Version 4.4.1.0 is a mix of new features and bug fixes. Highlights include: Risk of Bias reports and support for DOI and Keywords fields.

New Features:

Risk of Bias Reports. Question reports offer a new option in the "Run Report" window, allowing to return a Risk of Bias report (instead of the standard one). These reports return Risk of Bias figures/tables, based on the following assumptions: for each (question/)parent code included in the report, its child codes will need to follow the specific pattern for Risk of Bias.

• The first (top-most) code will need to signify (no/low risk),
• the second code will represent "high risk of bias",
• all following codes are considered as multiple variants of the "unclear risk of bias" category.

Important: Risk of Bias reports will not work correctly if the above convention is not followed correctly, this concerns the order of codes as they appear in the code-tree/codeset, not the report desing.

The Risk of Bias report includes two tables: in the first, each study included in the report is listed; the table shows, for each category, whether the study was classified as Low, High or Unclear Risk of Bias. The second table shows the percentage of studies that fell in each category.
We are in the process of analysing different ways to make the Risk-of-Bias code-set design and report design more process driven, so to help users by providing a streamlined procedure, but we couldn’t find a unique "always appropriate" approach so far; you may expect the Risk of Bias reporting features to evolve as we’ll collect user feedback on our first implementation.

Please see the section Creating a Risk of Bias report in the User manual for full details on how to use this new feature. You can find the User manual under Manuals in menu at the top of the screen.

Support for new fields: DOI and Keywords are now imported, stored, shown and exported (applies to new imports only).
EPPI-Reviewer native import filters (RIS and others) can now recognise the DOI and Keyword fields. Data is imported and saved on the database (when present). The "Item details" screen shows and allows to edit the two fields (also applies to "Linked Item" details, read-only). The direct import from PubMed now imports DOI and (MESH) keywords. The "Export to RIS" feature inside the main EPPI-Reviewer application also exports these two new fields.

NOTE: the "RIS export Utility" is able to interpret hundreds of tagged file formats and their specific, idiosyncratic and undocumented dialects (the small variations that different websites/databases use for their own exported files). As such, it is impractical and probably impossible to update all the filters to include the (currently ignored) DOI fields, we will try to update the most used filters in the near future, but it is not (and will never be) guaranteed that the DOI field will be handled correctly by the RIS export utility.

New Import Filter: EPPI-Reviewer includes a handful native filters that can be used to directly import different types of files, without having to go through the RIS export utility. These filters included one called "ERIC CSA" that has been superseded and is now removed. A new filter replaced it, called "Web of Science" this filter natively imports ".ciw" files, that are generated by Web of Science and other databases (these files are sometimes called "EndNote Desktop": EndNote and Reference Manager natively understand them). A number of users contacted us asking how to import such files; this previously required to use the RIS export utility and can now be done directly in EPPI-Reviewer.

New Importing option: Safe Upload.
New Features: in "Add/Manage Sources" window, "Import New" tab, there is a new "Safe Upload (slow)" checkbox (not selected by default). When activated, this checkbox allows to send the imported items in batches of 250 references each. This is slower than the automatic system (sends 1000 references per batch) but should be entirely safe, as it's very unlikely that 250 references will ever hit the size threshold that makes large uploads fail (considering that 1000 references almost never do). The resulting upload is slower, but should be rock-solid.

Bugfixes

In the "Reports" tab. Designing reports was sometimes difficult, due to a combination of odd behaviours that could make it quite tricky to make sure the report columns appeared in the desired order. Also, reports that contained a large number of columns (in the range of 20 or more) could sometimes crash, after editing the details of a single column. Both bugs are now fixed, and designing reports should be considerably easier.

In "Collaborate\View(stats)\Reconcile" a bug was preventing the Reconciliation data to be loaded in the Reconciliation window, this happened when some items carried codes that have been detached from the code-set (one of their parent codes had been deleted). Problem is now fixed.