Figure 1 shows the spatial distribution of the poorest and richest districts according to our classification. The top (non-poor) 60% of districts (marked in blue) are predominantly located in the city center, and the outskirts with the bottom (poor) 40% of districts are located in between (in pink). Most public health centers and all public hospitals are located in the top 60% of districts. Private facilities are not displayed on the map.

Fig. 1

Map showing the poverty status of districts within Addis Ababa city and the relative location of public health facilities. Note: the 116 districts are within ten sub-cities

Health Facility Distribution Across the Bottom 40% and Top 60% of Districts

Out of a total of 1133 health facilities in Addis Ababa, 384 (33.9%) health facilities reported at least once via DHIS2 between 2019 and 2021. Of these, only 133 (35%) health facilities (all 110 public and 23 private) were geo-located and used for the analysis. A total of 79 (59.4%) health facilities (42.9% health centers, 15.8% hospitals, and 0.8% specialty centers) are located in the top 60% of districts, while 54 (40.6%) health facilities (27.1% health centers, 8.3% hospitals, 4.5% clinics, and 0.8% specialty center) are found in the bottom 40% of districts.

Maternal and Child Health Services Coverage

Major differences in coverage were observed for all maternal health service coverage indicators. While coverage estimates in the top 60% of districts approximated universality, and in the case of C-section far exceeded population need, low coverage in the bottom 40% of districts is most notable for SBA (54%) and also ANC4 (67%). On the other hand, the C-section rate of 9% suggests that the need for C-sections (between 10 and 15% of births) is almost met in the bottom 40% of districts. On the other hand, both poor and non-poor districts are performing well in child immunization services. There is no significant difference between the bottom 40% and the top 60% in child vaccination coverage (Table 1).

Table 1 Aggregate maternal and child health services coverage in the poorest (bottom 40%) and wealthiest (top 60%) districts of Addis Ababa by poverty status, 2019–2021

Figure 2 presents trends of maternal health service coverage over 3 years among the non-poor (top 60%) and poor (bottom 40%) districts. ANC4 service coverage during 2019 to 2021 was higher in the top 60% of districts than in the bottom 40% of districts. Between 2019 and 2020, the annual rate of change in ANC4 coverage was 4.6%, but this decreased to − 0.4% in the following year. However, the overall average change in ANC4 coverage was 2.1%, which was not statistically significant.

Fig. 2

Trends of maternal health service coverage in urban Addis Ababa, 2019–2021. Note: ANC4, at least four antenatal care visits; IFA, iron and folic acid supplementation; SBA, skilled birth attendance; C-section, cesarean section; PNC, postnatal care within 2 days after birth; PAC, post-abortion care; SAC, safe abortion care

In districts in the bottom 40%, ANC4 coverage increased by 9.4% annually between 2019 and 2020 and by 2.4% annually between 2020 and 2021, for an average change of 5.9%. This suggests that ANC4 coverage is trending upward in both the top 60% and bottom 40% of districts, with a slightly higher rate of increase in the latter.

Skilled birth attendance in the top 60% (non-poor) of districts was 100% in 2019, 2020, and 2021. This implies that maximum coverage was achieved in the top 60% of districts. The situation was substantially worse in the bottom 40% (poor) of districts, where SBA coverage remained below 60% during 2019 to 2021. In terms of SBA, the annual change in the bottom 40% of districts from 2019 to 2020 was − 5.2%, and from 2020 to 2021, it was 6.9%. The SBA’s average change between 2019 and 2021 was 0.8%, indicating a modest upward trend.

PNC service coverage was also higher near 96.9% in the non-poor districts, while it remained below 63% in the bottom 40% (poor) of districts during 2019–2021. Overall, maternal health services coverage was considerably higher in the top 60% of districts than in the bottom 40%. Between 2019 and 2020, postnatal care utilization coverage in the top 60% of districts decreased by 0.3% annually. However, it increased by 7.5% annually from 2020 to 2021, for an average change of 3.6% over the 2-year period.

In the bottom 40% of districts, postnatal care utilization coverage decreased by 2.1% annually from 2019 to 2020 and increased by 6.6% annually from 2020 to 2021, for an average change of 2.3% over the 2-year period. Overall, maternal health service coverage remained higher in the wealthiest (top 60%) than poorest (bottom 40%) districts (Supplement Fig. 1 and Table 1).

In addition to using DHIS2, we analyzed five rounds of the Ethiopia Demographic and Health Survey (EDHS) from 2000 to 2019 to understand the disparity in coverage of key maternal health interventions between poorer households (bottom 40%) and richer households (top 60%).

Figure 3 shows that ANC4 coverage has consistently been higher for pregnant women in richer households than among those in poorer households in Addis Ababa over the past 20 years.

Fig. 3

Trends of maternal health service coverage for ANC4, IFA, SBA, C-section, and PNC comparing the bottom 40% and top 60% in Addis Ababa, EDHS 2000–2019. Note: ANC4, at least four antenatal care visits; IFA, iron and folic acid supplement; SBA, skilled birth attendance; C-section, cesarean section; PNC, postnatal care

IFA supplementation has been higher among women in poorer households than women in richer households, except in 2005. SBA coverage was higher among urban richer women than urban poorer women in the EDHS: 2000, 2006, and 2011. Notably, although SBA coverage has been lower among women in poorer households, there was a larger increase among this group from the 2011 EDHS (75.6%) to the 2016 EDHS (97.0%). The gap between the richer and poorer groups closed between EDHS-2016 and EDHS 2019.

Coverage for C-section was lower among women in the bottom 40% than among women in the top 60% across all five EDHS periods (2000 to 2019). Rates started low, at around 3.9% in the bottom 40% in 2000, before increasing to 11.7% in 2005 and beyond. The rate of C-sections is higher than the recommended WHO level among women in the top 60%, while remaining within the range of meeting the need (between 10 and 15%) among women in the bottom 40% between 2005 and 2019.

Postnatal care within 2 days has exhibited an inconsistent trend both among women in the bottom 40% and women in the top 60% over the past 20 years. Generally, PNC coverage is higher among women in the top 60% across all five EDHS periods.

As shown in Fig. 4, BCG vaccination coverage in the bottom 40% of districts was 92% in 2019 and 2020 and then increased to 97% in 2021. Penta3 vaccination coverage increased from 94% in 2019 to 95% in 2020 and declined back to 94% in 2021. From 2019 to 2021, the BCG vaccination rate in the top 60% of districts stays 100%. The Penta3 vaccine shows a declining trend, with an average change of − 1.0% between 2019 and 2021 and an annual change rate of − 2.1% between 2020 and 2021.

Fig. 4

Trends of child vaccination coverage in urban Addis Ababa, 2019–2021. Note: BCG, Bacillus Calmette–Guérin vaccine; Penta 3, third dose of pentavalent vaccine; MCV1, measles-containing vaccine dose one

Measles vaccination coverage decreased from 92% in 2019 to 90% in 2020 and remained at that level in 2021. Conversely, between 2019 and 2020, MCV1 grew at an annual change rate of 6.7%. Subsequently, there was a decline with an average change rate of 0.7% and an annual change rate of − 5.3% between 2020 and 2021 (P-value above 0.05).

Within the districts that comprise the lowest 40%, the coverage of BCG vaccination was 92.0% in 2019 and 2020, but it rose to 97.0% in 2021, signifying a 5.4% increase and an average shift of 2.7% from 2019 to 2021. The annual rate of change in Penta3 vaccine coverage was 1.1%, − 3.2% between 2019 and 2020, and 2020 and 2021, respectively. Between 2019 and 2021, there was a − 1.1% average change in Penta3. There has been a decline in the measles vaccine coverage between 2019 and 2020, with an annual change rate of − 2.2% and an average change of − 1.1%, suggesting a downward trend.

Overall, high vaccination rates are found in both the top 60% and bottom 40% of districts, indicating a narrowing of the immunization service gap between poor and non-poor districts. BCG coverage is generally trending upward in both districts (Fig. 4).

Spatial Distribution of Maternal and Child Health Service Coverage

Figure 5 below presents maternal and child health service coverage by district. The majority of the top 60% of districts have ANC4 coverage exceeding 60%, while coverage of ANC4 within the bottom 40% of districts falls in the range of 20–60%. The average of both the top 60% and bottom 40% of districts have high (> 80%) IFA supplementation coverage.

Fig. 5

Geo-spatial distribution of MNCH services coverage at the sub-city level, 2019–2021

Regarding SBA and PNC, an average of the top 60% of districts have higher coverage compared to the bottom 40% of districts. C-sections, which can be lifesaving interventions for mothers at risk, are higher in the top 60% of districts than the bottom 40% ones. However, while there is overuse of C-section services in some of the top 60% of districts, the service is underused in the bottom 40%.

In contrast to the distinct differences over several indicators of maternal health service, both the top 60% and bottom 40% of districts have high child immunization coverage greater than 80% BCG, Penta3, and MCV1 coverage in both classifications (Fig. 5).

Newborn and Under-5 Mortality

Figure 6 illustrates the levels and trends of NMR and U5MR per 1000 live births for urban Addis Ababa across the two decades from 1996 to 2015 (each data point represents the 10-year period preceding each EDHS: 2000, 2005, 2011, 2016, and 2019). These rates are analyzed for the overall population, as well as specifically for the poorest (bottom 40%) and the wealthiest (top 60%) households in urban Addis Ababa. These findings reveal that both NMR and U5MR exhibited substantial declines, with NMR decreasing by 65% and U5MR by 77% over the two decades. This translates to an average annual reduction of 5% for NMR and 8% for U5MR. However, the results diverge when considering household wealth status.

Fig. 6

Trends of NMR and U5MR in urban Addis Ababa, 1996–2015. Note: Data source: five rounds of EDHS surveys: 2000–2019, and each data point represents 10-year period preceding each EDHS: 2000–2019, and the year represents the midpoint within the 10-year period preceding each survey. NMR, neonatal mortality rate; U5MR, under-5 mortality rate

The decline in U5MR is characterized by a significant drop among the poorest (bottom 40%) households, declining from 144 to 45 deaths per 1000 live births, marking a 69% reduction at an average annual rate of 6%. In contrast, U5MR among the wealthiest (top 60%) households decreased from 88 to 14 deaths per 1000 live births, reflecting a 84% reduction at an average annual rate of 10% between EDHS 2000 and EDHS 2019.

Similarly, NMR experienced a 49% decline among the poorest (bottom 40%) households, decreasing from 51 to 26 deaths per 1000 live births, with an average annual drop of 4%. Among the richest (top 60%) households, NMR declined by 66%, reducing from 35 to 12 deaths per 1000 live births, at an average annual rate of 6%, during the same time frame.