I guess assuming PREVALENCE is not a correct idea; while assuming ERROR can be accepted. On the other hand, based on prevalence available in literature in certain area or ethnicity, a prediction can be accepted while sampling, that is what I understood.
However, I have seen people directly assuming the prevalence of particular disease at a certain place while calculating sample size. Any kind soul to clear my doubt, please?
Thanks.
I understand that the decision to select a value of prevalence in the calculation of sample size will depend on the degree of knowledge of the disease.
For diseases with few references, that nearest in space or time to ours, to calculate the survey sample size is selected.
If the disease has been extensively studied, there will be a greater number of references, and possibly there is a systematic review, or at least can select an average prevalence among several upcoming publications in space and time to our study.
Best regards
Paco Rivas
thank! however, I would go for available incidence records in local areas and try to calculate the prevalance rather assuming it.
Also, why not simply use this:
Sample Size = (Z-score)² * StdDev*(1-StdDev) / (margin of error)²
can't we?
I agree with Paco Rivas' suggestions. If one finds many references with a wide range of prevalence values for disease X, one could use
(1) an average prevalence
OR
(2) min, max, average prevalence values
to estimate sample size. For option 2, I would select the highest sample size estimate (if logistically feasible).
One should select the best guesstimate of prevalence value to estimate N.
Seems to me that the Bayesian approach is tailor-made for assuming prevalence.
Thansk Wayway.
John, I am not sure if I have understood you correctly. Just a little science student this side :/ sorry
Hi Ashok, What is your value for SD? and on what assumption?
Dear Dr. Biswas,
Prevalence (number of existing cases) is used for diseases of long duration. For acute diseases we must consider the "bias of prevalence", that is, there is subregister for these diseases, as they cure or kills rapidly. For acute diseases, it must be better to consider the incidence (number of new cases, in a period)
It is scientifically plausible to find the average of prevalences while determining sample size. However, the need to concentrate these prevalences on the geographic location under study should not be under-emphasized. The average of the prevalences published would rather provide a relatively accurate information. For instance, if you're researching on the prevalence and intensity of intestinal schistosomiasis in a province under Cameroon, you'll do well to source for relevant publications addressing the prevalence of the disease in different parts of that province because they share similar ecological factors that influence the disease.
- Badges
- Science topic
- Similar topics
- Medicine
- Disease
- Infectious Diseases