriate sick child care (intervention-adjusted coverage, medium purple bars in Figure 4), and fewer received quality sick child care according to recommended standards.

Effective coverage for ANC and sick child care in Afghanistan does not significantly vary between urban non-poor and urban poor women at any levels of either cascade.

Return to top
Jump to conclusion

DRC

Figure 5. Map of DRC showing percent of household population living in urban areas by province, and percent of urban household population living in urban poor clusters by province

As seen in the map above, many provinces in Democratic Republic of the Congo (DRC) have small urban populations. Even so, a large proportion of the urban population is poor in most provinces. More than half of urban households live in urban poor clusters in all provinces except Kinshasa (16%), Bas-Congo (2%), Kasai-Oriental (42%), and Nord-Kivu (18%).

Figure 6. ANC effective coverage cascades for urban non-poor and urban poor women in DRC

Service-contact coverage for ANC in DRC is high; more than 9 in 10 women received at least one ANC visit. However, when readiness of facilities to provide ANC is taken into account, input-adjusted coverage (light purple bars in Figure 6) drops below 50%. These measures are similar for both urban non-poor and urban poor women, but significant disparities emerge at intervention-adjusted coverage. Among urban non-poor women, 33% received the complete intervention of at least four ANC visits, compared to only 23% of urban poor women (medium purple bars in Figure 6). There is also a significant difference in quality-adjusted coverage of ANC, which is 17% among urban non-poor women, compared to 12% of urban poor women (dark purple bars in Figure 6).

Figure 7. Sick child care effective coverage cascades for urban non-poor and urban poor children in DRC

Sick child care in DRC is very low, and there is no significant difference between urban poor and urban non-poor children at any level of coverage. Only three in ten children under age 5 with diarrhea or acute respiratory infection (ARI) symptoms sought care at a health facility (gray bars in Figure 7), and fewer than two in ten did so from a health facility that was "ready" to provide sick child care (light purple bars in Figure 7). Intervention-adjusted coverage, which accounts for service contact coverage, facility readiness, and receipt of complete intervention, is 12% for both urban poor and urban non-poor children (medium purple bars in Figure 7). When quality is factored in, quality-adjusted coverage falls to just 6%.

Return to top
Jump to conclusion

Ethiopia

Figure 8. Map of Ethiopia showing percent of household population living in urban areas by region, and percent of urban household population living in urban poor clusters by region

For this analysis, both Tigray and Sidama regions were excluded. A large proportion of urban households in Ethiopia are poor. Of the ten provinces included, in only three–Harari, Addis Ababa, and Dire Dawa–do fewer than half of urban households live in urban poor clusters. that , over two-thirds (68%) of women live in urban areas, and most urban women are poor (84%). In five regions–Amhara, Afar, Gambella, SNNPR, and Oromiya–all urban households in this analysis are poor.

Figure 9. ANC effective coverage cascades for urban non-poor and urban poor women in Ethiopia

There are significant disparities between urban poor and non-poor women at three of the four steps of ANC effective coverage. To begin, 95% of urban non-poor women receive at least one ANC visit from a skilled provider, compared to 82% of urban poor women (gray bars in Figure 9). For intervention-adjusted coverage, nearly twice as many urban non-poor women received at least four ANC visits at a health facility compared to urban poor women (50% versus 27%, medium purple bars in Figure 9). Quality-adjusted coverage was also significantly higher among urban non-poor women than urban poor women.

The 2019 Ethiopia DHS was a mini-DHS and did not include the necessary data to calculate the sick child coverage indicators.

Return to top
Jump to conclusion

Haiti

Figure 10. Map of Haiti showing percent of household population living in urban areas by department, and percent of urban household population living in urban poor clusters by department

In contrast to Ethiopia and other countries in this analysis, five of eleven regions in Haiti–Rest-Ouest, Sud-Est, Nord, Sud, and Nippes–have no urban households in urban poor clusters. Grand-Anse has the highest proportion of urban poor households, with one in three urban households living in urban poor clusters.

Figure 11. ANC effective coverage cascades for urban non-poor and urban poor women in Haiti

Haiti is the only country where there is a significant difference in ANC effective coverage between urban poor and urban non-poor women at all steps of the cascade (e.g., confidence intervals do not overlap). Urban non-poor women have higher service-contact coverage, with 95% receiving at least one ANC visit from a skilled provider compared to 85% of urban poor women (gray bars in Figure 11). Overall, quality-adjusted coverage is 29% for urban non-poor women and 18% for urban poor women (dark purple bars in Figure 11).

Figure 12. Sick child care effective coverage cascades for urban non-poor and urban poor children in Haiti

Sick child care effective coverage measures are higher for urban poor in Haiti than for urban non-poor children at each step of the cascade. These differences are not significant due in part to wide confidence intervals, as few children under age 5 live in urban poor clusters, and fewer had diarrhea or acute respiratory infection (ARI) symptoms in the two weeks before the survey.

Service contact coverage is low in Haiti; fewer than half of children with symptoms of diarrhea or ARI sought care at a health facility (gray bars in Figure 12). Haiti has the lowest intervention-adjusted coverage of all six study countries, at 10% among urban non-poor and 15% among urban poor children (medium purple bars in Figure 12).

Return to top
Jump to conclusion

Nepal

Figure 13. Map of Nepal showing percent of household population living in urban areas by province, and percent of urban household population living in urban poor clusters by province

Nearly one-third of the population in Nepal included in this analysis live in urban areas, and very few live in urban poor clusters. Less than 10% of the urban household population are poor in all provinces except Madhesh Province, where 13% of urban households live in urban poor clusters.

Figure 14. ANC effective coverage cascades for urban non-poor and urban poor women in Nepal

Nearly all women with a live birth in the two years before the survey in Nepal received at least one ANC visit from a skilled provider for their most recent birth, regardless of urban poverty status (gray bars in Figure 14). Effective coverage estimates do not vary significantly among urban poor and urban non-poor women at any step in the cascade. Nepal has the highest quality-adjusted coverage of any of the six study countries, 37% for urban non-poor women and 34% for urban poor women, (dark purple bars in Figure 14).

Figure 15. Sick child care effective coverage cascades for urban non-poor and urban poor children in Nepal

In contrast to ANC, sick child service contact coverage in Nepal is the lowest across all study countries. Only 16% of urban non-poor children and 9% of urban poor children with symptoms of diarrhea or acute respiratory infection (ARI) in the two weeks before the survey sought care at a health facility (gray bars in Figure 15). Like Haiti, differences between urban non-poor and urban poor children in Nepal are not significant at any step in the cascade. The small number of children living in urban poor clusters in Nepal results in wide confidence intervals for the urban poor sick child care effective coverage estimates.

Return to top
Jump to conclusion

Tanzania

Figure 16. Map of Tanzania showing percent of household population living in urban areas by zone, and percent of urban household population living in urban poor clusters by zone

In Tanzania, while nearly two-thirds of the population included in this analysis live in urban areas, only a small proportion of the urban population are poor. Of the nine zones included in this analysis, only Lake zone has more than 10% of urban households living in urban poor clusters.

Figure 17. ANC effective coverage cascades for urban non-poor and urban poor women in Tanzania

More than 9 in 10 Tanzanian women with a live birth in the two years before the survey received at least one ANC visit from a skilled provider (gray bars in Figure 17). More than half of women sought care from a facility that was ready to provide ANC (light purple bars in Figure 17). Both service-contact coverage and input-adjusted coverage are the same for urban non-poor and urban poor women. Disparities emerge at intervention-adjusted coverage, (medium purple bars in Figure 17), though differences are not significant.

Figure 18. Sick child care effective coverage cascades for urban non-poor and urban poor children in Tanzania

There is no significant difference between urban poor and urban non-poor children under age 5 at any level of sick child care coverage in Tanzania. The gap in coverage between urban non-poor and urban poor children narrows at each level of the cascade. There is a nearly nine percentage point difference between urban non-poor and urban poor children with diarrhea or acute respiratory infection (ARI) symptoms who sought care at a health facility (gray bars in Figure 18). Later in the cascade, the gap in quality-adjusted coverage is less than two percentage points (dark purple bars in Figure 18).

Return to top
Jump to conclusion

Conclusion

Significant differences in effective coverage of ANC between urban non-poor and urban poor women were observed in half of the study countries: DRC, Ethiopia, and Haiti. These three countries vary greatly in the proportion of urban poor, from 84% of urban women in Ethiopia living in urban poor clusters to 43% of urban women in DRC and just 6% of urban women in Haiti who are poor.

No significant differences were observed in sick child care effective coverage between urban poor and urban non-poor children for the five countries in this analysis. This is due in part to small sample sizes of sick children. In addition, only the service contact coverage measure used in the sick child care effective coverage cascade can be calculated separately for urban poor and urban non-poor children. The other three measures come from the SPA, which did not collect information on the urban poverty status of clients. This means that differences in sick child care coverage between urban poor and urban non-poor children are likely larger than the differences observed in this study.

This study extends the literature on equity in effective coverage by comparing effective coverage cascades for two essential primary health care services by urban poverty status. Future research should improve on methodological limitations, such as using external measures of poverty to classify catchment areas around facilities to better illuminate intra-urban inequities.

Access the full report and brief

Riese, Sara, Shireen Assaf. 2024. Effective Coverage in Urban Poor Areas. DHS Analytical Studies No. 89. Rockville, Maryland, USA: ICF.

Code to construct the ANC and child health indicators used in this analysis can be found in chapters 9 and 10 on The DHS Program's Code Share Library on GitHub at https://github.com/DHSProgram/DHS-Indicators-Stata

The code for this analysis was adapted from earlier effective coverage analysis code which can also be found on The DHS Program's Code Share Library on GitHub at https://github.com/DHSProgram/DHS-Analysis-Code/tree/main/EffectiveCoverage

© Eric Isselée / Adobe Stock

© Fredy Thürig / Adobe Stock

© robertharding / Adobe Stock

© qojoo / Adobe Stock

This Adobe Express was prepared by The Demographic and Health Surveys (DHS) Program.

The information provided on this web application is not official U.S. Government information and does not represent the views or positions of the U.S. Agency for International Development or the U.S. Government.