The Lancet correspondence (should you wish to review it) is here:
https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(24)01169-3/fulltext
The relevant paragraph about estimating the death - direct and indirect - caused by the conflict to that point is as follows.
That citation is to this paper about the
Global Burden of Armed Violence.
Specifically, it appears to be referencing table 2.3 on page 40.
View attachment 351325
You will see that the numbers there for the ratio track with what they are claiming.
So it is a
single source (although that table is sourced to multiple other sources of data) that they are using to justify the range. From that range they did, indeed, pick a reasonably conservative number.
But this does clarify that all that letter to the Lancet is, is an educated guess about what the number might be.