Hi Heather,

These are very sensible questions indeed; there is some information about backups scattered in different places (one is here) but it's a good idea to summarise:
1) Yes, the EPPI-Reviewer 4 database is backed up in a multi layered way. The real world effect is that it is possible for us to restore it to a previous state, and retrieve/update information based on the restored copy.
Things to know: in principle, we could go back in time to any date (with precision of seconds) with no restrictions; however, we can't store all this data indefinitely, so (under normal conditions) we can usually go back to any moment of the past week, and then only selected intervals for another three weeks, an even more restricted selection for the months before and so on.
This procedure is fully tested and we know that it works and that is time consuming. A minimum of three hours are needed just to restore data, after that one will need to fish the correct data out of the restored “snapshot” and use it to change what is on the live database. Depending on the needs, this may require days of work. Therefore, we will charge users for this service if/when requested. Of course, if the restore is needed because of some fault of ours (such as a bug) we will never charge for it!

2) No, at the moment EPPI-Reviewer 4 does not allow to manually save and restore a review state. This is in our plans, but it is very difficult to design (because the restore side of it clashes with other features that we are still developing). It is a feature that is quite high in our priorities, but given the challenges it presents we can’t foresee a release date. There are some “persistence” features in EPPI-Reviewer, though; these were included also to compensate for the lack of central, review-based backup and restore features. For example, searches and collaboration comparisons are “static” in the sense that they don’t change with time, so it is possible to create searches to record a particular state (like “which items were coded with code X before the second screening round”) and keep a record of double-coding state through collaboration comparisons.

I hope this clarifies and will be happy to explain more if needed.
Sergio