The Spotify iTunes Model For AI In Healthcare
Artificial intelligence (AI) in health care has become the subject of both great promise and great hyperbole. Beyond buzzwords and a plethora of venture capital investments, AI and other mathematical techniques are beginning to emerge in a second wave of domain-specific systems of intelligence. The key missing factor has been a business model in the payer-and-provider community that enables the “best” (aka: most validated, most clinically proven, most workflow-integrated) models to drive an economic value both to the care paradigm and to the cost centers of health care data storage and analytical processing.
Cloud Vendors Are Beginning To Create A Model For Performance-Based AI In Health Care
As the cloud storage wars have entered health care, a question has emerged: Beyond security and the improved economics of cloud storage versus physical storage, what will drive revenue for cloud vendors? Furthermore, how will horizontal cloud vendors prove the validity of AI offerings to impact the business needs of providers and payers such as improved quality, decreased workload and better outcomes? From IBM Watson to Microsoft, a Symphony AI partner, as well as Amazon and Google, the question of how AI practically affects care is critical for most health care players.
However, there seems to be a model emerging that could advance domain-driven AI though these platforms into practice based on performance and, perhaps, paid by usage.
How Companies Could Democratize AI In Health Care
Artists who submit content to Spotify or iTunes are inherently compensated based on the number of times users listen to their music (and based on other factors such as location, type of subscription, etc.). If the audience deems that it's a quality song, then compensation flows back to the creator of the content.
YouTube and others have established a similar model. What is fascinating is that as health care players have begun to transfer their data cloud storage, and a major part of their suite of services includes AI platforms that charge based on graphics processing unit (GPU) compute that is run when AI models process data. While it is early, this creates the potential that as more AI models are developed and run with greater frequency, compute will represent the unit on which this value could be measured.
For AI model developers in health care, this could mean that, whether you are a grad student or a multinational company, your model could have a marketplace by which you’ll succeed if your accuracy and utility actually drive a financial component of the data economy in health care.
Further, this puts health care providers and payers in the position of “voting” for AI models that actually positively impact their business and patient outcomes. Imagine an oncologist with a “playlist” of AI models they might utilize for a subset of patients, trained in regard to new data and transparently judged based on performance. This type of model puts the physician and administrator in a critical role of judging utility based on output and performance.
One example of this that could play out focuses on the hotly contested field of AI for radiology. Multiple startups (Enlytic, Arterys, Zebra Medical, others) and large multinationals (Google Deep Mind, Varian, GE, etc.) are working to drive prediction and augmentation of human processing.
In the academic literature, these companies compete on the basis of accuracy and reproducibility. Ultimately, they are likely to compete on repeatability and performance. In other words, may the best algorithm win. There is no area of health care AI that is more compute-intensive than the analysis and processing of three-dimensional imaging files. In this way, not only prediction performance but perhaps also compute cost end up becoming competitive aspects of this market.
Cloud Vendors Are In An Exciting Evolution As An AI marketplace
What is perhaps most exciting to me is that, despite years of siloed data preventing health care experts from answering even basic questions, cloud vendors' movement signals that we have now entered the stage of AI-driven insight. AI is no longer relegated to academic literature or complex custom deployments. Through teaming with cloud vendors, AI companies will be able to make this work mainstream by:
• Filling in gaps in electronic health record (EHR) and clinical record data to create consistently structured data fields.
• Helping oncologists optimize clinical workflow and care coordination with documentation, alerts and predictive analytics.
• Using real-world data (RWD) insights to help avoid adverse events for people with life-threatening indications and select the right treatment for the right patient.
AI technology is ready to prove itself in health care. There is half a decade of testing, experimentation and advances. New studies on AI in clinical use, such as the recent evidence published at ASCO showing pre-trained modules can successfully predict chemotherapy-induced neutropenia and make cancer treatment safer are emerging every day. With cloud vendors and the future emergence of a Spotify-style model marketplace, it is possible for AI to finally become mainstream. To help physicians and patients every day with life-and-death decision making that moves health care forward.