The Ministry of Health (MOH) of Harmonia has notified the steering committee that they would like a survey with estimates for each of the 100 districts in the country, as well as a national level estimate. They are thinking that they would ideally have a confidence interval no wider than +/-5% in each district.

You request additional information including:

the primary vaccine dose of interest,
the expected level of coverage,
the number of households to visit to find an eligible child,
the proportion of eligible respondents who are likely to refuse and their estimate of intracluster correlation coefficient (ICC), which relates to the design effect (DEFF).

Note: DEFF is a measure of variability due to selecting survey subjects by any method other than simple random sampling. Usually, cluster surveys have a design effect greater than one, meaning the variability is higher than for simple random sampling.

The Committee provides the following information and requests that you calculate the sample size.

The Committee decides that DTP/pentavalent 3 is the primary vaccine dose of interest
They assume that current coverage is 85% for the purpose of the sample size calculation.
They will also have a secondary question related to the prevalence of home based records, but given that this is secondary it will not be used for calculating the sample size.
The design effect will be 4 (assuming 10 children per cluster and an intra-cluster correlation or ICC of 1/3), a conservative choice given that there is not a recent EPI survey that can be used to obtain the ICC.
On average the team is likely to have to visit 5 houses to find an eligible child and the committee anticipates that 15% of households with eligible children will either be away from home or otherwise not cooperative.

Given this information,

Q#1. what would be the overall sample size?

• (100)(265)(4)=105,000 will be overall sample size

Q#2. How many households would have to be visited?

• (105,600)(1.18)(5)= 621, 177 will be total H/H to be visited

Q#3. What do you think of this estimated sample size?

• this sample size is too large which require more budget

Q#4. How feasible will it be to conduct this survey?

• since the 6212 H/H will be visited in each stratum it takes long time & it is not feasible to conduct

Q#5. What are the trade-offs in terms of time, money, and quality of survey implementation?

• we can summarize analysis by province also reduce the sample size to be able to conduct the survey at reduced cost and time frame.