The The National University Health System recently introduced the possibility of a new AI platform that can, among other things, predict hospital bed availability two weeks in advance.
Its AI ENDEAVOR platform integrates data directly from the next generation EMR (NGEMR) to compute multiple AI insights.
It hosts a variety of AI tools, one of which can predict the estimated length of stay for each patient admitted to three NUHS public hospitals. The AI engine does this by reading patient history and doctor’s notes in real time, running up to 30 times per hour. It can also provide clinical insights into factors that contribute to a prolonged patient stay.
According to a press release, the accuracy of the AI has been validated using NGEMR data for the past six months. It is said that it can predict bed status two weeks in advance to optimize bed capacity and patient placement.
WHY IT IMPORTANT
The predictor is NUHS’s solution to increasing bed occupancy and increasing bed wait times at the ED. It allows doctors to intervene early to predict problems. It could, for example, flag patients who have stayed in the hospital for the past two weeks, allow medical teams to change their management, or plan for early transfer of patients to a community hospital for rehabilitation. power.
With the ability to read notes, vital signs, and other lab reports, the AI tool can also predict a hospitalized patient’s risk of deterioration.
NUHS plans to further develop AI’s ability to recommend care plans “that [can] change the trajectory of the patient’s disease process.”
Meanwhile, ENDEAVOR AI is also capable of automatically alerting administrators of elevated ED timeouts, allowing early human resource activation. The timeout can be reduced from 30 minutes to hours, depending on the resource implementation.
TREND TO BIGGER WOMAN
NUHS also revealed plans to evaluate and then deploy an imaging AI model to enhance the X-ray assessment of scoliosis. Approximately 7,000 x-rays are examined each year through Singapore’s Scoliosis Screening Program. Doctors are still measuring spinal curvature manually, which can be time-consuming and error-prone. Results may also take time to be communicated.
A new AI model is being designed to automatically measure the extent of scoliosis, improving doctors’ ability to interpret scans. Early trials show that the AI model reduces reporting times with reasonable accuracy, potentially helping to improve clinician productivity and communicate results and specialist referrals and treatment. earlier.
ON PROFILE
“We leverage AI to improve healthcare operations and outcomes, enabling clinicians to deliver faster, more accurate diagnoses and treatments. Today’s healthcare aggregates large amounts of data, but most of this data is analyzed retrospectively.With the technology in ENDEAVOR AI, we can now say it’s now possible to Ngiam Kee Yuan, associate professor and CTO at NUHS. Real-time data transmission feeds AI models that generate instantly actionable insights, resulting in better patient outcomes.
