Digital Health and the Direction of AI in Healthcare - Cover

Digital Health and the Direction of AI in Healthcare

Ahead of KLAS’ annual Digital Health Investment Symposium (DHIS), I recently spoke with experts at each of our sponsoring companies: Seth Kneller, managing director at TripleTree; Walt Breakell, executive VP at Marwood Group; Doug Greenberg, senior client partner at Korn Ferry; and Alya Sulaiman, partner at McDermott Will & Emery.

During our conversations, each expert had some great high-level thoughts on some of the most pressing topics in healthcare IT today. The focus of the conversations was digital health and their perspectives on the direction of AI.

What is the current state of digital health?

Seth Kneller: The term digital health can mean many different things when it comes to healthcare. From my perspective, some of the most interesting aspects of digital health are the capabilities that address rising costs, labor challenges, and care delivery inefficiencies. I continue to see strong market interest in companies that address these industry pain points, and I anticipate increased mergers and acquisitions in these areas.

Doug Greenberg: I agree that it is important to define what is meant by digital health, as the term is very broad and can be defined in many ways. Digital health can be used to improve operations, increase access, and drive both patient and clinician satisfaction. Additionally, digital health can be used to drive better health outcomes. Looking at digital health across all these areas, the industry is still in its infancy. A lot of organizations have purchased a variety of digital health solutions, but they have yet to realize the full value of these tools. Some organizations say they are very active in digital health, but when you look deeper, what they really mean is that they have purchased or are building some digital-led tools or solutions. From a big-picture standpoint, we are still very early in putting digital health solutions into practice to meaningfully change business outcomes.

Alya Sulaiman: I’ve been working in digital health for my entire career from the legal perspective. I’ve seen a dramatic shift in organizations’ willingness and enthusiasm to use data to inform decision-making and best practices throughout the healthcare ecosystem. That has largely been accommodated by a shift in the regulatory landscape with the 21st Century Cures Act and some of its regulations. We have moved from a world where some data sharing activities were sometimes permitted to a world where they are required. The resulting shift is taking a stronger and stronger foothold across the healthcare industry and especially in digital health.

KLAS has seen high energy in digital health. What areas are you most excited about, and how are these areas solving critical pain points in healthcare? 

Walt Breakell: We are finally moving toward a more seamless integration of health communication into our lives. Not that long ago, a senior citizen in Florida could possibly get numerous text messages a day from a pharmacy, primary care provider, oncologist, etc. Things are shifting to be a bit more streamlined. Shopping online from companies such as Amazon and Target is easy because they’ve created a great experience. Healthcare is commercial like anything else. It’s a service people need at the right price and in a convenient way. The healthcare industry is spending money on communicating effectively with people to educate them and get them comfortable so that they can make better decisions. We are streamlining the various disjointed ways that people interact with the healthcare system, and that gets me excited.

Seth Kneller: There is no doubt the COVID-19 pandemic ushered in a new era of tech-enabled capabilities across the healthcare industry. As I think about what’s ahead for digital health, I’m excited about the growth potential for tech-enabled solutions that make the healthcare system easier to use and engage with, including developments across three large sectors:

  • Clinical, including solutions for patient monitoring and the early diagnosis of conditions, imaging and AI-enabled image guidance, clinician burnout (e.g., physician assistants), and facilitating care outside the hospital (e.g., urgent care, ambulatory surgical centers, and retail)
  • Administrative, including companies that deliver real operational efficiencies (e.g., patient access, patient flow, and discharge), claims administration and automated medical coding, and data management and security
  • Drug discovery and development, including the opportunity to use AI as part of the drug development and commercialization process

Alya Sulaiman: It’s been fascinating to see general-purpose AI tools answer complex medical questions accurately and reliably with little specialized medical pretraining. Generative AI’s ability to engage with and reason through health prompts and questions to reach a decent and often sound outcome is exciting, even as several patient safety, legal, and regulatory considerations demand a rigorous and formalized review process that most organizations are not yet equipped for. Outside the direct patient care workflow, there are many high-impact, low-risk use cases to deploy AI tools in ways that make a meaningful difference for decreasing physician burnout, removing the burden of certain administrative tasks, and eventually helping improve patients’ ability to access care. I’m really excited about those use cases because they create lower-risk opportunities to experiment with AI in healthcare. From a legal perspective, those use cases are often more straightforward for me to counsel clients on how to develop and deploy AI responsibly.

Doug Greenberg: Everybody immediately looks to AI to help solve so many issues, and it will be fascinating to watch things over the next several years as AI ultimately helps us better manage illnesses and health risks and promote wellness. But before we can get to that future state, hospitals and health systems need to directly address fundamental operational inefficiencies, given the existing headwinds these organizations are facing. Further outsourcing, offshoring, embracing a cloud-first approach, and utilizing other advanced technologies will help healthcare organizations significantly reduce costs and optimize both clinical and nonclinical services, creating smaller and more agile and cost-efficient organizations and labor forces. At that point, we can shift our focus and capital to AI and its enormous potential, especially in areas such as harnessing AI to automate notes and medical records, increasing accuracy and helping to reduce clinician burnout.

Moving on to the world of AI, do you have any predictions for where AI will see the most growth, adoption, or validated outcomes in healthcare?

Seth Kneller: As I reflect on recent conversations with company leaders and investors, I think the potential for AI is profound. I believe AI companies will be successful if they can deliver solutions with a demonstrable impact. These solutions need to be actionable and deliver workflow-driven benefits and efficiencies. They should also have a significant impact on existing costs, particularly around labor, supply chain, and drug spending. Finally, the solutions should have highly complementary features and benefits and augment existing software systems and processes.

Walt Breakell: AI has been in healthcare on some level, whether that be ML or AI, for some time. It just depends on what segment of healthcare we’re addressing when we talk about AI’s impact. Years ago, academic studies with radiologists showed that ML and AI could improve efficiency and accuracy and even catch things that may have been missed because of human fatigue.

The clinical impact of AI is less interesting to me than AI’s ability to improve labor shortages in healthcare. At least in the early stages, AI has the potential to reduce labor that’s not clinically related, such as billing or coding. Some technologists are using AI to reduce labor or time. We’ll see more use of AI in billing, procurement, analysis of best prices, and the best utilization of a healthcare provider’s resources. AI may very well change less about diagnostic care and more about the delivery of healthcare and how people think about getting resources to the bedside.

Within organizations, people are organically going to start deploying and playing with AI tools. They’re going to figure out ways to make their jobs a little better. Things will bubble up, and there won’t be a hard break. AI is a tool just like the internet; it slowly becomes adapted into workflows. Like they do with any other innovation, people are going to slowly figure out how to use AI in their daily lives and then expand its use.

What should a healthcare provider looking at AI solutions be aware of? What does it take to be successful?

Alya Sulaiman: There are three questions I like to pose whenever I’m consulting an organization on an AI initiative or deployment.

  1. What does it mean for this AI integration to really work? Organizations need to get specific because oftentimes identifying intended outcomes and AI’s intended use can help us know what guardrails need to be in place for a responsible, safe deployment.
  2. How could this deployment fail and for whom? Answering “for whom” is important because these deployments could have an impact on more than just end users. AI can fail for different reasons with different effects on different people. For example, the deployment may impact individuals whose roles could be augmented or replaced by AI tools. The “for whom” question humanizes the potential risks associated with a particular tool. In a world where there seems to be a new framework for AI every week, it is critical to get down to the real impact of AI tools on human beings, whether those be patients, providers, or clinical workers.
  3. Does this AI integration introduce or amplify risks to the secure environment? I often interact with organizations that are just beginning to understand how data intensive certain AI deployments are. Depending on the terms associated with the underlying enterprise AI developer or model vendor that provides the tool, people may unwittingly engage in data disclosures, even temporarily, that are inconsistent with the expectations of their privacy and security program.

Seth Kneller: Providers need to keep in mind the hype factor, meaning they need to pay attention to which solutions have traction and which companies have something in production that generates value using “real” AI. There’s always the need to be prudent and understand what the solution does—to know what a company does really well and whether that aligns with your needs. Finally, I would say to think of AI as a smart assistant or something that augments and enhances an existing process or function. Organizations should think about the handoff from technology to humans. They should understand where the technology ends and identify where they still need people to manage things.

On the flip side, what does it take for a company with a healthcare IT solution, AI or otherwise, to be successful in this space?

Doug Greenberg: Successful organizations follow basic project management principles for change management and implementing new tools. They understand how clinicians and nonclinicians are going to effectively utilize new technologies and how those technologies may change workflows, and, ultimately, how people do their jobs. AI will only be as strong as its facilitator, and all organizations will need the right teams and support to drive its effectiveness. There’s still a large change management component that will need to be put into place to deliver the desired outcomes.

Seth Kneller: Large healthcare acquirers are placing high value on technology companies that deliver meaningful competitive differentiation and revenue or cost synergies while delivering value in a real, quantifiable way. Hospitals are under significant financial strain right now, and that means today’s environment is conservative and hungry for proven solutions. Hospitals want solutions that address key pain points and deliver a real ROI, such as physician satisfaction, improved patient access and scheduling, increased reimbursement, and reduced costs.

Alya Sulaiman: Success in this space requires a crystal-clear vision of how AI fits in with or enhances strategic priorities. It is not productive to be excited about AI for AI’s sake; be excited about what it can do for patients, your customers, and your partners. Invest in AI governance—a dedicated, centralized function to support the responsible assessment, incubation, and implementation of AI in ways that align with the organization’s risk tolerance and principles.


Want to learn even more from investors, partners, providers, and vendors? This is just a glimpse of the content that will be available at KLAS’ upcoming 2023 DHIS. For more information, please visit the events page on our website.

Photo credit: Kay A/, Adobe Stock