The issues were not technical
Having worked on two parts of npfit over several years, I could see not only that it was going badly, but why.
Basically the core "spine" and database part was terribly badly designed simply because it was based in a 'quick and dirty' demo setup that wasn't designed for scaling to production, but 'the authority' insisted in using that.
They also insisted on regular version releases so close together that there were two separate development streams leapfrogging each other. An utter and complete waste of resources.
The implimentation of the system at endpoints was mostly a matter of integrating it with one of a handful of existing systems.
The sensible approach would have been for a given contractor to integrate all instances of one system no matter what the location.
Instead it was done regionally, so each contractor had to cope with all the systems as well as accommodating varied (and contradictory) requirements for each trust - and sometimes for each hospital!
It could have been excellent. It should have worked, but it was crippled by inappropriate administrative restrictions.
Since then, technology has moved on, but political incompetence is still the same problem