AI support for Chronic Patients

Danish municipality part of trial in using AI-generated support in the caretaking of chronic patients through telemedicine

Consultants on the case

Aksel Larsen

Katarina Gesser

Brage Emil Kringelbach

Emma Espersen

Johan-Oliver Krause

Sector: Medical

Region: Scandinavia

Machine learning

Artificial intelligence

Identify

Challenges and opportunities

Imagine a healthcare system where treatment and care for chronic patients have undergone a transformation, focusing on proactive maintenance of health and quality of life, rather than, as it is now, firefighting when familiar symptoms flare up or new ones appear. Furthermore, imagine that as a patient, you are given more freedom to live your life despite chronic illness. This is the vision of the collaborative trial project with the participation of Middelfart municipality, The Danish Centre for Clinical Artificial Intelligence, Appinux, and Copenhagen Data. The project is made possible by the Danish Business Authority and The European Union.

Copenhagen Data was contacted to help drive the trial project and implement an existing AI model to act as a decision aid for healthcare workers, enabling more proactive treatment for chronic patients. We have created an implementation plan that includes both technical and non-technical deliverables to deploy the AI model in a trial municipality, Middelfart. Furthermore, we have identified the long-term roadmap for scaling the use of the AI model to additional municipalities and hospitals and have determined the prerequisites for successful scaling.

Qualify

Analysis and evaluation

Several initial operations are and were necessary before Copenhagen Data is able to realize the project. We started analyzing, planning, and evaluating different aspects of the broader project. In order to issue a functioning AI-component which has actual real-world applications for everyday use, we would need an immense degree of knowledge surrounding the field of telemedicine. It has proven vital to gauge the initial technical landscape of Danish national telemedicine offers, the political landscape of stakeholders - both local and national, and finally, the work patterns of the local healthcare workers. We wanted to ensure that the final solution was of use to the people who would be using it - the healthcare workers. We focused on creating a sense of actual assistance, while not tampering and inhibiting the authority and knowledge that the healthcare workers have.

We qualified three tracks crucial to achieving the vision and ensuring real-world application: Technical implementation, organizational anchoring in the municipalities, and securing evidence/research results for scaling to new municipalities.

Realize

Implementation and scaling

As of August 2024, we are in the realization phase, which involves the technical implementation of the AI model and the configuration of its integration with the national telemedicine infrastructure. Additionally, the project partners are collaborating to prepare the organization in Middelfart to fully utilize the AI model in their daily tasks to better the treatment of chronic patients. Lastly, the project is focused on obtaining the necessary research and evidence to meet the requirements for scaling to new municipalities. These three tracks are all crucial to achieving the project's vision and supporting the research findings presented below.

Initial findings from the use of the AI model in an operational setup

Findings

The AI-model have been in use in a private organization since 2020. Early signs indicate that the proactive possibilities, issued by the use of the AI-component, reduces the need of patient hospitalizations and emergency treatments.