Wanna read a Forest Plot ? (Meta Analysis contd...)Applied Econometrics

■ What is a Forest Plot?

Simply, A forest plot is a graph that compares information many individual, clinical or scientific studies studying the same thing.

If you guys are well acquainted with the prior knowledge of Meta Analysis (mine earlier blogs) then lets state:

Say, there are studies such as Ram et al., Sita et al., Siva et al., Abdul et al., Hanuman etal., etc. So in the plot we will also have a Summary measure of all these along with the odds ratio (odds ratio will be discussed later in the Qualitative analysis).

Now you need to break the ice. Here you go,

The Realtive Risk Ratio say, 95% C.I visually displays the study results. The right column of the plot displays the same results expressed numerically.

The horizontal lines in the figure denotes the length of the Confidence Interval. The longer the line, the wider the C.I and as such the less reliable the results are.

If all the horizontal lines crossed the vertical line, it shows that all the studies were in agreement and if the horizontal line doesn't cross the vertical it's an indication that were statistically significant differences betweem the studies.

The boxes on the intervals represent the effect estimates from the single studies. Usually, the bigger the box, the larger the sample size as well as more meaningful story.

Now, an outline of a diamond at the base of the forest plot represents a weighted average for all the studies but may also be in Odds Ratio. The lateral tips of a diamond represnt the Confidence Interval.

If the diamond touches the vertical line then the overall meta analysis results is not statistically significant. If the diamond is left of the (dotted) line, then there is a lapse of two episodes of interest in the treatment group (treatment group discussed in previous blogs - in RCT).

Similarly, if the diamond is to the right, there are more episodes of the outcome in the treatment group.

** Summary**

a) Names to the left: 1st author of the rinary studies

b) Solid Sqaures: RR or OR of the individual studies.

c) Solid Sqaure Size: Weight of the Study

d) Horizontal lines: 95% of C.I

e) Vertical lines: Line of no effect

f) Diamond: Overall treatment effect

g) Tips of Diamond: 95% of C.I

Thank you

P.S Comments and Suggestions are highly lauded.

Aditya Raz Pokhrel

Assistant Director,NRB

MBA, MA Economics, MPA(rng)