Navigating the Maze: Coverage and Reimbursement for AI-powered SaMD

Software as a medical device (SaMD) offers exciting possibilities for improving healthcare. But for developers, a crucial hurdle lies in securing coverage and reimbursement from payers. This landscape can be complex, especially for SaMD powered by artificial intelligence (AI) or machine learning (ML), as these can introduce unique considerations for value assessment. Let’s examine   the current state of SaMD coverage and reimbursement and how payers approach the value proposition of AI and ML in healthcare. 

What is SaMD? 

The U.S. Food and Drug Administration (FDA) defines a Software as a Medical Device (SaMD) as: “Software that meets the definition of a device in section 201(h) of the FD&C Act and is intended to be used for one or more medical purposes without being part of a hardware device.”  

This definition has two key parts: 

  1. Meeting the Definition of a Medical Device: The software must qualify as a medical device according to the Federal Food, Drug, and Cosmetic Act. 
  1. Standalone Medical Purpose: The software’s intended use is for one or more medical purposes, and it achieves this purpose without being physically part of a hardware medical device. 

The Evolving Coverage Landscape for SaMD 

Currently, there’s no single, standardized pathway for SaMD coverage. Reimbursement often hinges on whether the SaMD falls under an existing Current Procedural Terminology (CPT®) billing code or requires a new one. Here’s a breakdown of the main approaches: 

Existing Codes: If a SaMD fulfills a function similar to an existing service (e.g., remote monitoring), it might be reimbursed under the same code. However, this can be challenging as existing codes might not fully capture the unique value proposition of a SaMD and result in inadequate reimbursement. 

New Codes: Developers can petition for new CPT codes specific to their SaMD. This process requires demonstrating the unique clinical utility of the software, which can be complex and take 2-3 years. 

Molded Care Pathways: Some payers are exploring bundled payment arrangements for specific care pathways, potentially incorporating SaMD alongside other services. This approach offers flexibility but requires collaboration between developers and payers.  

Challenges in Assessing Value for AI-powered SaMD 

Payers are naturally cautious about reimbursing new technologies. When it comes to AI-powered SaMD, specific challenges arise in assessing value: 

Limited Evidence Base: AI in healthcare is a relatively new field, and long-term data on the cost-effectiveness of AI-powered SaMD can be scarce. Payers might require robust clinical data demonstrating the technology’s ability to improve patient outcomes or reduce costs. 

Black Box Problem: The inner workings of some AI algorithms can be opaque, making it difficult for payers to understand how the technology arrives at its conclusions. Transparency around AI decision-making is crucial for building trust with payers. 

Integration Costs: Implementing AI-powered SaMD might require additional investments in infrastructure or training for healthcare providers. Payers need to factor in these potential integration costs when evaluating the SaMD’s overall value proposition. 

How Payers Assess Value: A Framework for Success 

Despite the challenges, SaMD developers can demonstrate value to payers in a variety of ways: 

Focus on Clinical Outcomes: Provide robust clinical data demonstrating the impact of your AI-powered SaMD on patient outcomes as compared to standard of care. This could include improved disease management, reduced hospital readmission rates, or earlier diagnoses. 

Cost-Effectiveness Analysis: Conduct economic evaluations to show how your SaMD can reduce overall healthcare costs. This could involve demonstrating potential savings from reduced hospital stays or improved medication adherence. 

Focus on Interoperability: Develop SaMD that integrates seamlessly with existing healthcare IT systems, minimizing the disruption and additional costs for providers. 

Transparency in AI Design: Strive for transparency in how your AI algorithms function. This could involve providing clear explanations of how the technology arrives at its recommendations. 

The Takeaway 

The coverage and reimbursement landscape for SaMD is evolving, with AI-powered solutions facing unique challenges precisely because of how innovative they are. By focusing on robust clinical data, cost-effectiveness, interoperability, and transparency, SaMD developers can increase their chances of securing coverage and reimbursement, paving the way for wider adoption of these innovative healthcare tools. 

How HcFocus Can Help 

HcFocus provides a holistic approach to market access. For example, we help medical innovators draft regulatory submissions that make it easier to later obtain reimbursement from payers. Our expertise remains in creating and articulating the value of SaMD to both regulatory and payer stakeholders to drive approval, coverage, and reimbursement.  

Here are some resources to learn more: 

References 

AMA: CPT® overview and code approval 

https://www.ama-assn.org/practice-management/cpt/cpt-overview-and-code-approval

The Medicare Payment Advisory Commission (MedPAC): Medicare coverage of and payment for software as a medical service: An overview 

https://www.medpac.gov/wp-content/uploads/2023/03/SaaS-MedPAC-11.23.pdf

NPJ Digital Medicine: Paying for artificial intelligence in medicine 

s41746-022-00609-6.pdf (nature.com)  

FDA: Software as a Medical Device (SaMD):  

https://www.fda.gov/medical-devices

FDA: Device Software Functions Including Mobile Medical Applications 

https://www.fda.gov/medical-devices/digital-health-center-excellence/device-software-functions-including-mobile-medical-applications