Testing the Enterprise AI Waters: Healthcare’s Great Tech Transformation is Underway

Updated on December 27, 2021
Medicine doctor and stethoscope in hand touching icon medical network connection with modern virtual screen interface, medical technology network concept
Jeff Sorenson copy

Some leading hospitals have already started experimenting with deep-learning based enterprise AI that applies most directly to medical imaging. In a few years, we will be taking many more uses of AI for granted. How does the healthcare industry get from here to there? 

By Jeff Sorenson 

In 2020, nobody has healthcare AI completely figured out. But I’m certain that deep-learning tools available today will transform our field within the next five years, improving monitoring, detection, and ultimately treatment outcomes.

We are seeing certain health systems emerge as leaders, like Northwell Health in New York state 

AI goes into the storm

Northwell Health is New York’s largest health system. The organization serves multiple communities and has been at the world’s COVID-19 epicenter. Fortunately, the health system has been investing heavily in data operations and developing prowess with artificial intelligence. 

For instance, the system’s cardiac imaging team at Lenox Hill Hospital had been developing imaging automation and AI capabilities for more than 15 years. They were driven to go faster and faster by the health system scale, with the volume of imaging work accelerating day by day. Physicians pushed for a system to process access the data quickly to support the pace of everyday work. 

In addition to speed of care, another goal has been to ensure care team collaboration can happen in real time.  A physician in an imaging center should be able to talk to physicians anywhere in the health system with ready access to the right data for the conversation. Physicians reported this secure, real time capability as critical, especially when dealing with very complex cardiac disease. 

Going further, AI-driven automation has expanded to cardiac CTA imaging, calcium scoring, TAVI, mitral valve clipping, aortic surgery, congenital heart disease, and cardiac MRI evaluation.

As the current pandemic unfolded, the health system was able to work with a patient database featuring vitals, laboratory results, medications, and continuous data capture. The health system updated demographic information and used it to identify characteristics and track early clinical outcomes of more than 6,000 hospitalized COVID-19 patients in the New York City area. They used this agility to be among the first to recognize a correlation between COVID-19 and heart disease complications. 

With increasing health data, evidence-based prediction tools trained and validated properly and often can guide overwhelmed hospital frontlines and administrators to make informed decisions in a challenging time. With the best possible data and analytics connected in real time to medical imaging insights, the field of deep learning healthcare AI will be a key ally in rapid response innovation.

How to do it? 

In fields like radiology, cardiology, and oncology, much of the pioneering AI work is focused on health care imaging. But even in these fields, the vast majority of doctors aren’t using real deep-learning technology in their daily practice. That means, at least in part, that developers and SaaS companies have to increase the accessibility to these tools and reduce the barriers to adoption. Instead of continuing to focus on proving and refining the effectiveness to the point of perfection, more effort should be placed on seamless clinical workflow integration for physicians.

Hospitals and providers need to embrace systems that will support, store and analyze these enormous troves of medical data. It’s not a challenge health systems can take on alone. It will be born out of relationships between the health care and tech sectors. 

Returning to Northwell Health, they and others have developed venture arms to invest in and help drive AI and technology innovations drawn from the insights of everyday care. 

The day-to-day need 

We need to start by better identifying exact clinical needs. We can’t be designing in theory, we need to be designing where our products and services will be used: on the front lines of our medical systems, in hospitals and clinics and even in our homes. 

Think of AI not as a possible replacement for trained medical professionals; but rather a thousand sets of additional eyes — and a limitless memory — to expand our collective detection and diagnostic capabilities. 

AI will only be effective insofar as it empowers its users and informs clinical decisions in real-time. Hospitals won’t be wooed into AI through big ideas; it needs to start with a collaborative approach targeting specific pain points. AI is a technology with its own set of limitations and has often failed to replace entire jobs but has proven fantastically capable at completing entire tasks. For this reason, products that incorporate AI will generally be more helpful to physicians than products that don’t. 

This is true for large and small institutions alike. Ongoing efforts around interoperability and EMRs will open the way for smaller providers to draw on data from various health systems, while also contributing their own de-identified patient data to the AI training pool. 

If AI remains the domain of only the world’s largest research hospitals, it will never have the transformative power imagined at industry conferences or in Big Tech boardrooms. 

The real pioneers in health care AI are the frontline physicians that drive utilization because, more than anything, utilization will drive innovation in the years ahead. What AI really needs is utilization, which generates data that can lead to insights. It needs examples of what worked and what didn’t, both in the algorithms and how health care organizations deploy Enterprise AI. 

In turn, tech innovators will approach clinics offering purpose-built solutions, not big ideas built for a theoretical practice. 

There is already a tremendous amount of data generated in the course of people recieving healthcare, and these data demands are increasing exponentially. However, this data is not being properly labeled and stored in a way that is useful for training the next generation of AI solutions. AI can be used to solve this problem, too, by automating the very labeling necessary  to provide it the data needed to learn. Without utilizing AI this way, the data will soon become unmanageable for many health systems.

Turning data into AI-powered insights takes organizational discipline, as well as the right technology. This can be accomplished locally by the healthcare providers themselves, but if history is a guide, they would prefer out-of-the-box solutions from trusted technology partners that can increase the value of their data and use it to improve patient outcomes. 

14556571 1295515490473217 259386398988773604 o

The Editorial Team at Healthcare Business Today is made up of skilled healthcare writers and experts, led by our managing editor, Daniel Casciato, who has over 25 years of experience in healthcare writing. Since 1998, we have produced compelling and informative content for numerous publications, establishing ourselves as a trusted resource for health and wellness information. We offer readers access to fresh health, medicine, science, and technology developments and the latest in patient news, emphasizing how these developments affect our lives.