Paddy Padmanabhan, MBA

home

authors

paddy-padmanabhan-mba

The steady drumbeat of optimistic forecasts for artificial intelligence (AI) adoption across sectors has continued unabated this year. Analyst firm Frost & Sullivan fired the latest round with a report predicting that AI and cognitive computing will 

 for the healthcare industry by 2025. The report checks off all the usual suspects that will contribute to the savings: disease prediction, personalization of care pathways, claims processing and supply chain operations. Then comes the letdown: AI adoption in healthcare been slow due to “strategic and technological” challenges, and only 15-20 percent of end users have been actively using AI to drive real change in the way healthcare is delivered.

READ: The Rising Clamor for Explainable AI

The stubborn resistance to AI adoption in all industries is now a reality that everyone is coming to terms with. Consulting firm McKinsey says that only a small fraction of the technology’s potential value has been realized across all sectors — as little as 10 percent in some industries. To make sure that these downbeat forecasts don’t turn off potential clients, McKinsey helpfully points out that there is a wide and growing 

 between high and low adopters of AI, implying that the laggards need to catch up.

Something is worrisome about the whole picture.

First, let me address the vendor side of the picture and the current state of AI offerings. Unbridled marketing hype and inflated claims of “AI-enabled” solutions have confused many potential buyers and left others deeply skeptical. Black-box algorithms have created an “explainable AI” problem that I discussed in an earlier 

 here. Health systems that made early bets on promising technologies have pulled back sharply in the face of tepid, and often misleading, results. Healthcare is nothing if not conservative, and for two good reasons, if not more: Patient safety is more important than anything else, and there isn't much spare cash sitting around for expensive, cutting-edge technology experiments. Most health systems want results now, preferably yesterday.

I am not suggesting that AI has no takers or that there is no real progress. Our research indicates that many academic medical centers (AMCs) and some of our public health institutions, including the National Institutes of Health and the Centers for Disease Control and Prevention, are quietly investing in advanced analytics programs. Grant funding for AI projects is sustaining long-term research focused on finding cures for diseases. Public-private sector partnership is thriving, as are research partnerships between AMCs and big technology firms.

The story with commercial health systems is entirely different. Caught in the pincer grip of a rapidly changing reimbursement environment and an increasingly demanding healthcare consumer (not to mention serious nontraditional competitors waiting around the corner), these health systems are struggling to come to terms with changes in the marketplace and their own business and operating models.

Here are three key challenges for health systems looking to advance the ball with AI programs today.

1. Finding an organization model for the future.

, strategy guru Michael Porter of Harvard Business School and Thomas Lee of Press Ganey argue that healthcare has an org model problem, similar to that faced by industrial conglomerates in the mid-20th century. They argue that the way health systems are organized today, around physicians, is archaic. While the model works well in a fee-for-service reimbursement era, it will no longer work in a value-based care era with heightened consumerism.

The problem has its echoes in analytics efforts as well. The current org model makes health systems victims of data silos that refuse to come together, and when they do, houses them in standalone analytics divisions that very often do not report into the CIO of the organization. With multiple points of ownership for data assets, data sciences talent and computing infrastructure, even the most basic analytics programs struggle for traction at the enterprise level. AI requires an entirely different level of commitment at an organizational level and the establishment of specialist cross-functional teams with a common leadership.

AI requires vast amounts of data (and not insignificant amounts of financial commitments). AI programs may fail to take off simply because there aren’t enough data to work with to demonstrate real benefits. The widely publicized 

One way to address the issue of scale economies is to develop consortiums that can pool data for research and benefit from the results. The recent announcement by 41 health systems, led by Geisinger, Intermountain and others, to 

collaborate on reducing diagnostic errors

 could be an interesting model for the future. Similar initiatives are already in flight; Apple’s 

collaboration with several health systems

 across the country will deliver personalized health insights through the iPhone. An innovation collaboration by 17 health systems to improve Medicaid outcomes, led by Andy Slavitt, former Centers for Medicare & Medicaid Services administrator, and AVIA, is another.

Health systems are at a difficult crossroads today, forced to choose between continuing with the (currently profitable) status quo versus positioning for the future. Most health systems have made significant investments in their core systems of record, such as electronic health records (EHR), and are under pressure to optimize their use of the systems for improved returns.

But for programs such as advanced analytics and AI, EHR systems are not necessarily the right choice for advanced analytics. Other challenges may include partly depreciated investments in expensive data center environments built for the deployment of on-premise technologies in an era when the most advanced technologies are available almost exclusively on cloud infrastructure. The appetite for expensive AI programs with limited visibility to returns is even less when there is fierce competition among competing internal initiatives. Progressively moving workloads to the cloud and migrating on-premise software licenses to the cloud are some of the options available to continue to pursue innovation with AI.

Most health systems realize that a failure to adopt emerging technologies such as AI could doom them in the future as performance gaps increase not just among themselves, but more crucially between incumbents and the emerging crop of big tech firms such as Apple and Amazon, who are 

aggressively moving into traditional care delivery

 with owned and operated primary care clinics. For now, health systems have a choice: pursue competitive advantage and sustainability through advanced technology investments such as AI or risk becoming frogs in boiling water. I know the answer: Show some money — now.

The Big Unlock: Harnessing Data and Growing Digital Health Businesses in a Value-Based Care Era

, a healthcare growth and digital transformation advisory firm based in Chicago. You can follow him 

https://www.linkedin.com/in/paddypadmanabhan99/

Get the best insights in healthcare analytics 

How AI Is Shaking Up Healthcare, Beyond Diagnostics

Why Experts Are Developing AI, Medical Imaging Standards

Articles

There are three significant challenges that health systems need to address to overcome the artificial intelligence adoption gap.

Addressing AI's Adoption Gap Problem in Healthcare

There's no doubting the enthusiasm behind artificial intelligence- (AI) powered healthcare. A recent Accenture 

 found that 53 percent of healthcare executives expect to invest in AI, with 86 percent noting that their organizations are capitalizing on data to drive automated decision making at an “unprecedented scale.”

However, this survey also demonstrates the vulnerability that these executives are feeling. A stark majority has not yet invested in the ability to validate data sources across the most mission-critical systems, and 24 percent said they have already fallen victim to “adversarial” AI behaviors, including falsified location data and bot fraud. This raises alarm bells about the potential for abuse with AI in areas such as unnecessary profiling of patients or actual errors in diagnoses and treatment of medical conditions.

There is a growing realization that potential users of these advanced tools for clinical decisions want to see what’s inside the “black box” of AI to ensure that the technology is advancing care, instead of sowing mistrust on the part of patients and providers. In short, AI needs to explain itself.

The clamor for transparency into how an AI algorithm works is growing in the healthcare and patient communities. One concern relates to the potential for bias within the algorithm. As an example, bias can take the form of improper weighting of one demographic group over another, which could result in racial profiling and even push an inappropriate treatment option on the wrong group of patients.

Sheila Colclasure, chief data ethics officer at Acxiom, echoed this sentiment with me during a recent 

, as she explained why her organization now focuses on the idea of data ethics. “Data ethics is not just about minimum compliance but also about being ‘just’ and ‘fair,’” she said. It is essential to mandate data ethics at the engineering layer in today's cloud-based systems, she added. In the context of AI, this translates to ethics and transparency at the algorithmic layer, providing users of AI-based technologies with visibility into how such algorithms work.

Take the example of neural networks, which are systems that can process thousands or millions of binary data points to mimic human decision making and identify health issues based on patient data. This approach has shown promising results in imaging data for the detection of diabetic retinopathy. However, there is a recognition that unless a clinician knows how a neural network is arriving at a decision (such as predicting diabetic retinopathy through the analysis of images), any recommendations to the clinician will not be trusted. Without “explainability,” much of the potential of AI might be lost, as clinicians continue to rely on traditional methods of diagnosis and treatment.

A related rule of thumb: The higher the medical risk and complexity involved in a clinical decision, the less likely it is that experienced clinicians will defer to an algorithm to decide on their behalf. The widely publicized challenges of IBM's Watson Health platform in providing treatment recommendations for oncology is directly related to the lack of transparency around their cognitive techniques embedded in the platform.

The push for greater transparency is needed more than ever with uncertainty also building around regulatory and legal issues surrounding AI and other technologies used in healthcare. A recent 

 at automated systems used to perform surgery raised questions on just who is liable if these systems are not under the guidance of a physician and they make a mistake. Is it the physician using the device, or the manufacturer the one who should be held accountable? Moreover, how should regulations be crafted to address just such an issue?

Regulating (and Self-Regulating) AI in Software

 that is voluntary for companies developing software and medical devices. The goal is to protect the healthy growth of technologies such as AI and keep pace with fast, iterative designs, while still providing necessary oversight in the interest of consumer protection and patient safety. With the FDA’s recent announcements, this takes the form of a precertification program that utilizes a “trust-based approach to regulation for those vendors that have shown that they embrace a culture of safety and accountability.”

While still under development, so far the FDA is considering 12 

 to evaluate vendors including leadership, transparency, people, infrastructure and work environment, risk management, configuration management/change control, and so on. The plan is to tie key performance indicators to these categories to keep watch on an organization once they have been approved, but individual products would not need to be regulated.

Beyond the FDA, industry organizations have started taking the lead on defining guard rails for the use of AI in healthcare technology. The American Medical Association (AMA) passed a 

 on AI (its first) that starts to address the technology as “augmented intelligence.” The idea is to shift thinking on how to develop AI into something that enhances the clinical decision making of physicians but does not replace them. According to the AMA's May report, “Artificial intelligence constitutes a host of computational methods that produce systems that perform tasks normally requiring human intelligence. However, in healthcare, a more appropriate term is ‘augmented intelligence,’ reflecting the enhanced capabilities of human clinical decision making when coupled with these computational methods and systems.”

Whether it’s new terminology pushing for a rethink of what the goals of AI should be, a new regulatory framework, or new technology to make AI explainable to those who depend on it most, the reality is that transparent AI is the key to driving adoption levels higher. Moreover, it’s also critical if we want to protect public perceptions around AI, especially considering how much we’re going to need the technology if we’re going to fix our broken healthcare system. We are still in the early stages of opening up the algorithms for scrutiny—partly due to concerns around proprietary knowledge and intellectual property that most firms are reluctant to share. The growing clamor for explainable AI may be the best thing to happen to purveyors of AI-based technologies who face an increasing credibility problem.

Making AI transparent could head off public mistrust.

The Rising Clamor for Explainable AI

A friend who is the chief information officer of a large health system recently suffered a heart attack while competing in a race at a national level. As soon as he felt the heart attack coming on, he checked into the medical tent and was then expedited to a nearby hospital. At the hospital, the clinicians were ready and waiting for him. He recalled that as he lay on a bed and watched a stent go into his chest on a video monitor, he could feel his blood flow being restored and his muscles “fill and dance.” They discharged him the next day. I met him a couple of days later, and he told me he felt so good that he was already planning his next race. Considering the alternate scenarios that could have easily played out, my friend was very lucky.

It’s a simple fact today that much of our essential healthcare data resides in proprietary systems. Patients see multiple physicians, with data residing in multiple locations and platforms. There is no intelligent way today to adjudicate all of the data across all stages of a patient’s healthcare journey. The need for intelligent data integration processes has never been more acute. Consider that data are the oil that will grease the engines driving the transition toward value-based care, and yet health systems are struggling to integrate data within their own enterprises, and even then, it’s locked away from other sectors of healthcare behind proprietary standards.

LISTEN: The Hunt for the Orangeworm Hackers

“Data is at the center of all decision-making and all efforts to improve healthcare,” says Keith Dunleavy, MD, chairman and CEO of Inovalon, a cloud-based data analytics company that recently 

 the $1.2 billion acquisition of ABILITY Network, which is another cloud-based analytics solution provider with a network of more than 43,000 connected sites of care comprising acute, post-acute, and ambulatory providers.

, we discussed the massive scale of the problem. First, there are the different types of data, including claims data, clinical data, patient-reported data, patient behavioral data (whether a patient is following a recommended treatment, for example), data from medical devices and wearables, and even genetic marker data. Now, consider that the average patient typically has data residing in 12 separate technology systems within their healthcare universe and that these systems could be anywhere in the country. Data could reside in very sophisticated, cloud-based architecture, or it could live in a very basic diagnostic laboratory system.

So what needs to happen in the middle of all this data disparity and information silos? Our discussion centered on 4 key components.

 The technological capability to connect and aggregate massive amounts of data.

 The ability to analyze that data in petabyte volumes at incredibly high speeds.

 The ability to take the analysis from that compute and drive meaningful impact in the marketplace, at the point of care.

 The ability to take all of that impact and summarize it into data visualization and reporting and processes that are necessary for the back-end of healthcare to operate and be strategically directioned.

It's these 4 capabilities that enable what is referred to often in healthcare as the transformation from a volume-driven healthcare market to a value-driven healthcare market.

The Case for a National Analytics Backbone

This digital transformation to value-based healthcare is driven by the need for greater patient engagement, caregiver enablement, and increased operational efficiencies in health systems. But it will depend on a set of strategic enablers, one of which is a robust data and analytics capability. In an era of standalone electronic health records (EHR) systems and rapidly emerging new sources of data, and where healthcare technology solution providers look just as fragmented as the health systems, healthcare will need an entirely new model for data integration and analytics. Call it a national analytics backbone—one that can potentially fill gaps left by existing siloed systems.

What interested me the most about my friend’s medical incident was that even though he had traveled to a different state, the clinicians on duty could easily view his medical records through a data-sharing network enabled by the same EHR system where his original medical records currently reside. The instantaneous access to his medical records may have had a large role to play in the precious few minutes in which his doctors had to decide how to treat him.

My friend was lucky that the hospital where he found himself was using the same EHR as his own health system. The outcome of his care could have been very different had he ended up in a different hospital with a different EHR system that was not interoperable with its competitor. It’s possible that a national analytics backbone will help others like my friend avoid premature encounters with the “widow maker.”

Big Data, Analytics Ready to Meet Healthcare Challenges

Before Striking Outcome-Based Pharma Contracts, Payers Need Data

Computer Models Are Changing How We Craft HIV-AIDS Policies

A national analytics structure is important for healthcare, but it requires 4 capabilities to succeed.

How to Define a Model (and an Imperative) for a National Healthcare Analytics Backbone

Our latest healthcare technology survey is in, and the data are clear. After gaining the insight from 38 individuals at global technology firms, population health management companies, and digital health organizations, the majority of our survey respondents expect to see growth in healthcare. Of these respondents, 25% are targeting 30% or more in growth in 2018. But even more interesting is the fact that data and analytics was by far the single biggest area identi&#64257;ed by respondents (76.3%) for increased technology spend in 2018.

The underlying components behind this trend are simple: The march toward value-based care continues unabated. Improvements in population health are necessary to make value-based care a reality. And these improvements will not happen without big data and analytics, to actually glean actionable insights from the wealth of new information that’s now available.

For further proof of this trend, consider that, 

according to Healthcare Growth Partners (HGP

), their health information technology (IT) index gained an astonishing 25%. Health IT merger and acquisition activity and valuations hover at all-time highs. And you may want to sit down for this part: Global IT investment is up 849% since 2011, and investment in US health IT is up a staggering 583% during the same period. 

 the top 6 funded value propositions in 2017 and found that consumer health information was the most funded, as demonstrated by 5 of the 8 mega deals that happened during that period.

This Trend Is Heading in the Right Direction…

So where exactly is all of this investment in IT and analytics headed? Cognitive computing and its subset, artificial intelligence (AI), are playing a big role, with Accenture forecasting that the market for AI in healthcare is expected to grow 

. This is probably due to the fact that this forecast also predicts a $150 billion in savings opportunity for healthcare through AI applications to enhance diagnoses, boost efficiency and even accomplish such far reaching tasks as 

These types of learning systems can augment human thinking and reasoning capabilities to pull insights from structured and unstructured data alike. Effective variations of cognitive computing and AI never stop learning. They can identify hidden markers of disease that a manual review might miss, predict health population trends through a review of lifestyle data, and offer potential solutions to complex individual care challenges through deterministic processes based on the expertise of everyone who has contributed to the healthcare value chain.

We also need to pay attention to that hot new upstart known as blockchain. By using a distributed database consisting of groups of blocks that only approved members of a network can access, this technology offers the promise of much higher trust and transparency for healthcare data. This can enable silos of healthcare data to open up between systems, each with a potential piece that can unlock a revelation in terms of treatment for a chronic condition or a fix to current inefficiencies and restrictions in our healthcare system, which are driving up costs of care. Blockchain can also eliminate unnecessary third-party processing of financial transactions between payers and providers, driving costs of care even lower.

…But Here Are the Forces We Need to Watch.

Blockchain’s potential impact on interoperability cannot be understated, as interoperability is one of the critical issues creating a headwind against progress in healthcare analytics. Interoperability must move beyond simple interfaces to the point where all data, all the time, is fully accessible to cognitive computing and advanced AI. It's the only way these technologies can yield the results that all current investments in research and development are pursuing.

Healthcare data live in more places than ever before. From the proliferation of electronic medical records and the real-time information collected by wearable devices through the Internet of Things (IoT) to the financial histories locked away in payer systems, interoperability’s reach must expand ever further and cast an ever-wider net to be effective.

Regulatory burdens placed on healthcare organizations, such as patient privacy issues associated with HIPAA, can prevent this free flow of information. But healthcare businesses can also place restrictions on data sharing in the name of protecting proprietary assets. And IT resources can be stretched to the limit trying to incorporate the new data inputs that come online through IoT devices and new cloud-based technologies. All of these factors can come together to choke off the essential healthcare data raw materials that analytical algorithms are waiting for to create new diamonds of visibility.

Interoperability is a definitive an area of concern, as half of our survey respondents cited this challenge as the second-biggest obstacle to their IT spend growth in healthcare. And an even larger majority (60.5%) feel that enterprise readiness to adapt to new technology is another headwind against growth. If we’re going to unlock the power of healthcare analytics to its full potential, then it’s us, not just the algorithms, that need to evolve.

There's No Denying That Analytics Have Gone Mainstream

, you see analytics, AI, and blockchain coming up as recurring themes. These technologies are steadily becoming part of healthcare's mainstream consciousness, and we'll see them more and more as embedded parts of digital health applications and as a force to improve population health management. Looking ahead, we should not ignore the opportunity to apply analytics to address our pressing public health issues, such as the current 

. If we can overcome the interoperability chokepoints, then analytics-driven insights can transform how we address some of the heaviest burdens weighing down our system and start shifting many different healthcare trends to the right direction.

, a healthcare growth advisory firm based in Chicago. You can follow him @paddypadmanabha.

Spending on healthcare data and analytics is rising significantly, according to a recent survey. What factors are pushing the trend?

Paddy Padmanabhan, MBA

Articles

Addressing AI's Adoption Gap Problem in Healthcare

The Rising Clamor for Explainable AI

How to Define a Model (and an Imperative) for a National Healthcare Analytics Backbone

Why Data and Analytics Could See the Biggest Increase in Technology Spend in 2018