I am doing a cross-sectional study to determine the incidence of keratoconus in a risk factor group (the case group-the exposed). There will be another one as control. How could I calculate the sample size considering that there are no previous studies?
yes I didn't write the question in a proper way >>> I'm sorry >>>> I have edited it
the problem is that there is no studies investigating the prevalence nor incidence of keratoconus in this "risk factor group" >>>it 's just some case reports suggesting this relationship so what I am willing to do is to determine whether there is any relation or not?
first I want to thank you a lot for your caring :)
we are trying to find if ****** (or ******* , not decided yet) is a risk factor for keratoconus or not. Non of them has been investigated before
We will make two groups; one as cases ( ******** or ********) and another one as control.
-Is not this a cross-sectional study ?
-Is it necessary to test a pilot ? if yes what is the least number?
I agree with Miguel. What you are describing is a case-control study. If there is no previous literature of odds ratio of this risk factor (let's call it X) and keratoconus to estimate sample size, I would ask other experts in the field regarding the "ingredients" you need to estimate sample size. Second option is to estimate using the most conservative effect sizes or your best-educated guesstimate (e.g.,p1= proportion of people with keratoconus (cases) with risk factor X and p2= proportion of people without keratoconus (controls) with risk factor X). For details on how to estimate sample size comparing proportions, see Hulley SB et al., (4th ed) designing clinical research, Sample size chapters 5 & 6).
first, make it clear your study design then you can use EPI info software to calculate sample size for different study design ( Cross-sectional, case-control or cohort )
It might help you:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3775042/?report=printable
What if if you use P=0.5 if there is no previous study in the area. This will give you the maximum sample size so that it will minimize sampling error.
Hi I am doing a cross sectional study on the diagnostic value of a biomarker In cervical cancer.can I use the dobsons formula (z)2*p(1-p)/d2
And should p be the prevalence of disease or the sensitivity of the biomarker as per literature?
This formula (z)2*sn(1-sn)/d2*P where an= sensitivity as per literature and P = prevalence of disease is making my sample size too large which can't be reached due to many reasons primarily the time and finance
Can't I use the first formula.is it valid for a biomarker study?
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size