What is important parameters of sample size calculation in prevelance study?
population size
effect size from previous study
confidence limit
design effect
effect size from previous study
confidence limit
design effect
Leonid Hanin
In a typical prevalence study, the requisite sample size can be estimated from the following three parameters, provided that certain natural conditions are satisfied:
(1) sensitivity of a test for the condition of interest;
(2) specificity of the test;
(3) a desired precision of the prevalence estimator, as measured, for example, by an upper bound for its standard deviation.
For more details, see
https://www.mdpi.com/2227-7390/8/11/1900
(1) sensitivity of a test for the condition of interest;
(2) specificity of the test;
(3) a desired precision of the prevalence estimator, as measured, for example, by an upper bound for its standard deviation.
For more details, see
https://www.mdpi.com/2227-7390/8/11/1900
Larissa Adna
To calculate the sample size in a prevalence study, it is essential to define the expected prevalence. This can be obtained through previous research in a refined literature review or pilot studies. Another parameter is to determine the confidence level that will be accepted for the study, which can vary between 90%, 95%, and 99%. This depends on the outcome and design of the study. And the last one is the acceptable error. In this case, the error can be absolute in percentage points or relative (%). In general, when we keep the absolute error constant, a larger sample is necessary when the prevalence is 50%. For a constant relative error, lower prevalences require larger samples. It is possible to use websites such as OPENEPI or Stata itself through the "sampsi" command to calculate the sample size according to the study design. Depending on the study's sampling process, for example, cluster sampling, it is also necessary to consider the effect of the design.The design effect adjusts for the increased variability introduced by sampling clusters rather than individuals. Ignoring this adjustment can lead to underestimation of the required sample size and compromise the precision of the study estimates.
Wujoe
Yes simple size calculation in such studies is crucial to ensure that the estimate is accurate and reliable.
Birhanu
1. Expected prevalence / proportion of the problem represented by p
The proprtion can be obtained from previous similar studies or from prestudy survey.
2. Confidence interval
It is the degree of certainity
It is determined by the resercher
It is usually taken as the 95% CI.
3. Degree of precision(d)
It is the standard error that result is due chance
4. Non reponse rate
From the seleceted study population there will be chance of non response to the study.
Therefore 5-10% of contingency should be taken and add up to the calculated sample size.
n= Z2(p(1_p)÷d2