meta-analysis of diagnostic accuracy studies

[Shripada Rao]

Hi, how to pool the results for meta-analysis of a diagnostic accuracy test if there are only three studies? The enclosed excel sheet contains data on true positive, false positive, false negative and true negative from three included studies. But the 'midas' and metandi commands allow pooling only if there are at least four studies.

[Jared Greathouse]

Show us your data using dataex please.

[Shripada Rao]

author	year	tp	fp	fn	tn
author1	2012	13	22	14	58
author 2	2019	101	230	25	274
author 3	2017	90	16	22	116

[Shripada Rao]

Thank you Jared, here is the data, which I had enclosed in the excel sheet.

[Shripada Rao]

input str8 author int(year tp fp) byte fn int tn
"author1" 2012 13 22 14 58
"author 2" 2019 101 230 25 274
"author 3" 2017 90 16 22 116
end

[Shripada Rao]

I have tried to use dataex. Hope it looks ok

[Tiago Pereira]

You are trying to fit complex model that do not estimate well with few studies. I would suggest to perform a meta-analysis using the diagnostic odds ratio - via metan. You have also use -metaprop- to summarize proportions. In any case, avoid fitting complex model with little data. Results can be misleading.

[Shripada Rao]

ok, thank you