Using R for epidemiology stats

[Keno_Mentor]

So I’m doing a meta-analysis for the first time and I’m trying to get through some new (to me) statistical techniques using R. To test whether I’m used the code correctly, I’ve been using a few of the studies absolute data and calculating the stats and checking if I get the same results as reported in the studies. However, I keep getting slightly different results and I’m wondering why.

Is this normal? Using the same test but different software and getting slightly different results?

For Example:

When I use R, I get a RR of 0.51 (0.36-0.72) using the code below:

library(epiR)

RiskFactor ← matrix(c(28, 103, 253, 351), ncol=2, byrow = TRUE)
colnames(RiskFactor) ← c(‘Infection’, ‘NoInfection’)
rownames(RiskFactor) ← c(‘LiverDis’, ‘NoLiverDisease’)
RiskFactor ← as.table(RiskFactor)

RiskFactor

epi.2by2(RiskFactor, method=“cohort.count”)

Could I ask a few of you experts to use your stats tool of choice to compare?

[Jochen]

You exchanged the rows and columns in RiskFactor.

With epi.2by2(t(RiskFactor)) you get the result reported in the table.

[Keno_Mentor]

Silly mistake…thanks for pointing that out

[Keno_Mentor]

Could I try another? I’ve looked at this example many times and still get a different p-value:

[iiijohan]

To start with, the percentages are not correct.

> 2384/5228
[1] 0.456
>
> 777/1654
[1] 0.4698

Maybe there are non-reported missing values in the “Preoperative obstructive jaundice” variable?