Version 4.5.0.1 is an important update, containing two major new features that will be significantly expanded in future releases. EPPI-Reviewer is now able to integrate with the (open source) R statistical software, making it possible to implement a whole range of new data analysis functionalities. The current version also implements the first milestone in our joint project with the Cochrane Collaboration. Cochrane authors can now access EPPI-Reviewer via their Archie credentials, gain access to EPPI-Reviewer and work on their registered Cochrane reviews, free of charge.

New Meta-analysis functionalities

EPPI-Reviewer is now able to communicate with the R statistical software. This feature relies on Microsoft Azure Services, where a dedicated instance of DeployR (includes Revolution R Open), by Revolution Analytics will process data and return results back to EPPI-Reviewer. In this implementation, Meta-Analyses are performed via the Metafor R package, developed by Wolfgang Viechtbauer.
To access the new functionalities, in the Meta-Analysis tab, click "Edit/Run" for any existing meta-analysis, or create a new Meta-Analysis. This opens a new Window where it is possible to:
- Select the outcome measures, using a brand new tabular interface (more below).
- Specify an extended range of options, including fixed effects as well as a range of different random effects models. For the details, please refer to the user manual.
- Send the data to R and receive a report which contains all the results (including forest plots, funnel plots and more). Reports can be saved as Word or HTML files. The report also includes the complete R script, enabling users to verify the results, and/or replicate and modify the analysis separately.

The "Edit/Run" (Setup Meta-Analysis) window also features a brand new system to list, display and select Outcome measures. Outcomes are now listed in a tabular format, one outcome per row, while (configurable) columns show associated data. All columns can be sorted and filtered to aid finding the relevant data. For example, Outcome Classifications are now shown on individual columns: to show (and analyse) only outcomes associated with one particular class you can set the filter of the relevant column to 1. You may also add new columns to list arbitrary codes (not only Intervention, Control, Outcome and Classification codes); note that this will only work for codes that belong to the same set as the outcome. Upon changing the Outcome type and/or the filtering options, the system will automatically select the Outcomes that fit the current criteria so users can unselect outcomes as/when needed. Outcome selections and filter settings are saved when clicking "save meta-analysis".

The old system to analyse data and produce plots is still available, for compatibility purposes. To use it, in the "Edit/Run" (Setup Meta-Analysis) window, select the desired outcome measures and click "save meta-analysis". This closes the window and returns to the main "Meta-analysis" tab. Click "Run (old method)" to produce a report which uses the old routines. We do not have any immediate plan to remove this legacy functionality from EPPI-Reviewer.

Cochrane/Archie Integration

The logon screen now includes a new "For Cochrane Authors: click HERE to logon with your Archie credentials" link. This functionality allows anyone registered as a Cochrane author to logon on EPPI-Reviewer and gain almost unrestricted access to the software. As part of the Cocharne Author Support Tool (CAST) project, Cochrane Authors can now use EPPI-Reviewer, free of charge, for the following scenarios:
- Work on private non-shareable reviews.
- Work on reviews that are registered in Archie (provided that their Archie account has access to them and that they are not checked out elsewhere).
- Work on prospective Archie reviews. The role of EPPI-Reviewer within the CAST process is to support the production of complex reviews, as well as develop and evaluate new reviewing methodologies. Thus, EPPI-Reviewer allows you to mark a review as a Prospective Cochrane review (strictly on-demand). This allows the reviewer to perform experimental/scoping/exploratory work in preparation of what is expected to become a new Cochrane review. To access this feature you will need to get in touch with us, without exceptions.
Licensing for non-Cochrane shared reviews remains unaffected. In such cases, both the review and the accounts used to work on it will still need to be licensed individually.

When using the ‘logon via Archie’ feature for the first time, after logging on with your Archie credentials, EPPI-Reviewer will ask the (partially) logged on user how to proceed. Two options are available: link an existing EPPI-Reviewer account to the current Archie Identity or create a new EPPI-Reviewer account. We encourage users to link their existing EPPI-Reviewer whenever possible: people should only create a new account on the fly if and when they don't have an EPPI-Reviewer account.

After completing the first logon procedure, users will be able to logon to EPPI-Reviewer via either their Archie credentials (using the link mentioned above) or their corresponding EPPI-Reviewer account. In both cases, the resulting authentication will grant access to the same set of features.

Note: the ‘logon via Archie’ features requires the user to enable pop-ups for eppi.ioe.ac.uk, as it is necessary to open a separate browser window where the Archie authentication takes place. This system uses the oAuth (V.2) standard so EPPI-Reviewer never needs to handle the Archie passwords. Note also that after the first logon EPPI-Reviewer will occasionally ask users to refresh their authentication on Archie.

Working on Registered Archie reviews:
In order to gain access to an Archie review, users need to check the review out from Archie. This marks the review as read-only in Archie/RevMan, and provides read and write access in EPPI-Reviewer. Note that at this stage, EPPI-Reviewer only exchanges logon IDs and access rights information with Archie: no study data is passed from Archie or CRS to EPPI-Reviewer. Data exchange features are under development, the current version handles licensing and access control only.

Small changes/fixes:

- 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, the order could appear to be completely arbitrary.
- In duplicate checking, clicking "get new duplicates" now shows a confirmation window. As duplicate checking can take a long time for big reviews users asked us to add an additional prompt, in order to ensure one wouldn't trigger the procedure by mistake.
- In the start Window, if Silverlight is not installed, not available or not supported, the system shows instructions on how to gain access to EPPI-Reviewer. This page now includes new instructions, specific to Windows 10.
Bugfix: 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.
Bugfix: version 4.4.4.1 introduced a bug in the "hotkeys" functionality (Document Details window). The bug meant one had to re-select the relevant branch in the code-tree each time a code was assigned via hotkeys, effectively removing the speed advantage conferred by hotkeys. This problem is now solved.