One of your nightmare scenarios as you near the end of your PhD is discovering that someone else has already done pretty much the same thing. Those preparing for viva will be familiar with waking during the night with the eerie sound of your examiner saying "so, why didn't you look at Blogg's study of 2008 which reported similar findings ... I'm struggling to see what you're adding to here ?"
As well as checking the literature on a regular basis (see the post on literature reviews) ... it is a good idea to check for theses on the the same or similar topics. There is now a great on-line service to help with this process. It is run by the British Library and you can access digitised version of doctoral theses for free ... there's even a helpful search engine to narrow your search a little. It can be found at ...
Even if you are in the early stages of your PhD, looking at the finished article from time to time is a good discipline. I always found it reassuring to look at things which had passed ... it helps answer obvious questions such as length, format and depth. However, bear in mind that PhD theses vary in quality. Don't just read them blindly ... try applying the PhD assessment criteria (see earlier post) to check whether you think that the author has made a clear and unambiguous case in relation to the criteria.
For some other helpful thoughts check out the 2nd Edition of Research Methods.