New Processes ‘Close the Loop’ on Imaging Findings
By Stacey Kusterbeck
Lack of follow-up on incidental radiology findings, a long-standing concern in the ED, can be prevented with good processes, according to researchers at Case Western Reserve University.1 “Actionable incidental imaging findings are becoming increasingly prevalent as the use of imaging has increased in recent years,” says Tianyuan Fu, MD, the study’s lead author and a resident at University Hospitals Cleveland.
Inadequate follow-up on incidental imaging findings is a well-documented issue across health systems throughout the United States.2,3 “Leadership from the radiology department and university hospitals were concerned about the impact to patient care from actionable incidental findings being lost to follow-up,” Fu explains.
Fu and colleagues at University Hospitals Cleveland Medical Center implemented a closed-loop actionable incidental findings program. Based on imaging volumes from the last several years, they estimated the number of actionable incidental findings. “We found that at least a few thousand of these findings are encountered every year, which could be lost to follow-up,” Fu reports.
The next step was to implement a standardized protocol to manage those findings. “Due to resource constraints, we decided to implement our process first in the ED,” Fu says.
Now, radiologists report the findings through a standardized form integrated in dictation software. This automatically sends an email to a nurse navigator, who documents the findings and coordinates follow-up with patients, primary care providers, and specialists. From July 2021 to May 2022, 1,207 incidental finding reports were submitted, with the vast majority identified on CT scans. Ten new cancers were detected as a result of the program. “EDs and radiology departments across the country should appreciate that a systemwide, standardized process for handling these findings may be beneficial, both from a patient care and a financial perspective,” Fu offers.
When actionable incidental findings are caught earlier, and patients undergo the appropriate follow-up imaging study or physician visit, patients experience better outcomes. “Therefore, it saves a healthcare system money in the long term,” Fu says. “It is also the case that it lessens the risk of malpractice litigation being filed against the hospital.”
Another group of researchers analyzed CT reports from trauma patients discharged from an ED in 2019.4 The goal was to automate the recognition of incidental radiology findings with a machine learning model using natural language processing. “The popularity of natural language processing for gaining insight into textual data is increasingly being recognized as a powerful technology across healthcare,” notes Christopher S. Evans, MD, MPH, the study’s lead author.
Incidental findings often are recorded in free text or semi-structured text data. This makes automated tracking and recognition challenging, considering the data on incidental findings generally are not discrete elements recorded in the electronic health record. “We aimed to train a natural language processing algorithm to recognize prespecified incidental findings on a training data set of radiology reports that were manually reviewed for the presence of incidental findings,” explains Evans, associate chief medical informatics officer at ECU Health and assistant professor of emergency medicine at East Carolina University.
Machine learning and natural language processing can classify incidental findings in ED patients’ CT reports with high sensitivity and with high negative predictive value. “As new technology using machine learning or artificial intelligence continues to be developed for emergency medicine applications, it is critical that emergency clinicians are intricately involved,” Evans says.
Evans urges EPs to be involved in developing, validating, and maintaining models used to aid clinical decision-making in the ED, including incidental radiology findings.
Meanwhile, many departments are finding ways to revamp processes to reduce risks of incidental findings. “As we order more imaging and technology improves, we encounter more of these ‘incidentalomas,’” says Deepak Chandwani, MD, a practicing EP and chief medical officer at The Mutual Risk Retention Group in Walnut Creek, CA. Chandwani sees these allegations in malpractice claims:
• The EP failed to inform the patient of the radiology findings. “These patients may have a potentially treatable condition that was otherwise preventable with early intervention,” Chandwani explains.
• The hospital failed to institute a system for reporting overreads to the ordering provider, primary care physician, or patient. To ensure that no incidental findings are overlooked, Chandwani says that ideally, ED providers should read the entire radiology report. “Alternatively, they should work in conjunction with their radiology colleagues to determine a consistent location of documenting incidental findings in the radiologist’s report. Ideally, placing them in the ‘impression’ section to minimize the risk of missing them in the ‘body’ of the report,” Chandwani offers.
Chandwani suggests emergency providers use this practical approach for incidental radiology findings: Print the radiology report and give it to the patient. Inform the patient about the findings at the bedside. Document the findings in the ED medical record. Add the incidental finding as a secondary diagnosis. Lastly, says Chandwani, “provide discharge instructions regarding the incidental finding and suggested follow-up.”
REFERENCES
1. Fu T, Berlin S, Gupta A, et al. Implementing a streamlined radiology workflow to close the loop on incidental imaging findings in the emergency department. J Digit Imaging 2023; Jan 17. doi: 10.1007/s10278-022-00773-x. [Online ahead of print].
2. Spruce MW, Bowman JA, Wilson AJ, Galante JM. Improving incidental finding documentation in trauma patients amidst poor access to follow-up care. J Surg Res 2020;248:62-68.
3. Mortani Barbosa EJ Jr, Osuntokun O. Incidental findings in thoracic CTs performed in trauma patients: An underestimated problem. Eur Radiol 2019;29:6772-6779.
4. Evans CS, Dorris HD, Kane MT, et al. A natural language processing and machine learning approach to identification of incidental radiology findings in trauma patients discharged from the emergency department. Ann Emerg Med 2023;81:262-269.
University Hospitals Cleveland Medical Center, radiologists report imaging findings through a standardized form integrated in dictation software. This automatically sends an email to a nurse navigator, who documents the findings and coordinates follow-up with patients, primary care providers, and specialists. From July 2021 to May 2022, 1,207 incidental finding reports were submitted, with the vast majority identified on CT scans. Ten new cancers were detected as a result of the program.
Subscribe Now for Access
You have reached your article limit for the month. We hope you found our articles both enjoyable and insightful. For information on new subscriptions, product trials, alternative billing arrangements or group and site discounts please call 800-688-2421. We look forward to having you as a long-term member of the Relias Media community.