(Last updated on 2023-01-17)

SUMMARY

#1: Checked for three potential changes between rounds 3 and 4
1. In Round 3 (and before), countries with an “increase”: might have said “Not applicable” or might have not responded. More likely “Not Applicable”, since the % non-missing was same across EHS questions??
2. In Round 4, is there self-selection in participation?
3. Among those who participated, percent missing in R3 and R4 is similar, but higher than in R2.

#2 & #3: Checked for different percent distributions to compare cross-sectional snapshots
1. Overall stories are same across different approaches to restrict denominators - using “EHS” services as an example.
2. In Round 4, “increase” is minimal across services.
3. In summary, consider using an approach with minimal analysis/assumptions.
- % decreases, among all who participated in the survey (second figure under #3)
- use individual services as tracer (e.g., facility immunization, as opposed to immunization index)

P.S. Comparison for R3 vs. R4 presented. Overall conclusion is same, when R2 is included.

1. Considerations

1.1. Changes in Q

See Annex.

1.2. Changes in participation

  • In Round 2, 135 countries participated.

  • In Round 3, 132 countries participated.

  • In Round 4, 93 countries participated.

  • In total, 149 countries participated in either Round 3 or Round 4. Among them, only 76 participated in both rounds.

  • In total, 170 countries participated in any of the Round 2, Round 3, and Round 4. Among them, only 64 participated in all three rounds 2-4.

1.3. Among those participated, changes in non-missing for specific services

Question: is missing self selection, in either positive or negative way?

2. Check distribution betweem R3 vs. R4 - among those responded (i.e., pct excluding missing)

NOTE:
1. n for each service varies (about 70% of total), but more or less similar.
2. LEFT panel includes facilities that participated in each round (n=132 in R3; n=93 in R4).
3. RIGHT panel includes facilities that participated in both rounds (n=76).
4. Legend

3. Check distribution betweem R3 vs. R4 - among all who participated (i.e., pct including missing)

NOTE:
1. n is same for all services within a same round.
2. LEFT panel includes facilities that participated in each round (n=132 in R3; n=93 in R4).
3. RIGHT panel includes facilities that participated in both rounds (n=76).
4. Legend

Annex: Q changes between R3 and R4

ROUND 3:
7.a Please indicate the level of disruption due to COVID-19 during the previous 6 months (i.e. percentage of users not served as usual compared to before the pandemic) for the following service areas.

[ ] More than 50%
[ ] 26-50%
[ ] 5-25%
[ ] Less than 5% (including 0%)
[ ] Do not know
[ ] Not applicable

ROUND 4:
13.a Please indicate the current level of disruption due to COVID-19 (i.e. percentage change in service volume compared to before the pandemic) for the following service areas.

[ ] Decrease of more than 50%
[ ] Decrease of 26-50%
[ ] Decrease of 5-25%
[ ] No disruption (less than 5% increase or decrease)
[ ] Increase of 5% or more
[ ] Do not know
[ ] Not applicable