I don’t claim to be an expert but I can confidently assert that there are many, many spruikers for the advantages and possibilities of AI in the health sector.
This from the Harvard Business Review provides a flavour:
GenAI Could Transform How Health Care Works
by Ron Adner Dr. James N. Weinstein
November 27, 2023
Summary.
Consider how Napster, the networked file sharing system, upended the music industry. The emergence of generative AI language models like ChatGPT, has much in common with this Napster-initiated inflection point: a breakthrough technology with breathtakingly fast adoption, appropriation of other people’s data (OPD), and predictions of doom and obsolescence. Similarly, the generative AI revolution that ChatGPT has catalyzed is not going to be reversed. Leaders should look to three touchstones to calibrate their strategies and prepare for the transition: First, distinguishing between the role of AI in driving technology substitution and its role in ecosystem transformation. Second, preparing for the new organizational design challenges that will arise because of this ecosystem transformation. And third, crafting strategies that take advantage of new asymmetries that arise from new combinations inside and outside your own organization.
------
The word “disruption” is usually associated with technology substitution that offers a better way of accomplishing a given task. But its more profound realization is in ecosystem transformation that rewires and resets the boundaries across the old silos. This distinction is crucial in confronting the impact of generative AI.
Consider Napster. The networked file sharing system upended the music industry. Before Napster, music companies wrangled for years over how to engage with the digitization of music. And then Napster took the decision out of their hands — it broke the logjam. Initially, music industry leaders sounded alarms about rampant theft of intellectual property. But ultimately music-as-data led to a new golden age of profits as individual songs were re-aggregated into personalized streams by new actors like Spotify and Apple Music. The model shifted from album sales to predictable monthly subscriptions. Today more music is heard, by more people, in more places, with more profit to the music companies than ever before. Ecosystem transformation unlocks value.
The present moment, so full of debate about the impact and implication of generative AI language models like ChatGPT, has much in common with this Napster-initiated inflection point: a breakthrough technology with breathtakingly fast adoption, appropriation of other people’s data (OPD), and predictions of doom and obsolescence. And while OpenAI and ChatGPT, like Napster, may themselves be eclipsed by subsequent organizations and platforms, the generative AI revolution that they have catalyzed is not going to be reversed.
(Where Napster’s enabling of the appropriation and distribution of other people’s music was core to its value, ChatGPT’s appropriation of other people’s data for training its large language models is now the subject of numerous lawsuits. We fully expect the question of intellectual property will loom large in AI’s future. Our focus here, however, will be not on the original training data but on the new and proprietary data to which these learning models will be applied.)
How will the advent of large language models and other new AI approaches reset your sector, and how should leaders prepare? Our discussion focuses on the impact of AI on the U.S. health care context, but our broad points apply to every complex ecosystem wrestling with this new stage of digital revolution. From our positions as a technology executive and former health system CEO (Weinstein) and a strategy researcher and advisor (Adner), we present these ideas in the hopes that leaders might conceptualize new ways of strategizing and interacting.
Three touchstones will help leaders calibrate their strategies and prepare for the transition: First, distinguishing between the role of AI in driving technology substitution and its role in ecosystem transformation. Second, preparing for the new organizational design challenges that will be required for this ecosystem transformation to deliver its value. And third, crafting strategies that take advantage of new asymmetries that arise from new combinations inside and outside your own organization.
Technology Substitution vs. Ecosystem Transformation
ChatGPT broke the record for technology adoption, gaining 100 million users in two months. Most discussions have focused on the question of what tasks it will improve and what jobs it will replace. In other words: technology substitution and how to respond to it. But it is disruption at the ecosystem level that transforms the game and raises the biggest opportunities for change. By combining and analyzing data across previously disconnected silos, generative AI creates the opportunity to raise the bar on efficiency and effectiveness across the spectrum of health care delivery. Consider just three examples:
Billing and Claims
Administrative expenditure accounts for 15–30% of health care spending in the U.S., of which about half is consumed by hospitals’ management of billing- and insurance-related expenses. And even these estimates are unfairly low as they ignore non-dollar indirect cost, borne by patients and their families — the time spent fighting for insurance coverage and clarification on billing. Allowing artificial intelligence to break the silos between insurers, hospitals, and consumers would automate claims management, prior authorization, and even payment planning and collections, helping to eliminate a massive drag on system efficiency.
Resource Management
Health systems are plagued by long cycles of oversupply (buffers held in case of emergencies) punctuated by sudden shortages (when emergencies turn out bigger than expected) of equipment, medicine, rooms, beds, and staff. Poor management of patient flow causes unnecessarily long hospital stays and delays in admissions for those who in serious need. Lack of coordination with extended care and rehabilitation facilities increases time spent at the most expensive place for care, and puts patients at increased risk for hospital acquired complications. AI will enable cross-platform coordination across hospitals, systems, partners, and vendors to create higher resilience and better patient placement, lowering risk, shorten recovery times while improving outcomes and lowering cost.
Redefining Quality
A positive outcome from an avoidable surgery? Accurate results from an unnecessary test? These contradictions highlight the need for quality and performance measures that consider the patient and the patient journey more holistically. By incorporating the latest advances in medical science and real-world evidence into treatment recommendations and measures, AI stands to improve patient outcomes and raise standards in ways that reduce burden on both the patients and the system.
As you set your vision for an AI-enabled future, consider the balance of aspiration between inside-the-box substitution and cross-silo transformation. How is this balance reflected in your investment priorities: capital expenditures, operating expenditures, and capability development?
Ecosystem Transformation Requires Organizational Transformation
Change is held back by an inability among the actors who succeed in the current system to find their way towards a new equilibrium. High cost on one income statement shows up as high revenue on another. These income statements, literal and figurative, are determined by organizational boundaries, routines, and records. The AI-enabled transformations in billing, resource management, and quality described above all hinge on sharing data in novel ways. This novelty, however, gives rise to a new set of emergent organizational challenges.
Changing data access changes authority
Historically, the decision hierarchy that guides the reporting structure in an organizations was matched by a parallel information hierarchy — incomplete views across silos that may lead to suboptimal decisions, but that allow for clear decision-making paths and more efficient execution. This is true inside organizations (e.g., nurses do not have access to HR records) and across organizations (e.g., hospitals don’t have access to insurers’ financial records). But the promised benefits of transformational AI rely on crossing these silos. This means that, beyond concerns of privacy and security, true transformation will require organizations to rethink the informational foundations of authority. Once released across data pools, AI eliminates organizational information censors. This is a huge ecosystem transformation, shifting the focus from insuring the accuracy of content (“Is the data correct?”) to controlling the breadth of questions (“Who is allowed to ask what?”). This shift from censoring data to censoring questions implies a radical change in fundamental principles administration and management.
New information demands new metrics
New visibility into new data combinations open debates on relevant and appropriate metrics which, in turn, impact goals and incentives. The core questions of what is the definition of success, and who gets to define it, moves to the forefront. Consider surgeon productivity in a world where data can be viewed across silos. Do you measure the number of procedures they manage in a month? The revenue they generate? What weights do you assign? In a world of merged data pools and open inquiry, anyone with access can create their own new measures, and the system needs to find a way of settling on a new equilibrium.
Transparency creates new responsibility
A corollary to visibility across data silos is the expectation of more holistic decisions that take the broader landscape into consideration. Historically, a doctor’s recommendation of the “best” treatment was based on optimizing medical outcomes. But with an AI-facilitated view of a patient’s broader circumstances outside of the medical visit – the specifics of their insurance coverage; their work situation; their homelife situation – the notion of “best” can change dramatically. How to incorporate economic and social lenses into a medical recommendation, and to do so in an ethically and legally defensible way, will become a critical new requirement for both providers and insurers.
As AI enables more visibility across silos what parts of your org chart and your governance need to be revisited. You must be proactive in making sure that the upside of giving more people more access to information is not overwhelmed by the unintended downside of new sources of conflict. For every organization in the ecosystem this will lead to a redefinition of rules and roles. For successful organizations, this will be handled with forethought, not as afterthought.
Lots more here:
https://hbr.org/2023/11/genai-could-transform-how-health-care-works
So, while we can see the added value in general the specifics of just how this way work clinically are a little hard to be sure of.
When you contemplate typical health data which is often disjointed, incomplete, full of abbreviations and hardly clearly formatted (or even legible) just how clear value is to be obtained can be a little less than obvious!
It is probably these attributes that provide the scope for an effective AI to add the most value and to obtain the most clarity, reliability, accuracy and utility!
Just having these attributes in records would surely improve patient safety and confidence in decision making!
What more arcane transformations of a record are fascinating possibilities as are the representational possibilities that might be engaged! A ‘meta’ record what adds value to the original is a fascinating possibility and clearly conceivable!
Clearly we can move well beyond just ‘tidying’ a record up – but how far might be useful I will leave to the reader’s imagination. However, I sure do not think the decrepit myHR is a place to start from!!!!
I am very interested in what others might add, and just what sort of record tool(s) might be ideal going forward!
David.
No comments:
Post a Comment