Add evalue for meta-analyses?

[malcolmbarrett]

We may be interested in adding the E-value to the section on bias. The E-value is a measure of how strong unmeasured confounding would have to be to explain away results. It has an implementation for meta-analyses, as well as an R package: https://github.com/mayamathur/evalue/tree/master/EValue

cc: @mayamathur

[mayamathur]

You are welcome to take source code from EValue with attribution if you'd like (or just call its functions directly).