In HD4DP reopening a participation document is now possible.
In HD4RES when editing participation content without saving, a popup with warning 'unsaved changes' is shown to avoid for users to lose changes.
In HD4DP the possibility of downloading attachments in the registration is added.
When using the upload center, the cancellation of an upload is made available in the upload step.
The password encryption was improved to meet the current security standards.
Use full reference list cache when entering registrations and uploading documents for performance reasons.
The logging mechanism is only storing the necessary data to improve performance. In case of a problem full logging can be requested.
The elastic search used for minimizing the time to open the overview of the registrations is only indexing the fields on which a user can filter and sort. Bulk indexing for uploads is additionally enabled.
The same version of elastic search is used in all HD4* applications to facilitate the maintenance of all applications.
In HD4DP enable the possibilities of the actions button when multiple registrations are selected.
Missing translations in Dutch and French are added in the application.
Elastic search split status
In the admin panels the application gives a health status to see if elastic search is working properly, but this was not interpretable in a nuanced way, giving the impression it was down, while actually it was not.
Now we make a distinction between if a cluster is running or not and the amount of successfully indexed documents in elastic search. A NOK for the first is an issue to be looked at immediately, while the second doesn't need to be worrisome.
This distinction allows us and our partner Arxus to better monitor the status and health of the system for our users ( for more information on the Elastic Search monitoring ).
Task rescheduling was made more reliable
We noticed that sometimes when a task failed, it was not rescheduled automatically, f.e. when the database connection would be lost. This system was made more reliable for all applications, diminishing the chance that the applications stop running.
HD4DP checks GUI_HOST regularly to catch SLL errors
When a GUI_HOST is misconfigured in HD4DP, the auto-upgrade fails and users cannot use the application. Now the GUI_HOST parameter of a HD4DP is checked and posted to the catalogue. The upgrade will not continue when misconfigured.
HD4DP 'CodedContent' can be empty
Normally, in a DCD there is at least one field in the 'CodedContent', containing patient data, but in some registries this is not the case. In the latter situation, the data could not be processed by the applications. This has been modified so that the applications can handle an empty coded content.
Catalogue can handle special characters in the configuration of a DCD name
When a DCD name was configured with special characters, it was not possible to save or submit a registration for this particular DCD. This kind of configuration is now possible.
Bug fix for language issues
For one registry there was a mismatch between the language in the sidebar and the title header. The cause was an error in the place where these translations were retrieved.
When deleting a DCD, follow-ups can be deleted
Because of a bug, DCDs with follow-ups could not be deleted.
After elastic search rebuild status ‘Delayed’ is ignored
When uploading a CSV in HD4DP, when the data are saved (and not submitted), the elastic search status is put in ‘delayed’. This means it cannot be indexed, because it is not fully available in the database.
If the admin would rebuild elastic search, the status was put incorrectly in ‘dirty’ (which means it will be indexed). These data would get a status 'Error'.
Now when the admin rebuilds or repairs elastic search, delayed data will be ignored.
Validation for modulo 89
RIZIV started using modulo 89 for new RIZIV numbers this year. Our code uses a validation on the modulo to distinguish between a RZIV and a NISS number. Without a change into this validation we cannot guarantee the correctness of this identification. This validation has been expanded.
Timely resend messages from HD4DP
Before, when sending a message failed, the applications HD4DP/HD4RES would retry sending five times immediately after each other. Implemented this way, the function would not increase the chance of sending messages successfully, diminishing the number of messages in error. The retries are now spread over a broader timeframe.
Improve usage for the task scheduling
When restarting the applications, task executions were removed. We now retain this information for
debugging purposes, so that we can diagnose and resolve issues more efficiently.
Remove deadlock making task rescheduling fail
A deadlock exception on MSSQL server database was removed which caused the task rescheduling to fail sometimes and the sending and receiving of messages to stop running. Our task scheduling is now more reliable.
TPR Registration Workflow is not mapped
A piece of code was changed but because of an inconsistency between the code of HD4RES/HD4DP and HD4TPR, where in the latter the domain object was called differently, this generated errors.
Since this piece of code only impacted the frontend, which is not used for TPR, it was disabled for this application.
ClassCastException on resending failed messages
In HD4DP when a message cannot be sent, the application will retry sending five times. This was not working for HD4DPs running on an Oracle database.
Can't submit DCD with external reference list
A bug was resolved impeded the retrieval of external reference lists. As a consequence, records using these lists could not be submitted.
*A hotfix was added to the version 1.15.9 to resolve a bug with the CSV download feature.
Inclusion of both code and label in CSV download in HD4DP
Before, when downloading a CSV with records in HD4DP, only the code of the value was included.
Now in this download, both the code and the label of the value are shown.
French translation of Elastic Search menu were not correct
Before fix (screenshot was made before aforementioned changes to the status menu, see v 1.15.6).
Modification to SADMI functionality
SADMI is a hard coded reference list module used in projects such as RCT and Pacemaker. It uses an external reference list source at eHealth.
A modification of the labels of this module in HD4DP was requested by RIZIV.
HD4DP - Resubmit registrations with status Error
For DCD’s using registration code generation, new registrations sometimes end up in status “Error”. When this happens, the end user is blocked and an intervention by support is necessary. A feature has been added to let the end user re-submit records in status Error. This can be done either from the list view, or from the registration detail view.
HD4RES - status messages info available to all users
Every HD4DP installation sends technical status messages to the central HD4RES platform, on a daily basis. This is to test the health of the message flow between HD4DP and HD4RES. In HD4RES, admins can monitor the current status of the message flow. As of this version, this feature has been disclosed to all types of HD4RES users. Users can access this feature via the “Settings” page.
After clicking on “Status messages list”, they land on a page where they can check the status for each DP.
HD4DP - Reference-list options were not shown when there is only one item
When a reference list contains only 1 item, the dropdown representation didn’t work properly when filling out a registration in the HD4DP application. This has now been resolved.
HD4DP - Fix registrations with status Error
For DCD’s using registration code generation, new registrations sometimes end up in status “Error”. This has now been resolved.
HD4* - Disable html logging
HD4DP log files used to be saved in 2 formats, a plain text format and an html format. We have disabled logging in the html format, since this has shown to have an impact on performance. Logging to plain text files has been preserved, of course.
HD4DP - Refresh status in upload center when upload is queued
When doing a CSV upload via the upload center, your upload request can get queued. Previously, when this happened, the end user was not notified when the request gets dequeued and processed. Now, the screen is auto-refreshed and the end-user is notified when the upload request is handled.