'Meta-analyses and Forest plots using a microsoft excel spreadsheet: step-by-step guide focusing on descriptive data analysis' by Jeruza L Neyeloff et al, available at Article Meta-analyses and Forest plots using a Microsoft Excel sprea...
You are really good informative, please if you guide me a little bit more. Often we found only one value for specificity and sensitivity. How can we make its lower and upper limit through confidence interval?
I agree that SPSS is not the best software for doing meta-analysis and making forest plots. Nevertheless, one could do something like the following:
DATA LIST LIST / caselbl (a35) oddsrat (f8.3) lower (f8.3) upper (f8.3) .
BEGIN DATA.
"Trial 1" 0.806 0.373 1.739
"Trial 2" 0.222 0.054 0.914
"Trial 3" 1.412 0.396 5.036
"Trial 4" 0.929 0.304 2.841
"Trial 5" 0.913 0.394 2.111
"Trial 6" 0.385 0.168 0.880
"Pooled OR (fixed)" 0.682 0.460 1.009
"Pooled OR (random)" 0.680 0.432 1.072
END DATA.
* Create HILO graph .
GRAPH
/HILO(SIMPLE)=VALUE( upper lower oddsrat ) BY caselbl .
* Double-click the graph in the output window to open the chart
editor; then swap axes, and make other changes to make the
graph more presentable.
This is an old example I put together in 2001. Since then, SPSS added the GGRAPH command. If you look into using GGRAPH to make the HILO plot, you can probably get all the tweaks you want via syntax rather than manual editing of the graph.