CMS' New Technology Add-On Payment Ruling
In September, in a groundbreaking rule, CMS granted the first New Technology Add-on Payment (NTAP) status to an AI medical software company. The company, Viz.ai, received NTAP status designation from a three-year period based on its ContaCT (Viz LVO) AI algorithm. Viz.ai initially received FDA clearance in early 2018 for their LVO AI Stroke Platform, a computer-aided triage and notification platform to identify Large Vessel Occlusion (LVO) strokes on CTA imaging.
As a result of this ruling, providers who use Viz LVO to triage patients with suspected acute ischemic stroke can receive CMS payments of up to $1,040 per use. NTAPs are granted by CMS for novel and highly-priced technologies which provide a substantial clinical benefit, and which are not covered under the Medicare Severity-Diagnosis Related Group (MS-DRG). Importantly, the standard radiology reimbursement for interpretation of CT scans is not affected by the NTAP ruling, as the payment is an add-on to the DRG.
How it works
In cases of suspected acute ischemic stroke, the phrase “time is brain” is established scientific dogma. Viz.ai’s LVO AI algorithm which consists of a convolutional neural network, trained using machine learning techniques, that connects directly to the CT scanner and analyzes CTA images of the brain in the acute setting. If the algorithm detects a suspected LVO, it then sends a HIPAA-compliant alert, along with images in a DICOM viewer to the mobile device of the team of interventional stroke specialists at a comprehensive stroke center (CSC), who would be capable of performing endovascular reperfusion treatments such as mechanical thrombectomy and intra-arterial thrombolysis. According to Viz.ai, their LVO algorithm automatically identifies suspected LVO on CTA images of the brain with an accuracy of 90% in cases of proximal anterior circulation LVO.
Using this new workflow, the company bypasses several steps in the current serial stroke model and creates a parallel process model that aims to reduce patient transfer times and improve patient outcomes. In several small-scale studies, Viz.ai demonstrated that the use of their Viz LVO platform led to faster time in notification of the interventional stroke specialists, faster patient transfer times from primary stroke centers to comprehensive stroke centers, and a decrease in overall hospital length of stay and length of stay in the neurological ICU. In terms of improvement in clinical outcomes, in these studies Viz.ai demonstrated improved modified Rankin scores (mRs) at discharge and at day 90, and improved NIH Stroke Score (NIHHS) at day 5.
The past and the future
Prior to this ruling, many medical AI companies had received FDA clearance, and some even had FDA approval, but for most health systems there existed little to no financial incentive to institute and use these products. While many AI companies emphasized their efficiency improvements and clinical benefits, the lack of adequate reimbursement for use of their products remained a significant AI adoption barrier. Viz.ai’s success in achieving NTAP status, and the associated added reimbursement will likely result in more hospital systems adopting this AI technology with the hope of improving outcomes for stroke patients. The added financial incentive is particularly important during COVID-19 as many health systems try to mitigate the financial impacts of the pandemic.
Viz.ai charges users with a yearly subscription fee, and the NTAP reimbursement is designed to support health systems in covering that subscription fee. However, the reimbursement model is complex, as it applies only to Medicare patients, and only in cases in which the hospital system would incur a loss from the treatment provided to the patient. Additionally, the more stroke patients a hospital system sees, the less it will likely be reimbursed from CMS NTAP as the total reimbursement reaches the cost of Viz.ai’s yearly subscription fee.
What remains true, however, is that the results of this ruling will serve to provide AI companies with a pathway to adoption of their products and revenue generation. It still remains to be seen how much more NTAP reimbursement CMS will set forth for other AI companies, but we now have precedent. The reason that Viz.ai’s current platform works is because even when it fails and misses a stroke, the patient will receive the current standard of care — that is, review of the images by a diagnostic radiologist.
The role of diagnostic radiologists in acute stroke care
As previously noted, the model bypasses several serial steps in the current stroke evaluation paradigm to create a new parallel workflow. One of the critical steps that it bypasses, at least initially, is review and interpretation of the images by a diagnostic radiologist. The software sends the images directly to the stroke team via their HIPAA-compliant mobile application. The team can then make decisions about treatment in real-time. While the company makes note that images in their mobile viewer are compressed and intended for notification and informational purposes only and not for diagnostic use, it remains to be seen if and how the role of the diagnostic radiologist will change at centers using the Viz LVO platform for acute ischemic stroke care. Large-scale, multi-center studies are needed to systematically assess the scope of impact of Viz LVO platform on patient clinical outcomes.