CTcue version update: 4.4.0 & 4.4.1 

The latest version of CTcue is ready for rollout: v4.4. Below is a more detailed overview of some of the bigger changes in both versions v4.4.0 and v4.4.1, besides some minor improvements and bug fixes. To read the full release notes, click here for v4.4.0 and click here for v4.4.1. 

Reports plus in the Data Collector 

After releasing reports plus in the Patient Finder in version 4.1.1 and rolling out various updates in version 4.3.1, we are now bringing reports plus to the Data Collector. For now, this will remain an experimental feature behind a permission, to allow us to perform further tests on real hospital data and collect initial user feedback from a few selected hospitals. Once we are confident that the feature is ready for use by the broader public, we will remove the experimental flag.  

We have made a few further improvements to reports plus in both the patient finder and data collector.  

One major change is that you can now specify whether you are searching for concepts that apply or don't apply to a patient. If you are looking for patients with diabetes, then you will use the “applies to patient” setting. If you are looking for patients who don't smoke, then you can change the applicability setting to “does not apply to patient” and search for smoking. This will allow you to search for negated concepts while still using the same relevancy settings. 

Multiple concepts may be present in the same piece of text, so to make it clear which term is the best AI processing result, we have added additional highlighting to the selected term. 

To allow users to analyze the full context, we display all found concepts in a report when at least one concept matches the relevance filters that were selected. This means that if you have only selected relevant concepts (and thus deselected irrelevant and unsure concepts), then we will also display any irrelevant and unsure concepts when these are found in the same report. 

Reports plus terms are now taken into account in determining whether a patient matches the patient finder criteria. In the previous release, when adding multiple terms in the criteria, we would return patients that match other terms regardless whether they also match on the reports plus terms. Now that reports plus terms are taken into account as well, we would only return patients that match all selected terms, including the reports plus terms. 

The “AI processing” button has been renamed to “AI assisted search” for more clarity to users. 

We have added a description and labels to filters that impact the relevance of reports plus terms. This will help users understand what the impact adding certain filters could have on the outcomes while they are building their query. 

We have added some initial documentation to explain how confidence and relevance are determined. This documentation is accessible straight from the reports plus section in the app. We will continue to write full documentation about the reports plus feature in our knowledge bank and will publish this before the feature becomes widely available without the experimental flag. 

Improved identifiable data permissions 

We have had many requests from users to change the identifiable data permissions to allow for more granular configurations and in this release we have done just that. Access to identifiable data should be kept to a minimum as much as possible, so the main goals for change were to make sure that users only get access to the identifiable data fields that they really need and to enable administrators to monitor user access more easily. 

As part of CTcue's privacy-by-design approach, all data is pseudonymized. There are a few data fields that can be made available to users through permissions that can be set by the hospital administrator in the user profile: patient ID, zip code, place of birth, country of birth, family relations (more on this below), birth date and date at time of death.  

Previously, there were three options: no access to identifiable data, limited access to identifiable data based on DBC or full access to identifiable data. With full and DBC-limited permissions, users would have access to all identifiable data fields. However, in most cases users would only need access to some specific fields rather than all. So to minimize unnecessary access as much as possible, hospital admins can now specify which fields users get access to. Next to that, admins can limit this access to specific projects and in Dutch hospitals this can be further limited to specific DBCs. The permissions can be provided to multiple projects and DBCs at the same time. Be aware, though, that you cannot grant specific permissions for different projects. 

To make it easier for hospital administrators to manage identifiable data permissions, there is now an option to set an expiration date for the permissions. When the expiration date is reached, the permissions will be reset to no permissions. We have made it easier for hospital admins to monitor access to identifiable data directly in the user overview by adding a filter option for data permissions and displaying an icon that indicates the expiry date if applicable.

Family relations 

Another request that we have received from various hospitals is to link mothers to their children in the app. This would make it a lot easier to do any neonatal research, for example. Users have asked to see the pseudoIDs for children of a mother, so they can easily copy these over to another project to collect data for them.  

To facilitate this, we have now added a section called "family relations” to the patient information page. This section currently only contains mother-child relations, but could potentially be extended to other familial relations in the future, should there be any need and available data. The family relations section shows all known children for a female patient. This is presented in both a family tree style graphic as well as a list overview for easy copy-pasting. To help identify specific children, we also display the gender and birth year. The family relations information works both ways, so you can also identify the mother from the patient information of the child. Should that data exist in the EHR, then the family tree consists at most of three generations: mother of current patient, current patient, and children of current patient. 

To see the family relations information, the hospital administrator will have to enable the “family relations” permission in the identifiable data permissions for a user (read more on this above). 

Customise columns for all results questions 

Customising the columns for all results questions in the data validation table of the Data Collector used to require a couple of steps that were not very intuitive, even for experienced users. Now we have added a customise columns icon for all results questions on the overview page just like there already is for single result and multiple choice questions. With this button you can select which columns you want to show or hide for any of the terms in an all results question. This also makes it easier to control which columns you include in your export. When multiple categories are included in the same all results question, then you can select the relevant columns for each category by clicking the matching tab at the top of the modal. 

Patient information consolidated on a single page 

To prevent unnecessary duplication in demographics search results, we have consolidated all the patient information on a single page that can be accessed from the single patient view. Note that when you have included patient information in your search criteria, the patient information columns will not be visible by default to reduce clutter. However, you can always toggle the visibility of these columns through the ‘criteria selection’ sidebar. 

Multiple choice answer page redesigned (v4.4.1) 

The multiple choice answer screen in the data collector is a page with a lot of information and possible actions. For many users it proved to be confusing how to browse through different answer options and how to select a different answer for your question. To make this page easier to use, we have reorganised the different elements so that navigational actions are more clearly separated from selecting actions. Navigating between the answer options now happens in the left side bar instead of using a dropdown menu on the main screen. A “selected answer” label in the sidebar indicates which answer is selected, while a blue line in front of the answer option indicates which answer you are viewing. Viewing and selecting answers then happens on the main screen. When an answer option includes multiple categories, then the results for each category are displayed on separate tabs at the top instead of underneath one another. To provide more overview we now show the number of found results for each answer option in the sidebar and for each category in the tabs at the top. 

The redesigned multiple choice answer page: browsing answer options (1), browsing category tabs (2), selecting a custom result (3): 

New data fields 

We have made the following new data fields available within the app: 

• For the terms “administrative invoice” and “administrative medication” you can now search for the requesting and executing physician (Belgium hospitals only) 
• For the “admissions” term you can now filter by emergency (HiX, Epic and Nexus EHRs only)