Product Newsletter October 2021
The release of Discover 2.4.0 last week marks an important milestone for the product as it now allows customers to create insights across the Clinical Data Repository (CDR) domains.
The Discover Data Warehouse now includes star schemas for encounters, procedures, problems, results, and allergies, which enable ad-hoc analyses and dashboards to be developed on clinical data. Currently, this feature is available only to Oracle database customers, as future releases will support SQL Server.
Additionally, the underlying data warehouse has been redesigned to support ELT (Extract, Load and Transform). There are now two schemas in the data warehouse, a raw data store, and a presentation layer. By storing all the source data relevant to reporting in the raw data store, Discover can be used to answer virtually any question on CDR across the population. And if required, it can generate new aggregates or transformations without needing to access the source tables.
Dashboard developers and power users can now access the data warehouse via the new presentation layer. It is designed to only contain the fact and dimension tables used for analytics.
In the coming months, new out-of-the-box clinical dashboards will be added to the Discover Library.
If you would like to know more about the product or this release, please contact Anne O’Hanlon or Malvin Leong from the Discover team.
In Checklist 1.1, the Coordinate team has developed an exciting set of new features aimed at reducing time spent completing a checklist and improving communication for the various situations the patient might encounter.
Orion Health Checklists enable easy collaboration across the continuum of care and increase patient safety. A Checklist is a predefined set of action items or checks that can be added to a patient record. Each Checklist outlines the required actions that need to be taken for the patient in their current state or location. Examples of when a Checklist would be a vital tool include when admitting patients to a ward and transitioning patients from the operating theatre to recovery. The new features are:
– Bulk complete section (if appropriate): To avoid too many clicks, an easy “complete section” button has been added. Shown in A
– Check comment: To improve the accuracy of patients handovers, clinicians can now add a comment to specify something specific about a check. Shown in B
– Version History: it is now possible to review previous versions of the Checklist so that clinicians can see the state of the Checklist at every stage of the patient’s journey and follow their progress. Shown in C
Want to know more? Visit the webpage
As part of our consumer engagement offering, we have improved our patient forms and questionnaires capability to empower the consumer to take an active role in their wellbeing.
Patient forms enable a patient or related person (such as a patient representative) to choose and complete documents that have medical relevance. Completed forms are stored as part of the patient’s medical record and can be reviewed by clinicians. All forms are fully responsive and work on most devices to reach out to a bigger audience.
The most exciting features developed as part of this release are:
– The Form Dashboard: Here, the patient can find out what forms they need to complete and when they are due. This is designed to help the patient understand when to provide their clinical team with information.
– The Navigation Bar: This breaks down large forms into sections so the patient can navigate to specific sections, save their form as draft and come back to it at a later time to complete.
– The Conditional Fields: Questions in the forms can be configured to only show based on the patient’s previous answer. This reduces the noise of non-relevant questions and facilitates a better consumer experience.
– The Comment Feature: Patients can add optional comments to any field where applicable so they can always provide more information or ask questions to their care team. A comment will only appear if the patient wants to add one, saving valuable space on the form.
Updates for De-Identifier and Document Tagger
The De-Identifier team has enhanced several features on database de-identification such as filtered table list, supporting incremental and chunking for big tables, quasi attributes handling, and Oracle porting.
These have been added to support international customer commitments. The product has also added Differential Privacy to the algorithm set and this provides an alternative way to share useful statistical information about sensitive datasets.
For Document Tagger, the clinical concept linking now leverages the open-source BERT language model – first developed by google – for even better accuracy. Terminology support now includes the New Zealand adverse reaction manifestation reference set.
Orion Health Intelligence Services are working on a range of exciting research projects through Precision Driven Health
Enabling personalised self-care at home
Consumers have many digital tools available today that could support aspects of their healthcare, but there is no easy way to evaluate, select and engage with the ones that would be most beneficial to them. This research aims to understand the dynamics of patient engagement and interactions with existing and new technologies and tools to support self-care at home and is a collaboration between Te Hiku Hauora, Waitematā DHB, and Orion Health.
Smart Patient Cohort Builder (making use of the Document Tagger)
Patient cohort identification for clinical trials is currently still a manual and time-consuming task. Typical clinical trials have a set of inclusion and exclusion criteria that are not always obvious to detect – these criteria might not have been codified (structured), or they may be embedded in pages of free text (unstructured) in ways where basic string searches can’t easily work. In these cases, brute-force human effort, although extremely slow, is currently the best approach to identify patient cohorts for clinical trials. The goal of this research is to investigate whether an AI-assisted human-in-the-loop hybrid approach to patient cohort identification could significantly outperform human efforts alone, both in speed and accuracy.
Clinical Decision Support adoption in the Emergency Department
Clinical decision support (CDS) tools can improve patient outcomes by providing clinicians with timely information, usually at the point of care, to help inform decisions, reduce variation and potentially inequity in patient care. These tools work most effectively when embedded within the clinical workflow of users, which is often a critical factor in their adoption. This project intends to accelerate the path from research to practice for decision support tools, facilitating the adoption of both guidelines and data science findings. We will research, design, develop, test, and validate new software that reduces the barrier to adoption by focusing on current workflow.
Patients Like This
It is important to provide patients with reliable, accurate, and personalised information about their health conditions, to engage them in their own health. Whilst this usually works well for experienced doctors and receptive patients, not all doctors have the same wealth of experience, especially for rare conditions, and it can be hard to advise patients of the need to accept life-changing therapies or to adopt lifestyle changes. Consequently, we propose to develop an informatics dashboard to augment patient-doctor decision-making. This would quantify and demonstrate disease-related risk for an individual compared to other patients who have the same disease but variations in other health factors.