Sorry, Danny but I don't think that you are looking at things correctly.
In fact, if you work in the way you describe, you are working very inefficiently.
The old 'ohhh you can't touch my model because of liability risk' is a big myth. Propagated by those who do not have any clue about the legal or commercial realities. The simple way to get around the perceived legal barrier is to simply get the structural engineers to check the proposed changes at each design gate.
The whole issue of shared responsibility is not new to the industry. Traditionally, the lead consultant will be held responsible for all mis-coordination, errors or omissions made by the rest of the team way before BIM or even CAD showed up. It is just exacerbated by BIM as more is exposed for all to see. And these days, due the amount of follow on design provided by the subcontractors, there is now simply much more elements published up front.
MEP modellers are one the most inefficient disciplines in the industry at the moment. Their work is typically design intent only and has to be completely remodelled by the MEP contractor. You think that there are liability issues between consultants? Hah, the contractor usually has to adopt the consultants model and fix any problems etc thereby taking ownership of the design. Do you think that he will ring up the consultant and get him to amend things? He will just put the changes in on a submital and get them to review. And, the consultant on the other hand will just give him a 'no comment, proceed' reply knowing that the contractor needs sort things out regardless. This is accepted practice and there have been established contractual mechanisms for this for ages. Same as a lot of the other disciplines, but more acute for MEP.
And, the complete disconnect between the modeler and the engineer's calculations is another problem and incredibly wasteful. The public health or duct modeller models... and if someone wants to know what the pressure drop is at a particular location, the engineers would have to do a manual calc. No intelligent link between the physical and analytical models.The structural guys at least have the option to connect to an analytical tool via ISM. The modeler might even know enough to construct and maintain the analytical stick frame in aecosim to maintain continiuty between the p/a models.
Same goes for PH design. When the modeler models the system, he would start with the number of end fixtures like hand wash basins or WC's and the system would know what water and pressure demands are in the system at all times. As he strings the system together and eventually gets back to the water storage tank, he will know what what capacity is required or whether the pumps strong are enough etc... not wait days for the PH engineer to re-jig his excel sheet or worse re-do his paper calcs.
Same would apply to the structural calcs. The SE sees that the opening has changed. He has a prepared structural check that he just applies. This is called a Design Check Form in SCIA. He runs through the changes and accepts or rejects the changes like a MS Word user would with Track Changes, or what your BIM modelers have to do with clash detection. Hey, here is the kicker. If you can't make any changes to the other guys model, you end up having to do a sketch or some bumbling note in Navis or Navigator and send it over to be mis-interpreted or even worse go in for an expensive 'back-seat driving' session where multiple guys travel hours to sit behind the modeler's screen to ensure that he gets it right. Better to just make the changes in the model and flag it up for review.
I have seen MEP modellers chase their tails because they have all their models in separate Refs. They can never do anything clash free first time and have to wait for the export to Navis and the clash detection to come back, and go into another merry-go-round as they try to react to or second guess what the other MEP modellers will do in their separate models. Some projects shoot themselves in the foot even further by not allowing other disciplines to ref each others models unless they are shared, mutually preventing their modellers seeing what their counterparts are doing until the next share !! They are all paid by the hour so I can understand why no one has raised this to Bentley... no? Bentleys, the do nothing path is not always best.
As a result, some get hired at a higher rate because they are 'multi-discipline' modellers supposedly knowing enough about PH, Mech, fire and electrical to hop between models using the Activate tool. To use your logic, this should never happen due to the liability issues, but I can assure you it does.
Even then they are sooo slooow. It would be far better to be able to have multiple Ref attachment active.
Mstn can already do unification across multiple Ref files. Subtractions should be possible.
Auto-subtractions are already possible in Speedikon, Revit and PDMS. No moaning about liability there.
As mentioned above, it would be good to have the option. Its all about control. If it is not desirable... switch it off.
I would argue that since it is already common practice to ignore the smaller openings, then it would make sense to auto-subtract small pipes. Provide the functionality for auto subtraction by size. We could also include a parameter that allows the structural engineer to 'check' when he reviews the penetrations.
Workflow: Yes, it would be good consider this in more detail.