Skip to main content

Survey estimate of vaccine coverage

(go to Outline)

Another way to measure what proportion of children have received a vaccine is by carrying out a survey. This is, a random sample of children eligible for the vaccination is selected from the population of interest and their vaccination status is determined. If the sampling for such a survey is unbiased (that is, it produces a representative sample), then this is the most accurate method of estimating vaccine coverage.

Surveys can also collect other data on vaccinations, such as risk factors for not being vaccinated, reasons for not being vaccinated, and other health information. Measles vaccination status is collected on children in most health and nutrition assessment surveys in emergency-affected populations. Surveys after measles vaccination campaigns are often used to be sure that the campaign has achieved a sufficient coverage to prevent measles outbreaks.

Vaccination status can be collected directly from a vaccination or children health card completed when vaccines are given to the child. However, in some vaccination campaigns in stable populations and in many campaigns in emergencies, health cards are not issued or completed. In this case, mother's recall must be used to determine if the vaccine has been given to the child. According to different studies, mother's recall has been shown to overestimate vaccine coverage, underestimate vaccine coverage, and provide an accurate indicator of vaccination. It may also be less accurate the more vaccines the child has received and the longer time since they were administered. If the evidence of vaccination for a large proportion of apparently vaccinated children comes from mothers' recall, coverage is often calculated twice, once assuming children with only mothers' recall are unvaccinated and again assuming that they are vaccinated. The true coverage probably lies somewhere in between.

The World Health Organization developed a recommended method for vaccine coverage surveys (click here to open this manual).

1

A measles vaccination campaign had recently been completed in Badghis Province, Afghanistan. It targeted children 9 months - 14 years of age. A survey asked mothers of children 9-59 months of age whether their children had received measles vaccine during the time period in which this campaign was carried out. Among the 466 children in this age range, mothers of 277 reported that their child had been vaccinated in the recent campaign.

What is the survey estimate of vaccine coverage?

a)
b)
c)
d)
e)
CorrectIncorrect. The vaccine coverage is the number of children in the survey sample who received the vaccination (277) divided by the total number of eligible children (466) = 59.4%Your answer has been saved.
Check your answer

2

What might account for this poor coverage?

1) The vaccination campaign did not reach many children. This could have been due to the vaccination teams missing entire villages, or the teams missing children within villages in spite of having visited all villages. In fact, the survey results showed that among the villages included in the survey, some had nearly universal coverage and others had almost no coverage. This may indicate that entire villages were missed by the vaccination teams.

2) Unvaccinated children entered the population between the vaccination campaign and the survey assessment. In fact, many residents of Maslak Camp outside of Herat had left the camp and returned home to Badghis Province, so there probably were children who had arrived in Badghis after the campaign. However, only a few thousand households left Maslak Camp, while the population of Badghis included hundreds of thousands of households. So this alone probably could not account for a large part of the low coverage achieved by the vaccination campaign.

Check your answer

There are, of course, some limitations to the survey method of estimating coverage:

  • Because surveys take time to organize and implement, and data analysis and reporting takes time, survey estimates are rarely as timely as administrative estimates based on current program data. Therefore, quick program decisions, such as those needed to redirect an emergency vaccination campaign, cannot be based on survey data.
  • Surveys require additional resources to collect vaccination data. They do not rely on data already collected for other program purposes. For this reason, surveys can be expensive.
  • Surveys may rely on mothers' recall which can be inaccurate. However, in surveys done to evaluate coverage after an emergency vaccination campaign, the vaccination was often recent enough so that mothers' recall may be better. On the other hand, in surveys done during the acute phase of an emergency to measure the proportion of children who are susceptible to measles, vaccination cards may have been lost in the displacement and the vaccination itself may have been many months or years ago.
  • The WHO recommended method may have inherent biases which could result in an under- or overestimate of vaccine coverage. For a more complete discussion of this issue, see pages 79-85 in A Manual: Measuring and Interpreting Malnutrition and Mortality. WFP. 2006. (Click here to open this document.)