Digital health chronicles: Opening the gates of clinical AI for clinical decision making
|Digital health chronicles: Opening the gates of clinical AI for clinical decision making
Despite the global health situation, 2020 is a busy year for digital health worldwide. Alcimed highlights in these chronicles major events that are going to shape the digital health efforts for the coming years.
Viz.AI is one among many companies positioned on AI for clinical decision making. Actually, it is one among hundreds of these companies stating that through a relevant use of AI, one could significantly boost clinicians’ performance, simplify patient workflow, improve medico-economics, or outcomes. In fact, Viz.AI’s mission is to leverage AI to detect Intra Cerebral Hemorrhage and triage patients quicker to avoid strokes.
What makes it so special? On September 3rd, 2020, the company received a Medicare New Technology Add-on Payment (NTAP). This is a first, and it means that a large public health system is ready to cover AI-powered clinical decision-making systems.
What are the details and mechanisms of this funding? What is the impact for this crowded branch of digital health?
Viz.AI and its AI-based software for clinical decision making, what are the advantages?
Viz.AI is a Californian company with offices in Israel that develops and commercializes its Viz ICH AI-based software to automatically identify suspected ICH strokes on non-contrast CT imaging. The software can alert on-call stroke teams and facilitate patient workflow and triage, leading to significant improvement regarding Time-to-Treatment.
The software for clinical decision-making has been granted US FDA De Novo Clearance in 2018. According to the company CEO, the software is implemented in circa 500 hospitals across the US.
In the stroke environment, “time is brain” is a common motto, so such claims have huge clinical and medico-economic impact. In a paper published in February 2020, the company demonstrated on a cohort of 43 patients a reduction of 22.5 minutes in average of the transfer time from the CT angiogram imaging to entry into a comprehensive stroke center (132.5 minutes vs 110 minutes). Beyond the use of AI, the paper shows that in the interventional arm, the care pathway evolves, cutting out the radiologist in the process and going directly to the neuro-interventionalist, which can also contribute to delay reduction.
Towards Centers for Medicare & Medicaid Services (CMS) reimbursement
While AI, machine learning, and deep learning are full of promises, getting a financial incentive to use it is probably the best (only?) way to reach market for now.
So, back in September 2020 comes the CMS announcement of a NTAP for Viz.AI of up to $1,040 per use in patients with suspected strokes. That is seen as a major step forward as many of AI-based solutions for clinical decision making struggle hard to find somebody to pay for it, especially at such a price tag. And while AI, machine learning, and deep learning are full of promises, getting a financial incentive to use it is probably the best (only?) way to reach market for now. Where Viz.AI went beyond many players is that they did not stop at operational impact (reduction in time to manage patients) but hunted for clinical outcomes which ultimately convinced CMS to move.
The first comment to make is that Viz.AI has been submitting a business model based on subscription, which is uncanny in the medtech world (generally driven by fee-for-service business model). While the exact amount of the subscription is not known, CMS documents highlight a yearly bill of $25,000.
Now, about the $1,000 price tag that made the headlines. This price reflects the maximum amount of reimbursement a hospital can claim from CMS if Viz ICH has been used on an “eligible” patient. So if a hospital has less than 25 patients a year using the software, it will lose money in the process. If a hospital has thousands of eligible patients, it will use more of the software and may lead to high value generation. Fact is, there are so many exclusion criteria that even in large stroke centers, the probability of having thousands of patients requiring Viz ICH is very low (CMS data estimates somewhere between 2 and 3% of eligibility for all codes for stroke).
The point where hospitals may benefit from this is that they are offering to their patients better chances of getting a thrombectomy in due time, leading to better outcomes for the patient, savings from the system and surgical billing codes for the stroke center. Now the question remains: are hospitals incentivized here by the NTAP perspective ($1,000 of reimbursement for patients going by Viz ICH) or the higher number of thrombectomy rates leading to more surgical reimbursement?
The use of clinical AI for clinical decision making, what impacts on e-health?
Viz.AI succeeded in securing a revenue flow of somewhere in the magnitude of $25,000 for each equipped stroke center. This is already an achievement in the current state of clinical AI for clinical decision making market and business models.
What this story shows however is that the funding mechanisms allowing digital solutions to be covered is exceedingly complex. While AI and more broadly digital technologies support the Value-Based Healthcare concept, it’s not enough to get them reimbursed. What would be needed is a full revision of how we pay for health, and such revision should take into account the high versatility of healthcare systems worldwide. How would such a payment model cope elsewhere?
While Viz.AI paves the way for future innovators in the clinical AI environment, innovative business models are more than required to make it look more like The Shire than Mordor in the coming years. Interested in this topic? Discover our achievements in business models consulting for healthcare.
About the author
Amélie, Consultant in Alcimed’s Life Sciences team in France
Benjamin, Great Explorer Digital Health in Alcimed’s Healthcare team in France
Do you have an exploration project?
Our explorers are ready to discuss it with you