I'm not sure what other's practices are, but this is what I have found to be useful and hopefully scientific:
When presented with a patient who has a possible antibiotic allergy, eg penicillin and requries a similar antibiotic eg a cephalosporin some people advocate a IV test dose before a full dose is given.
This would minimise the effects of an anaphylactiod reaction, which is dose-related, but not a true anaphylactic reaction which can be triggered by a minimal amount of antigen.
My practice has been to give a tiny bleb of antibiotic as an intradermal injection, essentially making it a skin test. If there is no weal or flare 10mm or larger after 20 minutes, then I feel confident giving the full dose IV.
The advantage of this is that the antigen is not given as a systemic dose and thereby triggering a systemic reaction. It remains localised causing a local reaction only, yet still giving the required information.
I realise that even very dilute skin tests can cause anaphylaxis, but this technique is still safer than giving an IV "test dose".