Your 3 examples are great, but they often end up being huge amounts of consulting -- you need to integrate with the quirks of how the specific practice works, and whatever hardware and existing systems they have already. A lot of this is related to insurance as well.

I think huge conglomerates like Kaiser, which provide insurance and care through a single network, could go a long way to improving healthcare IT. At some point, they become big enough to actually become a device manufacturer, software developer, etc., or at least to allow a single product to take over their entire practice. It would almost make sense to have an IT-focused medical hedge fund which develops a great IT system, then goes around and buys practices in strategic markets to convert to the IT system; sort of like the medical borg.

Mobile interfaces to existing EMRs would be interesting -- either within a hospital, or for consults outside. Being able to quickly check and see where someone is in the process would be great. The problem is this either requires huge amounts of customization to work with existing workflow, or adjusting workflow. Radiology with PACS is really the only case where workflow has been totally changed by IT, and that was mainly to allow perhaps the laziest specialists of all to work from home or the beach.