AI can predict hospital stay lengths for people with learning disabilities

Hospitals thrive on precision. Every bed occupied is a resource tied up, and every prolonged stay is a challenge for both patients and providers. Now, a breakthrough AI model from Loughborough University is helping hospitals get ahead of the curve. Designed to predict how long people with learning disabilities and multiple health conditions will remain in hospital, this innovation could transform care planning, making stays more efficient and tailored to individual needs.

How does it work?

The AI model, developed under the DECODE project, crunches data from more than 9,600 patients with learning disabilities and multiple health conditions. Within just 24 hours of admission, it analyzes key factors like:

• Age
• Medication history
• Lifestyle choices
• Pre-existing conditions

Professor Georgina Cosma, an expert in AI for healthcare at Loughborough University, explains: "With early and accurate predictions, hospitals can plan better and provide more personalized care, ensuring fair treatment for all patients."

And it's not just theoretical — the model has already proven 76% effective in distinguishing between patients who need extended care and those likely to be discharged sooner.

Why does it matter?

People with learning disabilities already face major healthcare hurdles, from delayed treatments to misdiagnoses. Having a system that flags potential long stays early can mean better resource allocation and, ultimately, better patient outcomes.

The findings are eye-opening:

• Cancer tops the list of hospital admission causes for both men and women with learning disabilities.
• Epilepsy is the most frequently treated condition during hospital stays.
• Mental illness often leads to the longest hospitalizations, sometimes exceeding 129 days.

This isn't just about logistics — it's about quality of life. The NHS plans to use insights from this study to refine its risk prediction models, improving care coordination for vulnerable patients.

The context

The DECODE project is no small endeavor. Funded by the National Institute for Health and Care Research (NIHR), it brings together experts from seven UK universities and partner hospitals to tackle a long-standing issue: how to enhance care coordination for individuals with intellectual disabilities and multiple long-term conditions.

The next step? Scaling up.

Professor Thomas Jun, co-principal investigator of DECODE, shares what's ahead: "We're now expanding our study to include a more diverse group of over 20,000 patients across England to ensure our predictive model is as accurate and effective as possible."

The team is also seeking funding for clinical trials to test whether these predictions can reduce emergency admissions and improve patient well-being. And while the current model was trained using Welsh hospital data, researchers are now applying it to English hospitals to see if similar trends hold.