Region Hovedstaden to test artificial intelligence in psychiatric patient readmission prevention in Denmark

Upon discharge from psychiatric hospitals in Denmark, many patients experience a rapid return to inpatient care, with one in four readmitted within 30 days. To address this pressing issue, various strategies are being considered, including the potential use of artificial intelligence (AI). Starting next week, Region Hovedstaden will launch a trial to explore whether AI can assist in preventing premature discharges by ensuring that appropriate preventive measures are in place before patients leave the facility.

Professor Michael Eriksen Benros, from the Psychiatric Center of Copenhagen, explains that AI can help identify patients at high risk for acute readmission. By analyzing data from electronic health records, a model has been developed that acts as a virtual assistant for doctors, highlighting critical risk factors that may lead to readmission. These factors can range from unstable periods leading up to discharge, prior hospitalizations, psychiatric diagnoses, medications, and social circumstances.

The electronic health record system, known as Sundhedsplatformen, plays a central role in this initiative. It allows healthcare providers to document their work and enables patients to access their treatment information.

Despite the innovative approach, there are significant challenges in Danish psychiatry, including resource shortages and insufficient bed availability, often resulting in early discharges. While AI cannot solve all these problems, it may offer valuable support to clinicians navigating extensive patient information during their busy schedules. The project will initially test the AI model across four psychiatric departments in Region Hovedstaden, with the potential for broader application depending on the outcomes of the trials.

If successful, this initiative could extend across psychiatric services in both Region Hovedstaden and Region Sjælland, ultimately enhancing support for high-risk patients, including those prone to self-harm or suicidal behavior.