Sorry Sam, I'm going from a conversation a few months ago about data gathering, for an NHS project (secret squirrel and all that). I'm only working on feedback I received to my questioning (read interrogation. Oooops!) at that time.
I know for mass data gathering in commercial environments, systems are usually set up to passively gather data, based on day-to-day business activity.
I wish I had that level of confidence in the data. Data gathering is usually poor (NHS or not) - especially where much of it is governed by accurate coding.
Given the data generates a source of income, and the turnover of most GP surgeries as independent businesses requires external audit, I suspect it's a lot better than people keep giving it credit for.