As health centers work to lower reject rates, one technology is automating a once-manual process for better quality assurance and an easier experience for all.
The challenges of X-ray reject rates, or when scans are deemed unusable due to poor image quality, can pose an administrative nightmare — increasing cost, radiation dose, and inefficiencies for all. And the burden is especially high for patients, who may need to take repeat exams to get the right scan of the right area at the right time.
Repeat/Reject analysis requires hard, manual work — taking up to 7 hours of manual work to access reject data,1 which poses a drain to time, resources, and staff. While that’s a big improvement from searching through film bins where the technologists would manually review and dispose rejected films of X-ray images one-by-one, the “click-review-reject-repeat” method of reviewing digital X-ray images still takes time that medical professionals and patients don’t have.2
Recommended reject rate thresholds are low; the American Association of Physicists in Medicine (AAPM) recommends a reject rate of 8 percent or less for adults and 5 percent or less for pediatrics.3 Achieving those metrics can be difficult, but new technologies can help. Through quality assurance applications that automate once-manual processes, X-ray technologists, physicists, radiologists, physicians, and patients can enjoy easier, more effective, and less costly imaging procedures — while giving X-ray technologists the real-time feedback they need to get better, one scan at a time.4
To learn more about the clinical and operational impact of such technologies, we sat down with Katelyn Nye, a global product manager at GE Healthcare who focuses on X-ray artificial intelligence and analytics. Nye and her team have worked to bring quality assurance automation to health centers around the world, notably through GE Healthcare’s X-ray Quality Application.5
In what ways have the demands for X-ray quality assurance changed in recent years? What challenges do they cause?
“Ever since X-rays have existed, there have been procedures in place to try and keep quality high and to keep technologists trained and accountable. But as X-ray capabilities have become increasingly digital, volumes have gone up significantly, and a lot more people are getting scans. Departments are asked to be more efficient, even when there are hundreds of images generated per day. As a result, all of the reject information is stored individually on each machine — and quality assurance technologists have to go hunting on each machine to download the files, scan through them, understand why the rejects are happening, and try to find who performed the scans to conduct training because there’s not always traceability for that on the machines themselves.”
How can newer technologies help address those challenges?
“We’ve created a piece of software called AXIS [Analytics X-ray Ingestion Service], which is installed on-premise on a customer-owned server. It runs on a Windows virtual machine and can pull and analyze quality assurance data at selected intervals, and we recommend once every 24 hours. So instead of having to manually pull different log files, you’re automatically getting an aggregated database created for all quality assurance data from all your connected machines on a daily basis.”
What can healthcare facilities do with that data?
“The data goes through the X-ray Quality Application, which is a set of powerful analytics visualization tools that help users interpret and understand trends to drive action. That way, you can break down and understand the repeat reject analysis, as we call it — which means you can filter reject rates by devices, departments, technologists, reject reasons, or exam types to understand what’s contributing the most to your reject rate so that you can address the issues. For example, you could spend more time with individual techs to train them on specific positioning techniques — or employ department-wide trainings if an entire department struggles with a certain type of exam.”
What are some of the biggest drivers of rejected images?
“Patient positioning is consistently the leading reject reason. I’ve seen reports indicating that up to 60 or 70 percent of the time, it’s due to that reason alone. Sometimes, that’s unavoidable — say if you have a sick patient or someone in a ton of pain struggling to stay in position. But a lot of that can come down to training for technologists. It’s pretty common for us to see that the most senior X-ray technologists move on to CT or MRI for more advanced imaging, which may leave more junior technologists with less experience handling X-ray acquisitions.”
Do images ever get rejected that shouldn’t be?
“When images get rejected, they never make it to the radiologist — so it can be hard to tell without reviewing the rejected images with radiologists on a regular basis. The X-ray Quality App uniquely pulls the rejected images off of the systems and allows them to be easily viewed in a web browser. Previous research has shown that sometimes, reject standards are too conservative — you might show a radiologist a rejected image and they say, ‘I would have been able to read that. It didn’t need to be rejected.’ With our technology, though, you can actually have a radiologist as part of your quality assurance team who can log into the tool and analyze the images being rejected — and be an open part of the discussion to say whether or not they were appropriate rejects. You could then go back to the tech with that feedback.”
Why should technologists be excited about this technology, rather than intimidated for being called out on their mistakes?
“It’s an opportunity for professional development — to set goals and work toward them. But also, it’s really hard to know your own performance until you have some time to dig into the data, so one of my recommendations is to let the techs use the tool on their own to see their own data, review it, and have some self-correction and self-awareness that way. There’s also an opportunity to do buddy systems for teams to help each other — for example, I may be really good at one exam type, but struggle in another. Colleagues can work together to elevate their skills.”
What does that mean for patients? How can the X-ray Quality App make their lives easier?
“We have measured, at our beta sites, that on average it takes about seven minutes for exams, while with a reject, on average it takes about 10 minutes. So as a patient, your exam is going to be shorter and more comfortable. But also, if there have been three or four rejects earlier that day, it can slow things down in the department and extend the wait time for patients because things fall off schedule. There’s also dose to think about. X-rays do have a low dose already, but patients are rightly concerned about exposure to radiation, so they’d likely want to know what institutions are doing to reduce that risk, especially for pediatric patients.”
What are the benefits of this technology for pediatric patients?
“The AAPM recommends a lower reject rate for pediatrics because of the sensitivity around radiation. That’s especially for kids with chronic diseases who have to have a lot of X-rays. To think that their radiation would double because of a reject would be disheartening. You have really scared kids and parents who are difficult to keep calm and still. I think that technologists who work with pediatric patients really want to do everything they can to improve the experience for the patient and their families. Technologies to reduce reject rates are maybe one small way to help that.”
What’s the future of this technology?
"We’re working on the second release of this product, where we are expanding device compatibility, including exploring vendor-neutral capability and expanding features to add quality assurance tasks outside of reject analysis — such as exposure index and deviation index. Leveraging the same analytics platform, GE Healthcare is also building new multi-modality, vendor-neutral analytics offerings like Imaging Insights to expand into analytics revealing operational benefits, such as tracking whether you’re able to have X number more exams throughout the day because you’ve reduced the time on the machine, leading to reduction in cost. All of these additional analytic tools could give you the vision of throughput and scheduling to improve processes."
1. Using Small Data Analytics for Real-Time Analysis of Radiology Practices. Unleashing the Power of Small Data in Healthcare Organizations. GE Healthcare. http://content.gehealthcare.com/-/media/b930d54e1858409881b87d422e7b9402.pdf. Accessed Sep. 24, 2018.
2. X-ray Quality Application. GE Healthcare. http://content.gehealthcare.com//-/jssmedia/c2399c841c8d42229e294c711977bafb.pdf. Accessed Sep. 24, 2018.
3. Ongoing Quality Control in Digital Radiography: Report of AAPM Imaging Physics Committee Task Group 151. American Association of Physicists in Medicine. https://www.aapm.org/pubs/reports/RPT_151.pdf. Accessed Sep. 24, 2018.
5. X-ray: Quality Application. GE Healthcare. http://content.gehealthcare.com/en/products/x-ray-quality-application?cookie=906&account=36813&. Accessed Sep. 24, 2018.