Error with confidence interval calculations

[theanita1]

Hi everyone

Am trying to calculate sir_confidence_intervals and it's throwing the error
Error in summarise(): ℹ In argument: ci_min = sir_confidence_interval(...). ℹ In group 1: state = ACT. Caused by error in stats::binom.test(): ! 'n' must be a positive integer >= 'x'

Whilst I do only have small numbers in my results, I haven't come across this issue with other bug/drug combos with similarly small numbers

Here's my code for what I'm trying to calculate:
sau%>%
group_by(state)%>%
summarise(count = n(),
no. tested = n_sir(ERY),
no. R = count_R(ERY),
ERY = proportion_R(ERY, as_percent = T, minimum = 0),
ci_min = sir_confidence_interval(ERY, ab_result = "R", side = "min", as_percent = T, minimum = 1),
ci_max = sir_confidence_interval(ERY, ab_result = "R", side = "max", as_percent = T, minimum = 1))

and here's my aggregated data
/state/tested/resistant

• ACT/20/2
• NSW/173/28
• NT/29/6
• Qld/88/14
• SA/57/7
• Tas/15/0
• Vic/152/17
• WA/73/6

[msberends]

Hm, I wonder if this isn’t a bug.

So if I understand well, you have 20 isolates available where state = "ACT" of which 2 are resistant? Since susceptibility calculations require >= 30 isolates at default, it could be that the confidence interval calculation fails if this requirement is not met. I’ll look into this.

[msberends]

Ah but you have set minimum = 1… Weird. I’ll get back to it.

[theanita1]

Hi Matthjis

I don't know what changed, but the error has stopped appearing when running the above code, so wanted to give that feedback

Cheers