Logo and page links

Main menu

Helse Bergen HF

Helse Bergen HF’s AI project is based on figures showing that 10 per cent of somatic inpatients count for approximately 50 per cent of hospital stays and have frequent readmissions. This is not unique to the Helse Bergen HF numbers. It is a pattern documented elsewhere in Norway and the rest of the world.

ErlendHodneland_BW_300x300.jpg– Common to these regulars is that their physical state is bad. They’re not necessarily so old, but their health situation is complex due to multiple diagnoses across medical disciplines. Because of that, they do not always receive optimal follow-up, says Erlend Hodneland, who is a consultant at Helse Bergen HF.

The pilot project has shown that machine learning is able to predict with a relatively high probability who is in danger of being readmitted.

– If these patients are discharged with an extra follow-up label, we might give them better help from interdisciplinary health care teams. The goal is to reduce the risk for a quick return to the hospital, which is of course beneficial for both patients and the health service.

The sandbox process

In the sandbox, the project wants clarification on a number of issues related to legal basis, the patient's right to access and consent, as well as the right to confidentiality through the use of structured data sources where AI is intended to be included in the treatment of patients. These main challenges are also subjects relevant in future AI projects Helse Bergen HF might introduce.

The project plan is just around the corner, and will be published here shortly.