Artificial intelligence (AI) has two faces in healthcare.
One face sings the praises of AI as the tonic that will enable healthcare to
deliver better clinical outcomes at a lower cost and the second face is full of
skepticism and raises barriers to adoption at every turn. It is heartening to
see that a third face is emerging, the thoughtful and appropriate use of AI to
predict adverse health events; to identify and stratify patients in need of
health, social, and human services; and the application of AI in the automation
of tasks, activities, and processes.
To understand the likely evolution of AI-based automation,
it’s important to evaluate the interaction of humans and machines across these
five levels. At each level of automation, the following questions must be asked
and answered:
Who produces insights? – Does the human
or the machine (AI) analyze data and deliver insights from such analysis? Does
the human or the machine describe what something is, how it trends, why
something is happening, and what might happen next?
Who decides and how? – Once all relevant
analysis has been conducted, does the human or the machine make the decision
based on the derived insights?
Who acts based on the decision? – Finally,
a decision should lead to an action by either a human or a machine? The action
can be in the digital or physical environment.
Based on the responses to these questions, IDC has
identified the following five levels of AI-based automation:
Human Led – At the first level, it is
the human who analyzes the data using limited technology, such as tools for
only descriptive analytics; it is the human who makes the decision based on the
analysis (or experience); and it is the human who acts based on the decision.
Human Led, Machine Supported – At the
second level, the human continues to lead data analysis, decision making, and
action steps but is now more reliant on the machine across these steps.
Machine-led, Human Supported – At the
third level, it is the machine that is using a wide range of analytic and AI
techniques to conduct the analysis and produce insights. These insights are
reviewed by humans. The human still makes the decision based on machine’s
recommendations, and it is the human who acts based on the decision. However,
at this level, the machine acts to provide oversight over human decision making
and execution.
Machine Led, Human Governed – At the
fourth level, the machine analyzes data and produces insights without the need
for human review. At this level, the machine decides based on the analysis of
all available data and a framework of human-developed governance policies and
procedures. At this stage, it is also the machine that acts based on the
decision under the governance of humans.
Machine Led – At the fifth level, the
world has likely achieved general AI. At this stage, there is a full AI-based
automation without the need for human involvement. At this level, we need to
think of machines that set their own goals and understand all mathematical,
economic, legal, and other external constraints. Most AI academics and experts
in labs of commercial enterprises predict this level of AI to arrive no sooner
than in about 50 years.
In
recent years, one of the shortcomings in the commercial sphere of AI has been
the misrepresentation of the scope of possible automation. Too often, we hear
claims of AI systems automating end-to-end processes and predictions of massive
labor losses, this does a disservice to organizations trying to plan for the
appropriate level of investment in AI. There is a need for a pragmatic
framework that decision makers across industries can use to assess
opportunities and risks of AI-based automation. The levels of AI-based
automation must also be viewed in the context of the scope of automation. We
define this scope where:
Task is the smallest possible unit of work
performed on behalf of an activity.
Activity is a collection of related tasks to be
completed to achieve the objective.
Process is a series of related activities that
produces a specific output.
System (or an ecosystem) is a set of connected
processes.
IDC’s AI automation framework was developed to help wade through the hyperbole associated with AI. Our goal is to help provide a planning tool and key piece of vendor evaluations processes to fully understand the role AI is playing in software and guide strategic decision making.
Guest post by Deanne Primozic Kasim, research director, payer health IT, IDC Health Insights.
Who will be the next big health plan to merge with and/or acquire Humana? This has been the topic of a lot of industry conversation and I have actually thought about starting an online betting pool. In lieu of creating a “bracket challenge,” I have been informally asking opinions on this topic and the results are all over the map. A deal for either Humana, valued at $27 billion, or Medicaid provider Centene Corp., at $8.3 billion, would be the biggest acquisition of a U.S. payer in more than a decade. It’s an exciting time this summer with the upcoming Supreme Court decision on the legality of insurance subsidies on the federal marketplace, and an anticipated consolidation between Humana and another large insurer.
Why will health plan consolidation continue? Two major reasons:
Increasing Medicare and Medicaid memberships
According to the Centers for Medicare and Medicaid Services (CMS), Medicare membership is expected to reach 68.4 million in 2023, up from a projected 54.4 million this year. In addition, Medicaid will add 9.3 million people over the same period. Baby boomers are hitting Medicare eligibility at a rate approaching 10,000 per day. Part of the reason companies like Humana and Centene are attractive acquisitions to other payers is because of the size of their Medicare and Medicaid memberships. It is important to note that regardless of how mergers and/or acquisitions play out, payers with multiple lines of government-sponsored business will need effective processes and related IT tools for understanding the complex demographics and health needs of these members. Many Medicare and Medicaid members are high-risk/high-value patients who have clinical and non-clinical factors that can result in the development of costly and complex conditions.
IDC Health Insights estimates that the combined annual growth rate in the analytics market during the 10 years from 2010 through 2020 will be in the 8 percent to 11 percent range; this places analytics among the top areas of spending growth for hospitals and health systems during this decade. This attractive growth rate has led to numerous new products joining an already-crowded supplier landscape.
The U.S. healthcare provider analytics market has experienced rapid growth and change since the introduction of accountable care with the patient protection and affordable care act (PPACA) of 2010. Analytics are clearly a critical tool that will allow health systems to understand and respond to the business model change and disruption of accountable care, and many types of analytics models and tools will likely be useful to providers. This IDC MarketScape report focuses on analytics platforms that allow providers to examine clinical and financial data together, and to use this data to provide actionable advice for optimizing delivery of care.
Key findings from the report include:
Clinical and financial analytics take many forms. This report examines platforms that allow providers to approach analytics in multiple ways, with agile tools that may include clinical and financial analytics, text and data mining, population health analytics, cost and cost accounting analytics, performance and quality management analytics and dashboards, as well as data exploration tools that can be applied to as-yet-undiscovered questions. This report examines the flexibility of analytics platforms as well as the strength and weaknesses of individual analytics applications available on the platform.
No analytics solution will meet all needs out-of-the-box. Successful analytics programs will develop and nurture platforms that assemble and manage data, offer tools to ensure data quality, and offer applications that allow providers to explore and assemble data on-demand into analytics models that meet business needs, whether they are long-established business needs or spur-of-the-moment questions.
The only valuable analytics are actionable analytics. Analytics are only valuable if they make the right information available, at the right time, at the point of decision making. Solid data and data management approaches are the foundation of analytics platforms, but the rigor of data integrity processes must be balanced.
IDC Health Insights new report that describes the distinctive offerings and capabilities from six fraud, waste and abuse (FWA) services providers and the market they compete in. The new study, IDC Marketscape: US Healthcare Payer Fraud, Waste and Abuse Services 2015 Vendor Assessment,” (Document #HI253636) examines in detail the FWA consulting and business process outsourcing (BPO) services. Vendors evaluated in this report include: Emdeon, IBM, McKesson, Optum, SCIO and Xerox.
Priority for investment in payer fraud, waste and abuse solutions is rising quickly on payer executive agendas because of the accelerated rate of evolution of the payer marketplace in the wake of government-mandated reform. Investment in U.S. healthcare payer FWA solutions (software and services) is rising rapidly in 2014–2015. While already present in over 50% of payers, in the next few years, such solutions will become ubiquitous among payers of all types and sizes in both the commercial and government sectors in some form. In the next three to five years, many payers will continue to enhance their overall FWA defensive capabilities by investing in oftentimes complementary solutions from multiple vendors.
The services component of payer FWA solutions is becoming more and more important, and it is now common for payer FWA software to be delivered to buyers “as a service.” Technology improvements in analytics and software-as-a-service (SaaS) solution delivery have made it more feasible for vendors to provide rapidly scalable, outsourced services at decreasing costs with defensive capabilities that outstrip those available to all but the largest payers.
Because of these developments, payers are therefore seeking guidance on how to invest in the different types of service offerings available and how the major vendors compete with each other. This IDC Health Insights report uses the IDC MarketScape methodology to assess six vendors competing in the U.S. healthcare payer FWA services market. A detailed profile supports the evaluation of the offerings from each of the selected vendors. Each profile outlines a vendor’s offering components, evolution, strengths, and challenges.
As health IT continues to mature and providers continue to adopt technologies like electronic health records, the data collected from their use in the care setting becomes the most obvious reason so much energy is being put behind getting practices to implement the systems.
Judy Hanover, research director of IDC Health Insights, recently told me, though, that one of the biggest challenges faced by ambulatory and hospital leaders is that the data entering the electronic systems, in most cases, is unstructured, which makes it almost useless from an analytics standpoint.
Without structured data, Hanover said, quantitative analysis across the population can be complicated, and little can be compared to gain an accurate picture of what’s actually taking place in the market. Without structured data, analytics is greatly compromised, and the information gained can only be analyzed from a single, siloed location.
“There must be synergy between the data collected,” Hanover said. “We’re entering the period of structured data where we’re now seeing the benefits of structured data but still need to manage unstructured data.”
In many cases, critical elements of data collected — like medications, vitals, allergies and health condition — are difficult to reconcile between multiple data sources, reducing the quality of the data, she said. Unstructured data proves less useful for tracking care outcomes of a population’s health with traditional analytics.
For example, tax information and census data are collected the same way across their respective spectrums. All the fields in their respective fields are the same and can be measured against each other. This is not the case with the data entering an EHR. Each practice, and even each user of the system, potentially may collect data differently in a manner that’s most comfortable to the person entering the data. And as long as practices continue to forgo establishing official policies for data entry and requiring data to be entered according to a structured model, the quality of the information going in it will be a reflection of the data coming out.
Lack of quality going in means lack of quality coming out.
“In many cases, structured data is not as useful for analytics as we’d hoped,” Hanover said. “There are inconsistencies in the fields of data being entered in to the systems; and that affect data quality as well as results from analytics.
“As we move into the post EHR era, how we choose to leverage the data collected is what will matter,” she said. “We’ll examine cost outcomes, optimize the setting of care and view the technology’s impact.”
As foundational technology, EHRs are allowing for the creation of meaningful use, but once the reform is fully in place, the shift will focus on analytics, outcomes and benefits of care provided.
Currently electronic health records define healthcare, but health information exchanges (HIE) will cause a dramatic shift in the market leading to further automation of the providing care and will change how location-based services and clinical decision making are viewed.
Though some practices are clearly leveraging their current data, others are not. For them, EHRs are nothing more than a computer system that replaced their paper records and qualified them for incentives.
In the very near term, the technology will have to have more capability than simply serving as a repository for information collected, but will become a database of reference material that will have to be drawn upon rather than simply housed.
“Health reform is the end game,” Hanover said. “And there can be no successful reform without EHRs. They are the foundational technology for accountable care.”
The data collected in this manner will lead to a stronger accountable care model, which will once again bring the practice of care in connection with the payment of care.
Evidence-based approaches will continue to dominate care when the data suggests certain protocols require it, which means insurers will feel as though they are working to control costs.
Unfortunately, all of the regulation comes at an obvious cost at the expense of the technology and its vendors, said Hanover. EHR innovation continues to suffer with the aggressive push for reform through meaningful use as vendors scramble to keep up with requirements.
“There’s little or no innovation because all of the vendors are being hemmed down by meaningful use and certification requirements,” she said.
Product standardization means there are far fewer products that actually stand out in the market.
More innovation will likely only come following market consolidation in which only the strong will survive. Hanover suggests that in this scenario, survivors will focus on innovative product research and development and will take a leadership role in moving the market forward
Though vendors will suffer, users of the systems will likely face major set backs and upheavals at the market shifts and settles. Especially as consolidation occurs, suppliers disappear or change ownership, practices and physicians using these systems face the toughest road as they’ll be forced to find new solutions to meet their needs, learn the systems and try to get back to where they were in a meaningful way in a relatively short period of time.
Likely, deciding which system to implement may bear just as much weight as deciding how to use it.