Why the lower 95% confidence limit of PPV is 0.00 for calculating by the "epiR "package in a diagnosis test ?

[Dennis007]

Dear all,
Now we conducted a diagnosis test study for estimating the diagnostic accuracy a new method.
We calculated the sensitivity, specificity, Positive predictive value (PPV) and NPV by the "epiR "package for R.
As shown figure1, the lower 95% confidence limit of PPV is 0.00.
[0.07 (95%CI 0.00, 0.34)]

What's the meaning of this?
I guess the number was always display 2 decimal place？
Or the number is represented zero？

Thanks!

[mstevenson]

Hello Dennis

You're correct, it's a rounding issue. Two options. Firstly, in the latest version of epiR (2.0.46) there's an argument 'digits' which allows you to change the number of digits in the printed output for epi.tests. So use something like:

epi.tests(c(1,13,1,28), method = "exact", digits = 3, conf.level = 0.95)

Second option is to ask for summary on a named epi.tests object. For example:

rval <- epi.tests(c(1,13,1,28), method = "exact", digits = 3, conf.level = 0.95)
summary(rval)

Hope that helps.

Mark Stevenson

[Dennis007]

Dear Mark Stevenson
Thank you for your good advice.
It's amazing for solving the problem.

Best ragrds
Dennis