The Tories are clueless about this.
Let's be frank: any IT project in the NHS is likely to fail: too many organisations that work in different ways; too many people in an organisation that want different things. It's a mine field.
However, what the Tories are suggesting is not necessarily Google / MS solutions. Somehow, Cameron went out to see Google and got the impression that distributed patient records was better than the centralised approach of NPfIT. Moreover, he wants to create a 'market' within the NHS so that the 'best' solution wins.
I've seen this in the Fire Service, where you could make a fair living by re-implementing systems slightly differently for individual brigades. No centralisation meant that communications equipment in each brigade was different, the software was different and that meant lots of money for contractors and ISVs. This is despite the GD-92 communications specification, which aimed at inter-operability and meant, in principle, you could re-use the same system in different brigades; somehow, it never ended up like that, however.
What the Tories are planning is even worse. Imagine a company where if you went from one office to another you had to re-learn how to use the software that it is now essential you use to do your job. It's not just essential: lives are at risk. Well, the Tories want to create the situation. And the call it 'the market'.
Any IT project in the NHS is going to be hard. But the problem with government is that they will only deal with big contractors (EDS, eg). Smaller, more agile teams would probably solve these problems better. Software on this scale should be a living, ever-changing thing, with very small changes on a regular basis. What the big boys will produce is large, functional systems that will do what it's supposed to on paper. But not in practice.