Earlier this year, a thought leadership paper from consulting firm Frost & Sullivan highlighted the principal reasons why a cloud-based enterprise imaging platform is critical for achieving improved clinical outcomes and delivering true patient-centric care.
The paper specifically analyzed the objectives of the global healthcare system, the inadequacies of the traditional picture archiving and communication system (PACS) and vendor neutral archive (VNA) in the current medical imaging environment, and the benefits of combining enterprise imaging with a unified VNA for enhanced clinical outcomes without increasing the cost of operations.
Following the release of the paper, Siddharth Saha, vice president of research, transformational health, at the firm, along with an executive at healthcare imaging solutions company Novarad— Harold Welch, vice president of technical solutions—discussed with Healthcare Innovation the key findings from the paper as it relates to imaging informatics and what the future holds as artificial intelligence and data analytics have begun to enter the picture. Below are excerpts of that discussion.
What were the most important takeaways from your research?
Welch: The industry is requiring that this new-age [imaging solution] be secure, very easy to use and portable. The conventional physician sitting at his or her desk with a PC on a six-year line is going away. We are moving more toward a distributed model. So in addition to consolidating and creating a patient-centric consolidated view, making it portable and consumable at the right time and place is now essential.
Saha: One of the things we have been seeing in radiology is that there has been a bit of a rethinking about the role of radiologists. This profession, which does a great job in clinical decision support, and everything else that radiology provides, ultimately is making a big difference in patient outcomes and in value-based care. But when you think about the enterprise imaging strategy, it is actually allowing for the use of so much information that is generated at the enterprise level. So [an enterprise imaging solution] is a solution that is helping the profession make that much more of a difference in clinical outcomes. Ten years ago, people thought of radiology as a service to the clinical process—you can go to a scan and take a look at the reports. But radiology is much more than that today. An enterprise imaging solution allows for the radiologist’s contribution to the decision making table; it’s actually enhancing the role of the profession.
What are your findings regarding the insufficiencies with traditional medical imaging solutions?
Welch: We have walked through hundreds of facilities and the common thread we have seen from different disciplines in the hospital is that it’s very difficult to get to the content I need when I need it. Out of necessity—or out of hospital preference—we have created silos of information. As a result of this, we have a self-inflicted wound; to find the information that is universally applicable for clinicians, I have to go to three different systems. And I have to also find out, is this the same patient that had a cancer study here, or a broken leg there, and putting that together can be time consuming and inaccurate.
It appears as if AI and data analytics are now becoming paramount in medical imaging. Can you explain this trend and specific use cases that stand out to you?
Saha: We have been tracking this for a few years; we go to every industry event, from the Radiology Society of North America (RSNA) conference to many others, and we have not only seen the startup ecosystem full of an enormous number of companies developing applications—for a body part, to support a given modality like an X-ray, CT scan, or ultrasound—but also for disease-specific capabilities. Despite all the question marks and bad press, and there are some fundamental questions about the role of the radiologist, we are seeing solid evidence in being able to shorten decision making times. So something that used to take 30 minutes to process, such as looking at blood flow from an MRI in a cardiac application, now takes two seconds to process that data. And there is also enough evidence in terms of the contribution this could make to the decision making pathways.
Overall, we are seeing a lot of positive contribution from AI an analytics in this space, and there is no stopping what they can do. But I would express a little bit of caution over having the right kind of data sets from which the system will learn. Machine learning will pick up from historical data so you have to feed it the right kind of data sets. Some of the leading companies are doing a great job by exposing the system, and it then learns, picks up the right pathways, and makes the right decisions.
Welch: There has been a mixed understanding between analytics and AI in this industry. Analytics is applying algorithms and natural data patterns to an existing data to better understand it. By simply using analytics, you can accomplish a huge amount of progress and you can be very proactive in patient care. Once the data set is clean and once you have a confidence level in that, you can use AI to feed this engine that improves and creates better outcomes over time. It will be a process; none of the radiologists are scared to death that they will be replaced by [IBM] Watson because Watsons does exactly what it is told, and it’s a great predictive model—but it does not replace the intuition and experience that a good radiologist has.
A group of European and North American radiology societies have recently issued a draft statement on the ethics of artificial intelligence. How would you address this statement?
Welch: The cumulative knowledge, strengths and talents of those who contribute to this is much stronger than any individual effort. I don’t believe [AI] will ever completely replace the human mind and the ability to perceive and finalize a diagnosis, but the cumulative knowledge is stronger than any individual.
What predictions can you offer for how this space will continue to evolve in the next 12 to 24 months?
Saha: Everyone has heard of precision medicine, and now we are starting to see a new phrase known as precision imaging. This is about having the right kinds of tests, the appropriateness of the tests and what you do with them, and then allowing for the use of all that data as well as collaboration. So it’s about bringing through all the different capabilities; the culminative power of what this profession can contribute to clinical decision making.
Welch: We are looking at predictive models through our R&D division, rather than waiting for a patient to exhibit a particular symptom. Apple just made an announcement that everyone can walk around with something on their wrists to give them an indication if they have some form of a heart problem. This is the tip of the iceberg from a technology perspective, and for data collection, it is our responsibility to collect it and use it wisely. The algorithms that will allow us to preemptively capture and address health concerns are coming very quickly.