One change that recently took place, and one is planned for next week. Both changes are related to the information found in this blog post about Google+ account integration, and I'd recommend reading that blog post first:
The change that recently took place has to do with the format of the file names used in the Insight archive. Previously, the file names would start with the YouTube channel name. Starting recently, the names of the files in the archive all start with a 22 character user id, which corresponds to the yt:userId value mentioned in that blog post. This ensures that a unique identifier will always be used for the file's prefix even when the report is for a channel that is linked to a Google+ profile.
The change that's scheduled for next Wednesday, May 16 is the addition of a new column within each CSV file. The column is called "Display name", and it appears in reports immediately after the column called "Channel". "Channel" will always contain the same unique user id value that is used for the prefix of the file name (corresponding to yt:userId in the API's responses) and "Display name" will always contain the publicly-visible name of the channel (corresponding to the display/yt:display value in the API's responses).
If you have existing code that parses the CSV reports that makes assumptions about the number of columns or the order in which the columns appear, please make sure that you're aware of this upcoming change and adjust your code accordingly.