Right now, the allure of AI is irresistible to every industry — healthcare is no exception. But there is little sense in acquiring a shiny new engine to power a vehicle you do not have.
When it comes to practical use of powerful new technologies like AI, businesses in the healthcare sector are often burdened with a “modernization gap” that squarely sets desired benefits beyond reach.
Mind the gap
While not unique to healthcare, and more pronounced in certain areas, healthcare as an industry generally lags behind others in terms of meeting aspirational technology goals — whether it’s moving to the cloud, going digital, or just trying to keep up in the marketplace.
The reasons for this are legion, culturally ingrained, and well documented: legacy IT infrastructure and architectures, tight compliance and data security strictures, not to mention day-to-day functional realities that are very real matters of life or death. And of course, there’s the issue of finance and challenging operating margins that can curtail IT investment. Resistance to change and slow technological adaptation hasn’t necessarily been a choice in healthcare, so much as a bid for survival.
The proliferating crop of AI capabilities promise to aid in such survival efforts when organizations put them to good use. But the business must be equipped to enable such use cases. They have to bridge that gap.
Healthcare organizations that have already started modernizing their IT and data infrastructure are hoping AI can accelerate that journey. If they haven’t already started modernizing, they’re generally unsure of where to even begin to take the next step. And just having a better mousetrap is not the goal. The aim is to actually impact patient lives, enhance capability and experience, reinvigorate the business, enable new models, and innovate into a better future.
Ironically, achieving that goal actually requires adopting something of a post-technology mindset. It’s not about the IT, it’s about the outcomes — and how you secure return on investment while achieving them. A colleague recently told me about consulting with executives at a well-regarded hospital system and determining that they face some serious self-imposed obstacles to modernization. They were consistently preoccupied with tools and systems at the expense of what they actually wanted all that technology to do. His exact quote was, “tools are for fools.”
AI strategy requires data strategy
If healthcare organizations want to use AI, they need a strategy. AI is fueled by data, so a huge part of any AI strategy is data strategy. Data strategy requires understanding all data feeds (internal and external) — where it goes and how it is shared — and ensuring secure, privacy-compliant data access and governance. How you use your data is the primary concern. The systems and tools are secondary.
Data drives everything.
Preferably, your strategy will include a pragmatic roadmap for quickly pivoting to use cases and outcomes. And without exception, all of this requires a shift in mindset about capability, return on investment, and value.
To start with, the concepts of capital expenditure versus operational expenditure need to be reexamined under a modern lens. For example, most health organizations no longer need to buy servers for the basement that have enough capacity for the proverbial church on Easter Sunday. Cloud models and the vast majority of modern tools are both “pay as you go” and “pay only for what you use.” This changes the traditional math.
Additionally, health data strategy needs to include determining where to build versus buy and with which vendors and third-party services. This will involve reexamining those secondary systems and tools that comprise your data stack.
A lot of hospitals, for example, may presume that their EHR’s modernization efforts will supply the bridge for their own modernization gap, but that is highly unlikely. All the data used in a hospital’s complete suite of necessary business and upstream workflow applications — as well as a patient’s continuum of care data — is never going to be housed in an EHR. Epic is a major player in the healthcare space, but, with no disrespect, is not a leading analytics company. The most impressive new capabilities for a tremendous host of purposes are going to come from a much deeper pool of innovators. Microsoft isn’t throwing money at OpenAI for no reason.
For that matter, do you want to lock yourself into, say, Microsoft-bundled technology? Do you need hybrid or multi-cloud options? What if Google or Amazon or someone else offers functionality that more affordably serves your needs, or develops the next big thing in the AI space that you can plug into rather than have to build yourself? Should your organization be building any such tools in house? And, if so, for what purpose, at what cost, and how long before it can start delivering value? There are vast resources available from parallel data-driven industries and plenty of experienced third-party experts to tap for guidance, so no one has to go it alone to answer these questions. But they have to be framed directionally for your business — where it stands now and where it needs to go — and they have to be asked.
No matter what, healthcare data strategy has to incorporate the ability to evolve.
This applies to every facet of the healthcare sector, where we’ve all been doing more with less while trying to navigate momentous disruptions, be it Covid or AI. That goes for hospitals and health systems dealing with surges and staff burnout and cyberthreats as well as revolutionary breakthroughs in treatments, therapies, and preventions. That goes for drug companies rethinking how they develop new pharmaceuticals and biologics. That goes for insurers, and providers, and payor-providers restructuring for value-based care, managing chronic disease, and keeping people well.
It’s a matter of survival for us all.
Photo: zhuweiyi49, Getty Images
