I don't want to make my identity known to the writers of the papers
(Yes, also in my field reviews are supposed to be blinded on one side. But often very good guesses to who was the reviewer are possible. Sometimes it is obvious, and oftentimes I believe I could track down the reviewers at least to their groups because of the specific use of certain terms. And, yes, you could probably track me down because I also have typical questions. Personally, I'd prefer receiving review swith the reviewers' names (it is useless to thank reviewers A and B, as those two review all my papers, and everyone knows it - but I'd like to acknowledge reviewers by name) and to write reviews under my name as well.)
I see several possibilities here.
As was suggested already, name the issue, not the paper. You can also guide the authors to search terms that will lead to relevant papers.
There may be valid exceptions to this:
- Sometimes one wouldn't know from title and abstract that the paper is relevant, e.g. when some methodological point was presented in a paper about an application.
- Sometimes it is impossible to dig out relevant papers between other papers that use the terms differently or some combination of search terms happens also in irrelevant context*
Your question sounds as if you are well-known for the topic which you found missing in the paper.
- If you are The Big Guy for this topic, pointing to your publication does not compromise your anonymity - any other reviewer who is aware of the issue would have done it, too.
- If you are not The Big Guy, but maybe the only one in that field looking into this topic, odds are still you were asked to do the review because of this expertise. IMHO, the quality of the review matters much more than semi-existent (see above) anonymity.
In that case, I'd write the review so the authors can understand and address the issue easily. If you really think that this compromises your anonymity, you may write to the editor that you think your review is not anonymous, because ..., and possibly that he may decide to give your email to the authors and that they could contact you in case of further questions (IMHO it is much less work to answer a few questions that to have to review an additional time).
* e.g. "soft classification" in remote sensing is used ina certain way, which I took over into chemometrics. However, one very important classification method in chemometrics is SIMCA, the "S" for soft. It could be used as as soft classifier in the remote sensing meaning. But is usually isn't. So I got tons of hits with SIMCA. No hits excluding SIMCA, and after looking into a certain number of them and never finding it used in this "soft" way, I gave up and had to say that I didn't find any such application. If anyone knows such a paper, please tell me the proper citation. I don't mind if you're the author.