As part of a Master's degree I conducted a diagnostic test accuracy meta-analysis on the ability of the anion gap to screen for hyperlactataemia. For the summary ROC curve, I used the Moses-Littenberg model. Now, I would like a more statistically accurate summary ROC curve i.e. a hierarchichal or bivariate model. Unfortunately I do not have the skills to implement them myself and I have been unable to find someone with such expertise necessary at my University.
Does anyone have any ideas where I can find a statistician who might be abie to help me with this task?
Hi Stella.
I had some difficulty issues with statistic on the past.
If you really know your question and how you wanna to go for it, I would recommend to you to look over your maths or statistics department.
This can be the easiest way.
They are more difficult models to fit but not impossible. If you are used to statistical modelling and STATA there is quite a good guide written by the Cochrane Collaboration that I can send on to you. Not many people have this experience and it might not be possible to find anyone at your university, not even in the maths or stats department (though definitely worth checking). I have done one and can send on the guide if you email me: [email protected]
Dear Stella Glasmacher
I have made a HSROC plot in STATA, when performing meta-analysis on diagnostic test. If you have access to STATA the following article gives a a good introduction to meta-analysis on diagnostic test and will guide you through making a HSROC plot
1. Harbord RM, Whiting P. metandi: Meta-analysis of diagnostic accuracy using hierarchical logistic regression. Stata Journal. 2009;9(2):211-29.
Kind Regards
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Meta-Analysis