Top Health IT Analytics Predictions, Priorities for This Year

January 30, 2023 - As health systems make plans to address challenges caused or worsened by the COVID-19 pandemic, such as efficiency and patient access to care, many are turning to health IT tools such as data analytics and artificial intelligence (AI).

However, recent questions around ethical healthcare AI — including how health systems can implement it, the role of collaborative efforts, and how guidelines from various stakeholders like the FDA, the White House, and the Coalition for Health AI will influence policy and regulation — have presented healthcare organizations with additional factors to consider before moving forward with their analytics priorities.

Leaders from McKinsey, Deloitte, and Carnegie Mellon University discussed their AI and analytics predictions for 2023 with HealthITAnalytics, including what healthcare stakeholders should prioritize heading into the new year.

IDENTIFYING VALUE AND JUSTIFYING INVESTMENT

Many of the trends seen in 2022 and priorities for 2023 are centered around identifying the value of a data analytics or AI initiative and getting the investment needed to support those efforts.

“At an aggregate industrial level, there isn't yet a full appreciation and conviction on how much value [data analytics and AI] can really drive,” explained Venkat Inumella, partner at McKinsey & Company and leader of the company’s Healthcare Systems & Services Practice, in a Zoom interview. “Obviously, by now, every provider executive has heard [those] words, [but now] they want to do something in it. But if you don't clearly see a path to value, the investments don't follow.”

The amount of investment that health systems are making in these technologies is small relative to those of similar industries, but this is changing, Inumella noted.

These changes are particularly evident in the realm of AI.

“If you look at [the] last two, three years, the COVID-19 pandemic created the right set of conditions for digital transformation across the healthcare industry. AI is one of the biggest areas of investments that we have seen for that period, and healthcare organizations are using AI to improve the efficiency of internal processes,” said Kumar Chebrolu, healthcare data and artificial intelligence practice leader at Deloitte, in a Zoom interview.

Many of these internal processes have added tangible value in nonclinical areas, which have a lower threshold for risk tolerance, Inumella noted. Some examples include revenue cycle management, claims management and denial prediction, and staff scheduling.

However, there are examples of AI and data analytics adding value to clinical areas as well, such as improving clinical workflows, he noted.

James F. Jordan, distinguished service professor of healthcare and biotechnology management at Carnegie Mellon University and CEO of StraTactic, a life sciences, health IT, and healthcare delivery consulting firm, said that the COVID-19 pandemic has also highlighted the value of applying data analytics to population health and imaging analytics for telehealth.

Despite this, demonstrating how these technologies can add concrete value within a particular organization can be difficult.

“I think as we enter 2023, when you talk to the researchers in [the data analytics and AI] space that are starting to have a couple years of experience working with people in healthcare, they're recognizing that there's a gap between proof of concept and production,” Jordan said in a Zoom interview. “They can take these [technologies] and then go to Hospital A, and they can run it through radiology, and it works perfectly. Then they take it to Hospital B, whose information management systems are just slightly different, and it doesn't work. Yet I can take a radiologist from those two hospitals and move them around, and they would draw the same conclusion.”

He likens this recognition to the analogy of building a car, stating, “You can design a perfect engine, but that doesn't mean you have a car that has perfect transportation. That's one of the issues that we have [in this area], this gap between proof of concept and production.”

Jordan predicted that there will be more investment by health systems in these technologies moving forward, but these investments will most likely differ significantly.

“I think what we're going to see is the market is going to split where I think the question gets broken down into the larger players versus the smaller players,” Jordan explained. “I think these [larger players] have the scale and the size to be able to make the investment and get a return. And because of their scale and size, they make a lot more investment in standardizing the process of care.”

Inumella echoed this, indicating that he expects more investments in the core capabilities needed to launch data analytics and AI initiatives.

“What I expect and what I hope for is that the stable momentum continues. I think that we'll continue to see more systems set up teams, invest in the core capabilities that they have, and so on. And a lot more of them [will be] going after these use cases, if you will, that we talked about, where there is clear proven value... I think we'll continue to see more excitement and investment on the diagnostics front, the digital diagnostics front because as we chip away at this, we get closer and closer to that tipping point where it becomes a real thing and can [create] improvement,” he stated.

Inumella and Jordan recommended that health systems focus on finding tangible ways to justify investment in these technologies.

“What I hope we'll see much more of is systems investing behind the core capabilities that are still lacking in a lot of places,” Inumella said.

“It's easy to chase after some use case or value because you're talking to a startup or you talk to another system that did something right,” he added. “But in the ultimate analysis for the long run, you have to invest in the basics. So, that just means getting your data governance in place, investing in the basic cloud infrastructure, recruiting and building the necessary talent.”

Focusing on the data analytics side more so than AI has significant potential to quickly demonstrate clear value before moving toward more experimental and advanced technologies, Inumella and Jordan noted.

However, health systems must also continue AI-focused innovation to remain competitive in the future healthcare landscape.

“But all of this said, I think I would certainly have at least a small corner of my AI operation that's focusing on innovation, especially innovation that's unique to your system,” Inumella said. “Maybe you're operating in a market that has particularly unique challenges, or maybe you serve a demographic that can benefit from AI-based solutions differently than others. So, I mean, if you had a hundred dollars, put $10 towards this last pocket, but spend $90 on those first three is what I would say.”

DEVELOPING AN ENTERPRISE-WIDE STRATEGY

Identifying value and investing in the basics are necessary building blocks for developing an enterprise-wide strategy and scaling initiatives related to data analytics and AI.

“If I were to [make suggestions] to a provider or executive, I think the first thing, given the relative maturity of this industry, the first thing is to be clear about what you want out of this,” said Inumella. “I think the systems that do best have a bold aspiration, but they have a clear aspiration.”

He illustrated this point by highlighting an example he came across in his own work.

“A system that I had worked with said, ‘in the next three years, we want to create $200 million of value from analytics and AI. It doesn't matter which areas it comes from, but that's our aspiration. And now let's figure out how we do it.’ What that creates is an anchor point and kind of filter that allows you to de-prioritize things that are just shiny objects and go after things that are more prevalent,” he explained.

Chebrolu highlighted the recent trend of chatbots and generative AI being leveraged in healthcare and similar industries as potential aspects of an automation-based strategy for health systems.

“If you're looking at healthcare and AI , the three things that we always talk about are better access to care, quality of care, and affordability… [I would add] a fourth one — enhancing the human experience,” Chebrolu said.

“What's happening with affordability is when you have that human experience, you need a lot of people to support, provide, and consider these kinds of services for members or patients,” he continued. “Only some people can afford the effort of that particular consumer kind of service. How do you scale that? You have to automate it, and you have to deploy AI solutions, which will improve the experience.”

Enhancing the human experience is also crucial for an enterprise-wide strategy that benefits clinicians as well as patients, Jordan suggested.

“I think [health system] leadership has to really sit down and figure out what the master plan is,” he stated. “I [once] worked for a nonprofit where we would help companies, and sometimes we would have these executives and residents that came from industry, where they were the CEOs and the stars. And I used to always say, ‘You're not the star; you’re the producer of the director at best.’ And so I think we need to turn the tide from finance and administration driving behaviors and saying [to clinicians], ‘this is all you're getting,’ to some sort of collaboration where we get the clinical decision-makers back involved and start piecing together where the value-added activities are and have a master plan for being able to bring the humanity back to healthcare."

If enterprise-wide strategies do not consider the needs of the physician, fewer people will consider joining the healthcare industry, Jordan added.

IMPLEMENTING HEALTHCARE DATA GOVERNANCE

Even if a health system has a great enterprise-wide strategy that puts clinical decision-makers back at the heart of the data analytics and AI efforts, these plans won’t get very far without healthcare data governance, which refers to the people, processes, and systems used to manage data.

“Scaling comes with more data availability for testing, whether the solution is working as expected or not. That means you need a lot of datasets,” Chebrolu explained.

“Availability of data is very important in healthcare,” he continued. “The biggest thing is that the data availability is really, really a challenge. Without data interoperability among multiple organizations, you can’t look at the data that you have and [use] the AI. Again, when we talk about AI, we are talking about doing any kind of prediction, or recommendation, or forecasting, anything like that. In order to do that, you need all these data elements from various places.”

Data availability is one of many challenges, however, according to Jordan.

“I think when we look at our stakeholders, and what they're facing right now [in terms of challenges], I would say that on the data analytics side of things, it's the data quality and the standardization. They want to get things to talk together, and some of these data elements are in silos, and there's an inability to integrate them,” Jordan said.

This is especially challenging on the AI side because healthcare, unlike the technology industry, lacks a significant amount of data surrounding human behavior. The lack of data will be one of the biggest challenges in the short term, he explained.

Providers can take heart, though, as these challenges may not be as limiting as they appear, noted Inumella.

“You'll hear a lot about [this] if you talk to providers… this notion of data, which is: it's messy, it's not clean, it's fragmented. I think that's true. All of that is true. However, it's not as much of a roadblock as one might imagine. There's a lot that you can do with the data we have today that we still haven't done, I would say. But [data] continues to be the challenge, especially as we go further ahead,” he stated.

BUILDING TECHNOLOGY INFRASTRUCTURE AND ATTRACTING TALENT

These challenges in data governance and enterprise-wide strategy are often exacerbated by a lack of infrastructure and talent.

“Talent has been a continuous challenge,” Chebrolu noted. “For the last, I would say two to four years, a lot of organizations started talking about AI: ‘Let's do a small proof of concept and do it in a particular functional area,’ and they're trying to scale it to the enterprise-wide strategy, basically. In order to scale to the enterprise AI strategy, there are so many other things that you need to consider… you need talent at every level, from leadership to the data scientist, to the data analyst, to support from the business, and all the consumers, and everyone.”

Talent is not just lacking for healthcare AI and data analytics initiatives, but across health IT, Inumella added.

“[Healthcare is] an industry that historically is not deep on technical talent, not just in AI and data, but even in a broader IT, so that reflects in data analytics as well,” he explained. “That has changed, that is changing now, especially as some of the larger systems are poaching from other industries; they're grooming their own talent. But it's still very much a bit tricky, and I think that's made [it] harder in this industry because many systems are not exactly in talent hotspots.”

Building the appropriate IT infrastructure and cloud architecture are also key to the success of any data analytics and AI efforts in healthcare, as this infrastructure works in harmony with both data governance and human capital to bolster these efforts.

“I think some of the challenges that we have seen, and we'll be seeing in 2023, are around technology architecture and cloud adoption,” Chebrolu stated. “In order to do data analytics and AI, the data availability and data are very important, and a majority of healthcare organizations are going through this technology transformation, going toward a digital-first, AI-first kind of mentality. When they're doing that, they need to move all the data into a place, into the cloud kind of architecture, where they can use the data in the right way.”

“[Health systems] need to have the cloud infrastructure, [they] need to have a feature bank, [they] need to be able to consume that data to build these AI solutions,” he added.

Chebrolu suggested that health systems moving into 2023 should continue to grow their talent or to look for new talent to help lead data analytics and AI initiatives.

In fact, many health systems are already focusing on this, Inumella pointed out.

“I would say data analytics with provider systems is going or has gone mainstream. So, even compared to, say, two years ago… the number of systems that have chief analytics officers or chief data officers has really exploded. That's just not something that we had previously, which is great, and they have teams supporting them,” he explained.

Overall, the experts agreed that by focusing on value, strategy, data, infrastructure, and talent, health systems could take important steps to build or improve their data analytics and AI efforts, which have the potential to bolster clinical and nonclinical goals moving into 2023.