A patient enters the ED with severe lower abdominal pain and the ED physician knows just what to do. The patient is evaluated and sent for a CT scan. Just as suspected, the CT comes back positive for acute diverticulitis with a perforation. The patient is treated with drainage and IV antibiotics and the issue is resolved. But is it possible that this health system missed a care opportunity from this CT scan?
It turns out the patient has an extensive medical history. This was accentuated by his CT report, which on its own was two and a half pages long. In that report, the interpreting radiologist noted an aortic measurement of 3.7cm.
The ED physician did nothing wrong: they didn’t focus on the small aortic aneurysm (AA) as it was not the reason for the patient’s current, more critical issues, nor does this sort of finding necessitate immediate treatment – only follow-up to monitor growth. The acute pathology was dealt with and the patient left happy and, most importantly, without pain and additional risk of potential complications from the perforation. After all, the ED is flush with massive amounts of patient information: imaging and lab reports, pathology reports, test results and more- which can include any number of findings that are not associated with a patient’s current visit.
Even for the patient’s primary care provider (PCP), who receives the imaging report, there are many variables at play in health systems that can distract from necessary follow-up care. This includes an ever-increasing patient volume and healthcare staffing shortages to name a few. In short, there simply is not enough time in the day to execute a detailed read of every single report for every patient, for every test that comes in. Even if the PCP becomes aware of a smaller aneurysm, they may not know what needs to be done to manage the patient at that time.
But with AI, there can be a considerable shift in this unbalanced equation.
The Pitfalls of Untreated Aortic Aneurysms
The ED patient has gone back to life as usual, thriving in the five years that passed since their diverticulitis attack, with no issues necessitating another midsection CT scan. The aortic aneurysm, however, unbeknownst to the patient, could now be at risk for potential rupture given the predictable growth rate of the pathology. As there are typically no symptoms associated with an AA until it is so large that it is causing pain or ruptures, timely intervention is crucial.
What AI Brings to Aortic Aneurysm Care
The truth is that with inundated physicians throughout the health system, whether that be PCPs, ED physicians, or others, there are not many great answers to preventing patients from being lost to follow-up besides utilizing technology. Because patients easily “fall off the map” when it comes to AA follow-up, it’s crucial to have a tool that can account for:
- Finding AA cases
- Managing these patients
- Ensuring appropriate timeliness of intervention, when necessary.
This is where the clinical effectiveness of natural language processing (NLP) shines. In short, NLP allows computers to pull data from human language (dictated or written), and pass that information to relevant clinicians who can then make decisions based on that information. In practice, this means that AI would flag AA findings in dense radiology reports of patients who entered the health system, even those having had imaging done for other reasons. Following the identification, the AA finding would then be elevated to relevant specialists, allowing them to conduct any necessary follow-up care, whether that is precautionary measures like medication and dietary changes and more time-based needs such as clinic visits, additional imaging and at the appropriate time, possibly surgical intervention.
Reshaping the Future of AA Care
Thanks to the advancements of modern medicine, AA care is in a relatively good place. As long as the pathology is caught in a timely manner and managed appropriately, it often doesn’t warrant immediate or emergent surgical intervention, allowing for treatment options such as endovascular aneurysm repair (EVAR) (which is less invasive than open surgery), as well as helping patients improve their overall health with medication, smoking cessation, diet and/or exercise. Physicians in the know can monitor their AA patients for years before intervening. That said, it’s incredibly easy for these patients to be lost to follow-up, either because there aren’t enough safeguards in health systems to keep these non-urgent findings easily visible, or the finding is buried deep in a radiology report that physicians simply could miss entirely.
In conversations I’ve had throughout health systems and at conferences, this is a recurring theme amongst vascular surgeons: there has to be a good answer to follow-up, or lack thereof, for AA patients, especially ones that are not currently considered “urgent.” The advent of AI tools can deliver on this need, giving health systems, physicians and patients the ability to stay in contact and drive optimal care decisions, ultimately improving outcomes.
Click here to learn more about Aidoc’s impact in cardiovascular treatment.
