"Don't Believe The Hype"

Artificial intelligence (AI) is riding a huge wave of hype in the healthcare industry. Everyone seems to be talking about the transformative potential of AI to cut costs, improve efficiency, boost diagnosis accuracy, and enhance patient outcomes. AI has captured the attention and imagination of healthcare providers, investors, Big Tech companies, and entrepreneurs.

However, it's important to view the hype around emerging technologies like AI with a critical eye. While the promises seem bright, there are still challenges and barriers when it comes to implementing AI in real-world healthcare settings. As with any new technology, it's wise to balance the excitement with pragmatism and evidence-based evaluation. This is where Gartner's "Hype Cycle" offers useful guidance. This analytical tool tracks the typical trajectory of emerging technologies from the "Peak of Inflated Expectations" down into the "Trough of Disillusionment" before coming back up to the "Plateau of Productivity." When it comes to AI in healthcare, there is no doubt we are still riding high on the inflated expectations wave.

Some deservedly caution that AI in healthcare may soon plunge into the trough of disillusionment before realizing its full potential. There are still obstacles related to data privacy, algorithmic accountability, regulatory approval, and physician trust in AI recommendations. Healthcare AI startups also face the stark reality check of whether their products actually improve health outcomes. Does this mean we should abandon hope in healthcare AI? Absolutely not. The hype exists because AI truly does have immense potential to transform areas like medical imaging, personalized treatment plans, patient monitoring, and automated documentation. But we must approach these use cases with clear eyes, focusing attention and investment on applications supported by peer-reviewed evidence and clinical validation.

Over the next 5-10 years, we will hopefully see AI slowly deliver on its promises to augment (not replace) clinicians to provide more accurate, efficient, and affordable healthcare for all. But we will get there not through hype — but through continuous real-world testing and staying grounded about the true near-term capabilities of the technology. If investors and companies keep these reasonable expectations in mind, AI can steadily make its way to the "Plateau of Productivity" without first suffering a plunge into disillusionment.

When it comes to putting money behind AI in healthcare, there is always a decision point for investors on whether to fund early or wait until the technology demonstrates greater maturity and real-world validation. Investing early allows you to claim a stake in what could become a transformational company. However, many of those bets will also fail or require many years of patience as the technology slowly claws its way up the Hype Cycle. Later stage investing certainly comes at higher valuations and reduced risk/reward ratios. But the odds of putting money behind an AI startup that actual improves clinical and financial outcomes is substantially higher.

Determining optimal timing for investing in new healthcare AI tools requires evaluating tradeoffs between being an early adopter and waiting for more advanced solutions. Becoming a pioneer allows you to help shape the technology based on your needs while potentially staking a competitive edge. However, it also means bearing substantial risk that the AI solution may not deliver on expected benefits or integrate smoothly across your infrastructure. There are major costs associated with changing workflows, retraining staff, and managing complex implementations too early.

The alternative path is to wait - allowing innovators and early adopters to test, refine, and regulatorily-approve the most promising AI solutions. This avoids the pitfalls of new, unproven technologies. But you also lose out on the competitive advantages of being a front-runner. Moreover, by the time you implement, you'll pay a premium price as vendors scale pricing to match mainstream demand.

Given these tradeoffs, prudent hospital CIOs often follow a "fast follower" approach. They run small pilot programs to evaluate emerging AI tools while building knowledge and capabilities. Then they wait for a tipping point of clinical validation and integration readiness before fully deploying an AI platform across their organization. This allows hospital systems to balance the benefits of gaining experience while avoiding major near term risks and costs. They give up the pioneers' edge but have reliable solutions ready to scale when the time is right. With healthcare AI still climbing the hype curve, this pragmatic fast follower mode often makes the most strategic and economic sense for major hospital buyers.

Over the past two decades, healthcare organizations have poured tremendous capital into next-generation platforms like electronic medical records (EMRs), data warehouses (EDWs), health information exchanges (HIEs), and most recently cloud infrastructure. Early proponents argued these technologies would drive progress through data-informed decision making, coordinated care, efficiency gains, and more. However, the industry has very little to show for these multi-billion dollar investments. Adoption of platforms like EMRs has certainly improved accessibility of health data. But costs keep rising without corresponding gains in patient outcomes or experience. Too often, new technologies simply layered complexity atop an already fragmented system without fundamentally improving care delivery.

Technologies like natural language processing and predictive analytics also promised to unlock insights from health data to enable personalized medicine. But most organizations struggle turning data into meaningful improvements. Just implementing complex systems required major tradeoffs in time, money, training, and workflows for overburdened health staffs. As we consider the next wave of promising healthcare technologies like AI, the lessons from this last era should temper our expectations while clarifying priorities. Data and tech are not silver bullets. Platforms alone will not cut costs or save lives without broader ecosystem change. This next generation of tools must focus relentlessly on enabling healthcare professionals to deliver better care, not just promising data-driven management mirages. Success will require an integrative approach where technologists and clinicians co-design solutions around patient-centric priorities - not siloed products that force complex workarounds.

Past misfires like EMR and CDS (Clinical Decision Support) implementations drained resources while leaving care inequality largely unaffected. As healthcare pursues high-tech aids like AI, investors and innovators must have clear eyes regarding this challenging history. By learning lessons from previous failed efforts, we can build a next generation of technologies focused on transparent, proven benefits for patients. If solutions don't directly link to better outcomes, improved clinician workflows, and higher satisfaction, buyers should view claims around AI and data-driven care with well-deserved skepticism.

There's no doubt healthcare needed major digital transformation and investments like EMRs to set the data liquidity foundation. Other consumer-focused industries digitized years before under pressure to enable ecommerce, improve customer experience, and leverage data analytics. Healthcare languished behind due to unique privacy constraints and lack of market pressures. EMR adoption forced this complex industry to begin catching up. In that sense, the last decade and a half of platform investments achieved a necessary end despite steep transition costs. Clinical data finally flows more fluidly so future innovations can realize productivity and quality gains through add-on decision support, predictive analytics, telehealth, and more. Much like digital infrastructure in banking, transportation and retail fundamentally improved how those industries function, so too will digitized health data and processes in the long run.

However, the way many healthcare organizations implemented those platforms helps explain limited returns so far. Too often EMR rollouts focused on mandates over user needs, failing to address clinical workflow integrations. Coupled with demanding data entry and subpar UI designs, this directly contributed to drops in physician productivity, bottlenecked operations, clinician burnout, and detracted from patient-centered experience.

So ultimately both perspectives hold truth. Heavy technology investments laid vital foundations to unlock future innovation. But poor change management and adoption strategies resulted in major short term productivity declines and engagement challenges. As healthcare adopts the next wave of AI and analytics, this mixed history offers an important reminder. Successful adoption equally depends on user-centric design and workflows as the technical capabilities themselves. By co-developing solutions with doctors, nurses and patients, future health technologies can avoid the pitfalls of previous implementations focused more on databases over end user needs.