As per our practice, we remove an equivalent diluent volume. I strongly recommend the calculation based on the final volume as one of the patient rights is right dose.
As you mentioned, we can’t always guarantee that this small fraction is not significant for all preparations considering, for instance; electrolytes and narrow therapeutic index drugs. I’m not sure if there’s a documented evidence for a contrary opinion.
Also it’s advised to make sure that the final volume is within the concentration stability window as per known references and manufacturer recommendations. Attached is an interesting paper showing supporting examples.