Version 4.4.4.1 includes new "On Request" integration features with Cochrane systems (Archie), these are part of our joint project with the Cochrane Collaboration, see here for some background information. As far as public features go, this release includes new functions to configure reviews and to copy existing Codesets across reviews. Some small changes and improvements are also included, as usual.
Review Setup Wizard.
When opening up a new review, if the review contains no Codesets, EPPI-Reviewer will propose to open the Review Setup Wizard. The same wizard can be opened at any time from a "Configure Review/Codesets" button in the "Codes" toolbar, visible in the main window ("Codes" tab). The wizard allows to proceed in two ways:
1. Users may import Codesets in bulk, using a couple of Template Reviews. In this release, only two templates are available: "Standard" and "Minimal". We can add new templates at any time, so we would welcome suggestions.
2. Users may import Codesets on an individual basis, chosen either from the list of public Codesets or from the reviews where the current user has Administrative rights. Once again, if you have produced one or more Codeset(s) and think they may be useful to other EPPI-Reviewer users, please let us know, we would be happy to include them in the list of publicly available Codesets.
Archie Integration Features
These features are currently in Beta, available to all users that are using EPPI-Reviewer, but on request only: please contact us if are working on a Cochrane review and wish to gain eraly access to these features. Assuming this first phase will not highlight major issues, these features will become public on the next EPPI-Reviewer release.
The highlights are: Cochrane users (people who have an Account on Archie) will be able to use EPPI-Reviewer to conduct their Cochrane-related work without charge (there will be no need to pay for subscription renewals), this is done by allowing users to authenticate in Archie from within EPPI-Reviewer (so to certify their status as Cochrane authors). Users will also be able to work on registered Archie reviews and mark them as Checked-out (for edit) in Archie. Finally, users will be able to carry out some exploratory work on reviews that are expected to eventually become registered Cochrane reviews.
Minor Improvements and Bug fixes.
Direct Search In Pubmed: NCBI is shutting down the old SOAP API services which EPPI-Reviewer was using to natively search on Pubmed. Therefore, a brand new "behind the scenes" system to search and import results from Pubmed is now in place. From the user perspective, it might be a little faster, but no significant change should be visible: everything under the hood is new, but in terms of functionality, no change should be visible.
Codesets Description: Codesets now have a "Description" field. This is visible at the bottom of Codeset trees in the same box that shows the description for Single Codes. The description can be edited when a new set is created and by right-click on the Codeset root and "Properties". Before this change only single Codes could be associated with a description.
In document details, ticking a code Checkbox or clicking the “Info” button (which would tick the checkbox if it is currently unticked) now “selects” the code, this makes the code description appear in the box below the codesets tree. Before this change, the currently active code in the tree would not change, so the description shown could have nothing to do with the code that had just been applied.
In document details, the "Infobox" (the window that is used to add information about an Item-code association) can now be moved freely, before it was locked at the centre of the screen. The window also remembers where it was last shown, so re-opening it for a new item/code will make it appear in the same place. Following some user feedback we think this may help filling in the infobox content while keeping other contents of the screen visible.
In the "Find similar documents" window, it is now possible to select multiple terms and delete them in bulk. Before this change, irrelevant/unwanted terms had to be deleted on a one-by-one basis.
Also in "Find similar documents": there is a new system to avoid search failures. When searching against a large number of references and many terms, the search may take a long time, causing a timeout error and returning zero hits. There was already a system in place which was dropping the least relevant terms from the search, so to mitigate the problem.
Starting with this release, in case the above fails (the search has an initial 2 minutes to run), the search will now automatically drop even more terms (picking the least relevant ones) and get executed again, this time with 4 minutes of timeout period.
As a result, (weighted) term searches now can take up to 6 minutes, but should almost always succeed. Note that term truncation starts kicking-in when the list of terms is above 120 terms or more. Truncation is performed on the basis of the terms length: how many terms will be included in the search depends on their overall length.
In the Edit / Create Outcome window: the drop down menus for Outcome, Intervention, and Comparison now list the available codes in Alphabetical order. Before this change (applied on July 24), the order could appear to be completely arbitrary, this happened when the codes listed appeared in different branches of the Code-Tree.
Bugfix: importing date fields. In some cases, and with some particular combinations of source and file-format, EPPI-Reviewer was mishandling the various date fields and was prone to miss the year information in particular. This bug has been fixed, and most new imports should benefit as a result.
Bugfix (added on July 24): in the Edit / Create Outcome window, the drop down menus for Outcome, Intervention, and Comparison could list codes that were deleted as part of larger Code-Tree branches. This problem is now solved.