Background
Since 2007, the National Quality Register for Back Surgery (NKR) has registered more than 60,000 back surgeries and contributed significantly to quality improvement and research in back surgery. The register is now working on developing prediction models at the individual level based on artificial intelligence (AI), and wants to investigate whether such prediction models can be used in routine clinical practice as decision support integrated in the regular workflow.
DIPS ASA has provided EHR solutions since 1989, and is now the country's largest supplier. DIPS wants to be able to offer a solution for decision support based on data in the national quality registers, and sees NKR as an interesting partner in the development of such a solution.
The project is strategically anchored in Helse Nord RHF's Strategy for artificial intelligence and in UNN and UiT's joint Consortium for patient-centered artificial intelligence (CPCAI), and organizationally anchored in SPKI
Aim
Main aim
To develop an AI-based solution that uses the data from NKR for decision support to select whether to perform surgery or not. The tool must be fully integrated in DIPS and available in the clinician's regular workflow.
The main aim is achieved through the following sub-goals
- To develop a AI-based algorithm that uses the data base in NKR for precise prediction of outcomes after back surgery at the individual level.
- To develop an integration for data exchange between DIPS and NKR
- To develop a user interface for data registration to NKR via DIPS.
- To develop a clinical decision support tool for selecting patients for back surgery, and user interface for this in DIPS
- To conduct clinical trials and validation of the decision support tool
- Feasibility (phase 1) study to ensure that the solution has the necessary applicability and that it is accepted by clinicians and patients.
- Prospective observational study (phase 2) to investigate whether the solution is safe and whether it can provide the expected improvement in patient selection
- Prospective randomized study (phase 3) for assessment of effect sizes
- Nationwide register study (phase 4) for final assessment of effect at population level
In addition, it may be relevant to expand the project to also include development, integration of an evidence-based guideline in decision support.
Project partners
A number of partners are involved in this project. These include NKR, SPKI and the Norwegian Centre for e-Health Research, all of which are organizationally affiliated with UNN. UiT is represented through the Department of Clinical Medicine and the Machine Learning Group. DIPS ASA and their sister company Deepinsight AS are involved as commercial partners. The project takes place in close collaboration with Helse Nord ICT.
SPKI contributes with project management in this project.
