Using AI to Improve Patient Care
The benefits of AI in health care
As AI and machine learning continue to evolve, we remain dedicated to using this groundbreaking technology to benefit our patients. That means staying on top of the latest advances and prioritizing ways to implement them responsibly.
Below are a few real-world examples of how UCLA Health is using AI to improve the quality of care we deliver. Our team is at the forefront of using AI to prevent disease, identify at-risk patients earlier than ever before and ensure patients have access to lifesaving care.
Identifying patients with a higher risk of chronic kidney disease
Chronic kidney disease (CKD) affects about 15% of adults in the United States . Untreated, CKD can lead to kidney failure. Early treatment can often help restore kidney function or slow the progression of the disease. Certain patients have a higher risk of rapidly declining kidney function, yet there is no established method of identifying these at-risk patients.
The UCLA Health Faculty Practice Group’s Office of Population Health and Accountable Care (OPHAC) and the UCLA Clinical and Translational Science Institute (UCLA CTSI) worked together to find a solution to this problem. They developed a machine learning model that detects early patterns of kidney function. These patterns can predict worsening symptoms of CKD.
UCLA Health providers are using this technology to evaluate patients with kidney disease. As a result, our kidney specialists are better able to prevent CKD progression by offering comprehensive preventive care.
Reducing unplanned hospital admissions
Our team is committed to reducing unplanned admissions to our hospitals and emergency rooms. By reducing these admissions, we aim to decrease the overall costs of health care and reserve hospital resources for visits that are truly necessary.
The OPHAC worked with the UCLA Clinical Translational Science Institute to identify patients at risk of unnecessary hospital visits. Together they developed a machine language model that flags people who have a higher chance of being admitted to the hospital within the upcoming year.
Our health care providers and staff reach out to these patients early and often so we can address issues before they arise. We assess each patient’s health, offer guidance and schedule appointments for preventive care. The technology enables us to be proactive and help our patients avoid future health issues that might lead to an unnecessary hospital stay.
Encouraging colonoscopy screenings for people with a higher risk of colon cancer
People with a normal risk of colon cancer should have a colonoscopy once every ten years. Patients who receive abnormal colonoscopy results need to schedule follow-up colonoscopies at least every three years. An abnormal colonoscopy can indicate a higher risk of developing colorectal cancer. But sometimes, these follow-up recommendations go unnoticed. Some patients might not be aware of their risk status and may not schedule another colonoscopy within the suggested time frame.
To address this issue, the collaborated with researchers in the UCLA Health Division of Digestive Diseases. They worked together to identify these patients using automated natural language processing (NLP) tools. These tools gather results from the patient’s colonoscopy reports and pathology reports. The NLP tools then merge the results in the patient’s electronic medical record (EMR). UCLA Health providers see a notification in that patient’s EMR and work with them to schedule a follow-up colonoscopy and other necessary tests.
Safeguarding patient privacy
With the use of AI technology for patient privacy monitoring, UCLA Health Office of Compliance Services can achieve more than humanly possible with fewer resources. This technology enables us to rapidly detect potential privacy violations and unauthorized access from both internal and external sources. We use AI to prevent potential data breaches and safeguard our patients’ confidential health information.
Maximizing efficiency in doctor-patient communications
Online patient portals allow doctors and patients more ways to communicate about test results, medication questions and general health inquiries. Yet the volume of messages has increased significantly. Health care providers spend a considerable amount of time responding to these messages. This additional time adds to providers’ workloads and means longer response times for patients.
UCLA Health, in partnership with our EMR vendor, is now using a generative pretrained transformer (GPT) to address this issue. This generative AI tool drafts responses to patients’ messages. The doctors then review the messages and edit them as necessary. UCLA Health is leveraging this advanced technology to enable doctors to respond to their patients with accurate information more quickly than ever before.