Table 2 Sampling procedures – study 1

1

Based on the design effect (1.5) applied in the sample size calculation and the total sample size (434), it is methodologically sound to recruit 20 participants per cluster (health facility). Accordingly, 22 health facilities will be selected

Based on the design effect (1.5) applied in the sample size calculation and the total sample size (408), it is methodologically sound to recruit 20 participants per cluster (health facility). Accordingly, 20 health facilities will be selected

2

We stratify facilities by ownership (public and private). We regard non-governmental health facilities as private facilities as the proportion of women who give birth in those settings in Addis Ababa is very low

We stratify facilities by ownership (public, private, and faith-based)

3

We allocate the total sample size to the two sectors (public and private) based on the proportion of women who give birth in those settings (74% in public and 26% in private facilities). Accordingly, 330 and 114 women will be surveyed from public and private facilities, respectively. As per the cluster size in Step 1, 16 public and 6 private facilities will be selected

We allocate the total sample size to the three sectors (public, private, and faith-based) based on the proportion of women who give birth in those settings (75% in public, 20% in private and 2% in faith-based facilities). Accordingly, 306, 82, and 20 women will be surveyed from public, private, and faith-based facilities, respectively. As per the cluster size in Step 1, 15 public, 4 private, and 1 faith-based facilities will be selected

4

We allocate the 16 public health facilities to hospitals and health centres based on the distribution of births in public health facility-based births in the city – 48% in hospitals and 52% in health centres. Accordingly, 7 hospitals and 9 health centres will be randomly selected. Selection of the public hospitals and public health centres will again be made proportionally based on the number of childbirths at individual facilities in the same category (hospital or health centre) provided in the last quarter of 2022 (this information is available from Addis Ababa Regional Health Bureau)

We allocate the 15 public health facilities to hospitals, health centres, and health posts based on the distribution of births in public health facility-based births in the city – 34% in hospitals, 63% in health centres, and 3% in health posts; given the small proportion of births in health posts, we will merge these with health centres. Accordingly, 5 hospitals and 10 health centres will be randomly selected. Selection of the public hospitals and public health centres will again be made proportionally based on the number of childbirths at individual facilities in the same category (hospital or health centre) provided in the last quarter of 2022 (this information is available from Conakry Regional Department of Health)

5

We select the 6 private hospitals based on probability proportionate to size using the number of childbirths at individual private hospitals provided in the last quarter of 2022

We select the 4 private facilities based on probability proportionate to size using the number of childbirths at individual private facilities provided in the last quarter of 2022. Similarly, one faith-based facility will be selected based on probability proportionate to size using the number of childbirths at individual faith-based facilities provided in the last quarter of 2022

6

Eligible women accessing antenatal care in sampled health facilities in their third trimester of pregnancy will be consecutively invited to participate in the survey until the required sample (n = 20) from each selected facility is achieved

Eligible women accessing antenatal care in sampled health facilities in their third trimester of pregnancy will be consecutively invited to participate in the survey until the required sample (n = 20) from each selected facility is achieved