I want to study a prevalence of a resistance mutation in Ascaris parasites that was detected in Brazil. I looked into the literature, and no one has screened for this mutation before in my country (Canada). So I thought to do a pilot study to see if we have this mutation or not in Canada. I don't know how many samples to collect for such a pilot study. All papers I read so far do not use a sample power equation to determine the sample size, they are using a random number of samples without a specific statistical method.
Basically, I'm going to collect only Ascaris-positive samples from clinics (with consent) and screen for the mutation with nPCR
- How do I determine the sample size for such a study?
- What is the name of the statistical method to use? A reference would be great to understand the point of view
THX
As you wish to determine prevalence of a condition with a specified precision .You need to have a guess at the likely prevalence.
Then use the method as descibed in the attached.
I thought if there was no previous prevalence then I would assume that it's 50% and I would keep my confidence level at 95% with a 5% margin of error.
I found this web tool that was made to calculate sample size for pilot studies (Calculation of sample size in pilot studies (crutzen.net)) by Crutzen et al. 2015, and I used the numbers above in a fractional format to use it and it gave me the sample size to be 4.3 ~5 samples, is this correct?
- Is it not too low?
- Do people get low numbers like this but they collect 100 samples or more or as much as they can just to satisfy a journal to get accepted even though the number required was as low as 5 samples?
I guess 5 is the sample size to reveal at least one mutation if the prevalence of mutation is 50%: the formula you used is not the right one
https://food.ec.europa.eu/system/files/2018-11/ad_cm_asf_btsf-asf_20171025_pres-04.pdf
- Badges
- Science topic
- Similar topics
- Papers