The rapid growth of artificial intelligence (AI) technology has created new chances to improve healthcare systems across the United States. One place where AI has made a clear difference is inside Electronic Health Records (EHRs), especially by adding Best Practice Advisory (BPA) alerts. These alerts, powered by AI, help doctors make decisions, improve how they work, and keep patients safer. For medical practice managers, owners, and IT staff in the U.S., it is important to understand how AI and BPA alerts work together in EHRs to help reduce medical mistakes, improve care coordination, and use resources better.<\/p>\n
This article talks about how AI-driven BPAs are now part of EHR systems to deal with problems like delayed follow-up on test results and unsafe opioid prescriptions. It shows what leading healthcare centers have done and explains how these tools can be used in clinics to support safe and timely care.<\/p>\n
Imaging tests are very important in finding diseases early. But a common problem in healthcare is missed or late follow-up on incidental findings. An incidental finding is a result found during an imaging test that is not related to the reason the test was done. Some of these findings can be serious, like lung lumps or adrenal gland problems, that need quick follow-up care.<\/p>\n
Northwestern Medicine, a large U.S. health system, created an AI program to fix this problem by adding it to their EHR workflows. Their AI uses Natural Language Processing (NLP) to read radiology reports and find mentions of incidental findings that need follow-up, especially for lung and adrenal issues. When the AI finds such a result, it sends a BPA alert in the doctor\u2019s EHR screen to guide them on the right follow-up steps.<\/p>\n
From December 2020 through more than a year, this system checked over 460,000 imaging studies. About 23,000 reports, or 5% of these exams, had lung follow-up advice. This means on average 68 findings are flagged each day that need quick clinical action.<\/p>\n
Before the AI was added, there was no special workflow to track these incidental findings. This led to delays and patient harm that could have been avoided. The AI-driven BPA alerts remind doctors in real time while they work. This helps them avoid missing important results. The system also contacts patients using online portals. If patients do not use portals or don\u2019t have a primary care doctor, nurses call them to make sure they know what to do.<\/p>\n
Mozziyar Etemadi, MD, PhD, Medical Director of Advanced Technologies at Northwestern Medicine, called delayed follow-up on incidental findings a type of \u201cpreventable harm.\u201d His team showed how common these findings are and why solutions are needed to lower risks to patients and healthcare costs. Missed lung findings alone are linked to more than $43 million in yearly U.S. malpractice payments.<\/p>\n
When building their AI system, Northwestern Medicine used a new method to label and train the AI. Nurses and frontline clinical staff who were on light-duty and could not do normal clinical tasks were trained to mark radiology report data in-house. This made sure the data used to train the AI was high-quality and clinically reviewed. It also avoided paying for outside services.<\/p>\n
The success of this program has led to expanding it beyond lung and adrenal findings. Now it also covers incidental findings in the liver, thyroid, and ovaries. This broadens the chance to reduce preventable harm.<\/p>\n
The opioid crisis in the U.S. is still a serious public health problem. Since the COVID-19 pandemic, fatal drug overdoses have gone up 29%. Some states like Ohio have their highest opioid overdose rates in ten years. Doctors face the hard job of managing pain well while lowering the risk of opioid misuse and overdoses.<\/p>\n
Cleveland Clinic has made important changes to improve opioid prescribing safety by putting AI-driven clinical support tools inside its Epic EHR system. Their Enterprise Pain Management Committee, led by Dr. Brendan Patterson, adjusts clinical workflows and sets up Best Practice Advisory alerts to warn about opioid risks and suggest safer options.<\/p>\n
Key parts of their method include:<\/p>\n
These upgrades have helped doctors reduce unsafe prescribing as overdose risks rise. Dr. Eric Boose, Associate Chief Medical Information Officer at Cleveland Clinic, said that customizing workflows all the time helps keep a balance between patient safety and work speed. Putting these tools inside the EHR avoided adding extra work for providers, which can block change.<\/p>\n
<\/p>\n