Version 22.214.171.124 includes two new features and a bugfix. The main new feature allows to export coding decisions in a machine-readable format (JSON), extending the functionality of the (quick) coding reports. The other new feature is aimed at users of the "Document Discovery Service" hosted by the National Institute for Health and Care Excellence (NICE), it allows to update the URL field of references, taking in the results produced via the document discovery service.
Export Coding Reports in JSON format: this is an experimental feature, a stepping stone in the road towards new enhanced interoperability with third-party tools. The longer term aim is to produce an API to allow other tools to exchange data with EPPI-Reviewer and naturally extend EPPI-Reviewer abilities to do the same. For the time being, users now have the possibility to export the coding data associated with any list of items, via the "Get Coding Report(s) for Selected Item(s) that have been completed" button (main toolbar). This functionality now includes an "Export results as JSON text file?" option. Clicking it will allow to save a file directly (does not show the file in the internal report viewer).
The exported JSON files produced in this way have the following high-level format:
- Files begin with a full representation of the selected Code-Sets, this ensures the exported information can be fully reconstructed while importing the data on the receiving software.
- The list of items follows. This includes all the main bibliographic fields that describe items in EPPI-Reviewer.
- For each item, a list of "ATTRIBUTES" is included, containing all the IDs of codes that belong to the current item. This allows to match an item to the correct codes as represented at the beginning of the file. The list includes details entered in the Info Box as well as full text associated with a given code.
- Each item them lists the outcomes associated with a selected code-set, the outcomes objects are represented in an extensive manner, including raw data and calculated values (Effect Sizes, Confidence Intervals and the like).
Overall, programmers and data scientists should find the format fairly self-explanatory. We have a few collaborators interested in using the files produced in this way. Our aim is to collect feedback so to inform our design of fully featured APIs in the future.
Import URLs from the NICE (National Institute for Health and Care Excellence) "Document Discovery" Service. As our collaboration with NICE grows, new features directed to NICE users and collaborators are starting to reach the production stage. In this case, the new feature can be used only by people who have access to the "Document Discovery" Service (other features are in the pipeline and will benefit all users). The functionality is made available in the "Manage sources / Import new" window, via a button in the "Manage sources" tab. Once the results of a document discovery exercise are exported from the NICE pages, the resulting file can be imported into EPPI-Reviewer. This will update the URL fields of the citations listed in the uploaded file. We are considering automating this system via API calls.
When exporting the list items to an HTML table (main toolbar, "print" button, "Export Current page/selection"), if items were selected (not selecting anything exports the full page) the list of items would stop updating: changing page and/or loading a new list would actually request and receive the new data, but the list of items wouldn't change. This bug is now resolved.