Incidental findings in diagnostic imaging are unexpected problems that show up but are not related to why the test was done. These findings may need more tests or treatment. Many times, these findings are missed or the follow-up is delayed, which can lead to diseases getting worse and legal problems for doctors and hospitals. Research shows that missed follow-ups on lung findings alone cost about $43 million each year in lawsuits across U.S. health systems.<\/p>\n
<\/p>\n
Northwestern Medicine created an AI system that works with their electronic health records (EHR). This system uses natural language processing (NLP) to find incidental findings in radiology reports. It then sends alerts called Best Practice Advisories (BPAs) to doctors during their normal work. In one year, the system checked about 460,000 imaging studies and flagged 23,000 cases needing lung follow-ups, about 68 cases each day.<\/p>\n
<\/p>\n
These alerts help stop delays in patient care and reduce harm by reminding doctors to act quickly. The alerts are part of the doctors\u2019 usual systems, so they don\u2019t need extra software. This helps doctors follow patient care rules better.<\/p>\n
Creating successful AI alert systems needs many healthcare workers to work together. At Northwestern Medicine, teams from radiology, quality assurance, patient safety, nursing, primary care, informatics, and process improvement all joined the effort to build the AI system.<\/p>\n
<\/p>\n
Radiologists give expert knowledge needed to understand imaging results. IT and informatics teams connect AI with electronic health records. Nurses help with patient contacts and follow-up, especially when patients do not use online portals. Quality and safety teams make sure alerts follow best care guidelines and reduce mistakes.<\/p>\n
<\/p>\n
One key part was training nurses and staff on light duty to mark (label) radiology reports. This helped create better AI training data without paying outside companies. This team effort made the AI system more accurate and reliable.<\/p>\n
It is important that AI alerts work inside the normal systems doctors already use. Northwestern Medicine\u2019s system shows alerts inside the electronic health record that doctors use daily, so they do not need to open other apps. This makes it easier and faster to act on findings.<\/p>\n
<\/p>\n
The alerts also guide doctors step-by-step to order the right tests or specialist visits. Patients can see study results and follow-up advice through online portals. For patients without portal access or a regular doctor, nurses call and help set up follow-ups. This ensures care continues smoothly.<\/p>\n
AI can improve safety in many areas beyond imaging follow-ups, including medication safety and helping doctors make better decisions. Federal groups like the Agency for Healthcare Research and Quality (AHRQ) show that clinical decision support (CDS) paired with computerized provider order entry (CPOE) helps reduce medication errors by making orders more consistent and automated.<\/p>\n
<\/p>\n
Most medicine errors happen when doctors prescribe drugs. The wrong dose is the top cause. AI and machine learning are used to reduce extra alerts that annoy doctors. Too many alerts can cause alert fatigue, where doctors ignore or turn off many warnings.<\/p>\n
<\/p>\n
One study showed that machine learning cut alert numbers by 54% but kept accuracy high. This means doctors can focus on the most important alerts. Reducing alert fatigue helps catch safety issues and makes doctors more satisfied.<\/p>\n
<\/p>\n
Hospitals using CPOE with CDS have seen a 78% rise in stopping medicines that may not be needed or could harm patients. This improves outcomes and lowers bad drug reactions.<\/p>\n
Even with benefits, using AI in healthcare is not easy. One problem is performance drift, where AI becomes less accurate over time because data or clinical situations change. For example, changes after events like a pandemic can affect AI accuracy.<\/p>\n
<\/p>\n
Many AI tools work like \u201cblack boxes,\u201d meaning doctors don\u2019t fully understand how AI makes decisions. This causes some distrust and hesitation to follow AI alerts. Experts recommend customizing AI and testing it regularly in the local setting to keep it working well for specific patients and clinics.<\/p>\n
<\/p>\n
Constant monitoring and feedback are needed to adjust AI systems and keep them accurate. This shows why teams of doctors, IT staff, and quality experts are needed to watch AI tools and make needed changes.<\/p>\n
Missed incidental findings can harm patients and cause costly lawsuits. The $43 million yearly cost from missed lung follow-ups reported by Northwestern Medicine shows how serious this is.<\/p>\n
<\/p>\n
Using AI alert systems to reduce missed or late follow-ups can stop serious health problems and save money for healthcare providers. Also, making patients part of the process with online portals improves trust and helps patients stay involved in their care. This extra step helps stop follow-ups from being missed.<\/p>\n
Multidisciplinary Team Formation:<\/b> Include clinical experts like radiologists, doctors, and nurses along with technical teams early to build AI tools that fit real work settings.<\/p>\n<\/li>\n In-House Data Labeling:<\/b> Use your own clinical staff to mark data if possible. This improves quality and lowers costs.<\/p>\n<\/li>\n Workflow Integration:<\/b> Make sure AI alerts appear inside the EHR systems doctors use every day. This helps increase use and quick response.<\/p>\n<\/li>\n Patient Communication:<\/b> Use patient portals or calls to keep patients informed about results and follow-ups.<\/p>\n<\/li>\n Ongoing Monitoring:<\/b> Plan to keep checking and updating AI systems with feedback from users to keep them working well.<\/p>\n<\/li>\n Address Alert Fatigue:<\/b> Use machine learning or rules to filter alerts and reduce overload on doctors.<\/p>\n<\/li>\n Compliance with Policies:<\/b> Follow federal rules like SAFER guides and CMS requirements when using AI.<\/p>\n<\/li>\n<\/ul>\n AI is also helping with routine tasks, lowering paperwork, and improving clinical workflows. Automation tools can scan lots of patient data in the EHR to find needs for medicine changes, duplicates, or stopping medicines. They also run order sets and reminders automatically and help with medication checks.<\/p>\n <\/p>\n For example, computerized provider order entry (CPOE) with clinical decision support reduces serious prescribing errors. When AI filtering is added, the number of alerts doctors get goes down without hurting patient safety. This helps doctors work better, lowers burnout, and supports safer medicine use.<\/p>\n <\/p>\n AI also helps fix medication lists and find mistakes when patients move between care settings. These errors are common and can be dangerous. Good AI can also predict bad drug reactions by using large medical records and past patient data. This can help prevent problems.<\/p>\n <\/p>\n Medical administrators and IT managers should keep these benefits in mind when updating technology. Starting AI use in important areas like order entry, medication safety, and follow-up reminders can improve quality and efficiency step by step.<\/p>\n AI alert systems for incidental findings and patient safety need more than just machines. They require healthcare workers, IT experts, patient safety staff, and leaders to work together. This combined effort makes AI tools more accurate and helps them fit well into clinical work, reducing missed care chances.<\/p>\n <\/p>\n In the United States, where healthcare is complex and regulated, systems built by teams following federal rules are more likely to succeed. These systems can lower preventable patient harm and costs, as shown by Northwestern Medicine\u2019s example. Medical leaders and IT managers should think about using similar approaches when adopting AI in their practices.<\/p>\n The article addresses the problem of delayed and missed follow-up on incidental diagnostic imaging findings, which can lead to patient harm and increased healthcare costs.<\/p>\n<\/p><\/div>\n<\/details>\n The AI system uses natural language processing (NLP) integrated with the electronic health record (EHR) to automatically identify radiology reports with incidental findings requiring follow-up and triggers alerts within the physician’s workflow.<\/p>\n<\/p><\/div>\n<\/details>\n The AI facilitates physician decision-making by identifying reports and triggering alerts but does not make clinical decisions, which remain the responsibility of radiologists and ordering physicians.<\/p>\n<\/p><\/div>\n<\/details>\n Physicians receive a Best Practice Advisory (BPA) alert directly within the EHR, which displays findings and provides workflows to order appropriate follow-up studies.<\/p>\n<\/p><\/div>\n<\/details>\n Patients receive notifications through their online portals with study results; if they do not use the portal or have no primary physician, follow-up nurses manage direct outreach to ensure care continuity.<\/p>\n<\/p><\/div>\n<\/details>\n In one year, over 460,000 imaging studies were screened with 23,000 lung findings flagged requiring follow-up, demonstrating the prevalence of incidental findings and effectiveness of the AI alert system in managing follow-ups.<\/p>\n<\/p><\/div>\n<\/details>\n Northwestern Medicine used trained nurses and front-line staff on light-duty to annotate and label relevant radiology report data in-house, ensuring high-quality, expert-reviewed data effectively and cost-efficiently.<\/p>\n<\/p><\/div>\n<\/details>\n A multidisciplinary team from Radiology, Quality, Patient Safety, Process Improvement, Primary Care, Nursing, Informatics, and others collaborated to design and implement the AI follow-up alert system.<\/p>\n<\/p><\/div>\n<\/details>\n Integration ensures alerts appear in the existing physician workflow without requiring additional software access, improving usability and response time to incidental findings.<\/p>\n<\/p><\/div>\n<\/details>\n Yes, the system is being expanded to cover hepatic, thyroid, and ovarian findings requiring follow-up to further reduce missed or delayed care across more conditions.<\/p>\n<\/p><\/div>\n<\/details><\/div>\n<\/section>\n","protected":false},"excerpt":{"rendered":" Incidental findings in diagnostic imaging are unexpected problems that show up but are not related to why the test was done. These findings may need more tests or treatment. Many times, these findings are missed or the follow-up is delayed, which can lead to diseases getting worse and legal problems for doctors and hospitals. Research […]<\/p>\n","protected":false},"author":6,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[],"tags":[],"class_list":["post-142147","post","type-post","status-publish","format-standard","hentry"],"acf":[],"aioseo_notices":[],"_links":{"self":[{"href":"https:\/\/www.simbo.ai\/blog\/wp-json\/wp\/v2\/posts\/142147","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.simbo.ai\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.simbo.ai\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.simbo.ai\/blog\/wp-json\/wp\/v2\/users\/6"}],"replies":[{"embeddable":true,"href":"https:\/\/www.simbo.ai\/blog\/wp-json\/wp\/v2\/comments?post=142147"}],"version-history":[{"count":0,"href":"https:\/\/www.simbo.ai\/blog\/wp-json\/wp\/v2\/posts\/142147\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.simbo.ai\/blog\/wp-json\/wp\/v2\/media?parent=142147"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.simbo.ai\/blog\/wp-json\/wp\/v2\/categories?post=142147"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.simbo.ai\/blog\/wp-json\/wp\/v2\/tags?post=142147"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}AI-Driven Workflow Enhancements Relevant to Healthcare Administration<\/h2>\n
Final Notes on Collaborative AI Design in Healthcare Settings<\/h2>\n
Frequently Asked Questions<\/h2>\n
What is the primary healthcare problem addressed by AI in the article?<\/summary>\n
How does Northwestern Medicine\u2019s AI system detect incidental findings?<\/summary>\n
What role does the AI system play in clinical decision-making?<\/summary>\n
How are physicians notified of incidental findings in the AI system?<\/summary>\n
What measures are taken to ensure patient awareness of incidental findings?<\/summary>\n
What were the results after implementing the AI system at Northwestern Medicine?<\/summary>\n
How was the large data labeling task for the AI system managed?<\/summary>\n
What departments collaborated in developing this AI system?<\/summary>\n
What is the significance of integrating AI alerts directly into the EHR?<\/summary>\n
Is Northwestern Medicine planning to expand the AI system to other diagnostic areas?<\/summary>\n