GPs invented medical records
We have been keeping them, quite safely and usefully, for some decades.
Some of us even start talking about things like medical informatics, and some of the reasons why a pile of records from several databases made by several orgs for several purposes turns out to be just a pile when you amalgamate them.
In Exeter in the '80s the experiment of building a record for use by specialties and GPs was tried. It didn't work. The current effort does not include any solutions to the problems that killed that.
Medical records are not the only thing that we should be looking at, they are relatively easy, accept that there are several sorts of Practice software, note that the laboratory is a good place to hang on to lab results, that the PACS of the local imaging department can usefully hold the pictures and reports, do some caching and devise a framework perhaps with a PKI to allow my machine to ask the machine of a GP in Aberdeen (it won't work BTW since that is another country, come back South to Carlisle for examples) questions such as "IsADiabetic(ID)" or "ListAllergiesAndAvoids(ID)" at need, and send a list of those to the patient as well in a quarterly statement.
No, how about a bit of informatics, of computing, of weak AI, or at least Hinting Engines and Opinionated Systems (one of those would model many of my colleagues quite adequately, and they are easier to build than the Eschaton I think)
Or if you want a working system, take copies of the GPL'd US VA system - called VistA for the last couple of decades, install that and then work on from there...
Because of all the long-term infrastructure systems that should be built of Open Source, healthcare IT is hard to find a better example.
What is being done at the moment is driven by money and power, not by cleverness in IT or medicine. Which is a pity.