News center

How next-generation real-time data analytics can drive agile radiology departments

Why being simply ‘data-aware’ is no longer enough

Chris Meenan
Nov 12, 2019 - reading time 8 mins

Head of PerformanceBridge, Royal Philips


With a background in Imaging Informatics, Chris Meenan is Head of PerformanceBridge at Philips. Previously, he was CEO of Analytical Informatics, Inc., and has a background in the design, implementation and support of innovative technologies in both commercial and clinical environments.

Radiology administrators and other clinical leaders are under intense pressure to drive operational efficiency across organizational boundaries. Limited access to information can be one of the biggest challenges to making better, faster decisions in order to achieve short- and long-term operational cost savings in healthcare. To do that, radiology administrators must move beyond simple reporting – that provides a limited view of what has already happened – to next-generation data analytics that provide a broader view that can help them predict what will likely happen.

 

Recently, I spoke with AHRA (Association of Medical Imaging Management) president Chris Tomlinson about challenges and strategies for becoming truly ‘data-driven’ in order to optimize radiology practices and drive operational efficiencies in imaging.

Chris Meenan, Philips: In the recent survey we did with over 100 AHRA members, 98% of them said that, with reimbursement increasingly tied to metrics, they feel pressure to drive efficiency across organizational boundaries. And, 91% of respondents believed that better access to data would positively change the way they run their practice overall and tie into larger organizational goals or value-based care. That’s great validation of the need to help radiology departments better utilize real-time data analytics. What are your thoughts on this?

 

AHRA President, Chris Tomlinson: You’re right there is tremendous pressure to drive efficiency across organization boundaries but I believe it’s for a different reason. In the larger ecosystem of a hospital, there has been a massive investment of IT and financial resources into Electronic Medical Records (EMRs). This might make you think – wow, that will help radiology leaders better leverage data! But actually, what’s happened is that investments in that radiology infrastructure, whether that’s PACS or analytics, hasn’t kept up with it as a result. 

 

Now, for many hospitals, there is a consolidation of IT resources into a central informatics or IT group that is focused on large data warehouses for the masses. Consequently, radiology-centric data that you really need is often very difficult to get or not detailed enough to use. Often, you only get a piece of the data you need from your Radiology Information System (RIS) and sometimes it’s not even clean data. That is not enough to drive efficiency and that is where the pressure comes in for radiology administrators. With an enterprise focus on EMRs, it’s harder to mine our imaging data at the modality or machine level, but we need that sequence level detailed data, in order to run our radiology practices more effectively. 

 

Chris Meenan, Philips: In the AHRA survey, the majority of respondents (66.96%) said their decision making was based on historical data (spreadsheet reports from the past month or more). This confirms what we’ve seen that the typical process of data analysis in radiology is limited in scope and very much manually driven. As a result, reporting is very labor and time intensive, not standardized, costly, and most importantly – delayed. Delayed reporting doesn’t make radiology departments agile. What’s your experience on using historical versus real-time data? 

 

AHRA President, Chris Tomlinson: In my experience as a radiology administrator, to improve efficiency in a radiology department you need to reduce variability and improve turnaround time and, to do that, we really need modality-level or machine-level data. By that I mean, we need to have data on things like: when a patient entered the room; or how long it took to set up the scan; and, how much time was there between sequences.  

 

If I can squeeze 5-10 minutes out of each MRI sequence and reduce variability – that’s huge – but that’s based on historical data. Historical data helps you find out where your problems are but then, how do you recover those operational issues all the time? How do you repair the system? What radiology administrators need is a way to make it easier to see it happening in real-time, and make adjustments for it and get back on course as quickly as possible. 

 

Most radiology leaders are more focused on the retrospective data than the real-time ‘repair’ data. However, if you can turn retrospective data into a useable, predictive format and couple it with the real-time ‘repair’ ability, that’s really where you can make a difference because you can re-allocate resources or move across scanners as needed, when needed. I believe that’s really where the opportunity is because it enables radiology leaders to run the department more strategically.

 

Chris Meenan, Philips: Exactly! Let’s talk a bit about ‘waste’ in imaging. Research suggests that 60-65% of the annual spend in radiology is in operations and that up to $10-12 billion of that is potential waste (such as wrong test, repeat exams, poor image quality, etc) [1]. We know that real-time data-driven practice management actually saves costs in various areas such as reducing the need for repeat imaging or rescans. And, it can even increase revenue by providing insights that help radiology administrators better understand which types of services are most needed for a hospital’s particular patient population. Interestingly, the results of our AHRA survey show that more than half (54.6%) of respondents said cost was the key obstacle for implementing real-time data-driven practice management strategies.  What are your thoughts on how can we best reduce ‘waste’ in imaging and support the adoption of data analytics solutions to achieve cost savings that can be reinvested in areas such as patient care? 

 

AHRA President, Chris Tomlinson: Most healthcare organizations want enterprise-class solutions so the cost of any radiology-centric data management solution is an obstacle, unless it aligns with the broader clinical care themes such as Length-of-Stay (LOS) or turnaround time that align with care pathways that resonate with CIOs. In terms of ‘waste’ in imaging, I see that a lot of the spend in imaging is due to over ordering or inappropriate imaging studies. Right now, if someone orders an MRI of the brain and then someone orders a CTA on top of that – I would argue that defensive medicine is being used and we need to push back on that to reduce waste and annual spend in radiology. 

 

As a radiology administrator, if I have the ability from an imaging perspective to readily look at data or metrics and see how many inappropriate studies were ordered and how much of the time are we able to redirect orders to reduce waste or cost – that’s very valuable insight. It enables me to look beyond the data and optimize by knowing how many studies can we move from in-patient to out-patient to reduce cost. As an example, I would say LOS is one really good metric that can be used and one that radiology can really impact positively. Radiology can see things from upstream systems and push back or help prioritize next steps for patients which correlates with LOS and impacts cost of care overall.

Chris Meenan, Philips: It was interesting to me that one of the most important areas for radiology practices to address was tracking imaging system utilization (31.25%) and the second was improving patient scheduling backlogs such as wait times and no-show prediction rates (24.11%). These two challenges often go hand-in-hand as the lost revenue opportunities as well as sub-optimal modality utilization for radiology departments as a result of patient no-shows can be significant.  We talked with Dr. Puneet Bhargava at the University of Washington in one of our recent podcasts about what data can tell us and the surprising things he learned on the impact of patient no-shows. What did you think of some of Dr. Puneet’s comments in the podcast in particular on how we can use data analytics in general to help predict no-show rates?

 

AHRA President, Chris Tomlinson: At the end of the day, it’s not just doing the right thing by the patient, which we are all of course focused on, but it’s also to impact imaging outcomes in a positive way. What’s really compelling about using data analytics is that, in some cases, there could be some negative health consequences or outcomes, if imaging isn’t delivered on time. So, using data to find ways to ensure greater compliance and avoid missed care opportunities for patients, as Dr. Puneet was able to do, is critical as it maximizes the chances of better outcomes and healthier people. It’s just another example of how real-time data analytics can help us recover, adapt and optimize, not just for efficiency’s sake or cost savings purposes but for better patient care as well.

 

Chris Meenan, Philips: One thing that I find really exciting is how artificial intelligence (AI) will help enable data analytics in the future especially in areas such as missed medical appointments that we just talked about, where the number of variables impacting patient no-shows can be significant. Individually, these variables or predictors are not enough to effectively predict patient no-show rates but collectively they can be very informative and predict how likely patients will show up for an imaging exam. This is where data analytics tools, predictive modeling and AI can be most helpful in helping radiology departments become agile in other areas as well. I was happy to see that in the AHRA survey 94% of respondents said that transforming their radiology practice from being simply 'data aware' to becoming 'data driven' was either 'important' or 'very important'.

 

AHRA President, Chris Tomlinson: It is very important and I agree with Philips’ approach of moving away from historical data and towards leveraging real-time data insight to drive better reaction to systems and dashboards as that is more transformative. To really become ‘agile’ you must be able to see it happening in real-time and be able to re-act and re-optimize based on real-time data and learn from historical data that is in a useable, predictive format. That combination of both predictive analytics and real-time optimization is powerful because it enables a more adaptive or holistic reaction to systems. 

 

The other thing we can begin to use AI and analytics for is to look at the care continuum for the patient. I’m hoping we can utilize AI to help us model care pathways even better and answer questions such as: how can we reduce variability; or reduce tests; or, most importantly, what’s the best test and the right pathway to move this patient through the care continuum most effectively? Going forward, I see predictive analytics will help us better map to patient needs with the most efficient and cost-effective use of our resources for continuous improvement.  

 

For example, if there are certain AI algorithms that can help us predict the likelihood of risk for patients that would be very beneficial. By that I mean that if we can use AI to ascertain that a particular image finding for a patient has a high likelihood of risk and take action to move it up the list, or send it to the right specialist instead of a generalist – because our predictive analytics has flagged it, not diagnosed it but flagged it with predicted outcomes – that is important. It would have a huge impact in the radiology domain in terms of turnaround time, and getting the right study to the right radiologist at the right time. 

 

Radiology leaders have to make the case to C-suite leaders that investments in radiology-centric data management solutions that help us get the insights we need is critical to driving efficiency and reducing cost, variability or meeting key value-based care reimbursement metrics. We have to make sure that we are able to leverage the data from enterprise EMR systems so that imaging’s potential impact doesn’t get lost or minimized as we shift toward value-based care. It’s true that being simply ‘data-aware’ is no longer enough; the real opportunities for growth and value for imaging is when we become data-driven.

 

[1] Peer60; Unnecessary Imaging.

About Chris Tomlinson

Chris Tomlinson

President Association of Medical Imaging Management (AHRA)

Chris Tomlinson is the Enterprise Vice President of Radiology/Imaging, Clinical Lab & Pathology, Emergency & Hospital Medicine Service Lines at Jefferson Health in Philadelphia and is also the AHRA 2019-2020 President.