Artificial intelligence (AI) and machine learning (ML) are increasingly gaining a foothold in the medical device landscape. Their transformative potential is particularly evident in medical diagnostics, where AI models and algorithms leverage clinical data to identify and diagnose diseases with greater accuracy.
But as AI-driven diagnostics rapidly evolve, are payers keeping pace? Can reimbursement models adapt quickly enough to support widespread adoption?
A/ML in diagnostics |
| AI/ML-enabled diagnostic solutions offer significant opportunities to enhance diagnostic accuracy and detect early signs of disease onset (e.g., cancer, stroke or cardiovascular diseases), enabling timely intervention. Other potential applications include prognosis of disease progression, prediction of treatment response for personalized medicine, and risk assessment (e.g., complications, mortality). These technologies analyze complex medical data, reduce human error, improve pattern recognition, and automate anomaly detection across various data types – from imaging to pathogen data. They assist clinicians in making more informed decisions, improving both the effectiveness and efficiency of diagnostic workflow. |
The US Food and Drug Administration (FDA) has granted marketing authorization for a total 1,016 AI/ML-enabled medical devices to date, including more than 550 new medical devices in the last three years alone. The list of applicants includes medtech companies, startups, and large IT players outside the life sciences industry, such as Apple Inc. and Samsung Electronics Co.
CADe/CADx (computer-aided detection/diagnosis) software in radiology has emerged as a dominant category, experiencing 17.8% growth in the past three years. Other rapidly growing areas include gastroenterology-urology (+41%), such as AI tools used for colorectal cancer screening, and anesthesiology (+34.4%), such as AI tools for sleep apnea detection.
Payer coverage lags behind, but momentum is building
While more and more products come to market, payer coverage is still at an early stage. Most AI/ML-enabled solutions have yet to secure broad, national-level reimbursement. Though national funding is still lagging, national payers have slowly started to open up their pathways to accommodate AI. Meanwhile, most funding is still provided on a subnational and/or temporary basis such as in scope of innovation fundLet’s explore a few country examples of initial success stories where these devices received national, subnational. or temporary funding:
- United States:
The American Medical Association (AMA) has made significant strides in integrating AI into the reimbursement landscape by continuously expanding their Current Procedural Terminology (CPT) codes for AI-driven medical procedures. Most AI-related codes currently fall under Category III, which are temporary codes designed to track emerging technologies. However, a key milestone is the publication of a permanent Category I CPT code 92229 for AI-based assessment for diabetic retinopathy and macular edema, which has also secured national reimbursement from CMS.
Similarly, AI-enabled coronary plaque analysis, associated with Category III CPT code range 0623T-0626T, received transitional pass-through funding from CMS through local coverage determination, with Category I CPT code creation and permanent coverage anticipated by 2026.
- Germany:
While there is no precedent of broad national coverage of AI yet, there are success cases at a subnational level. For instance, contextflow, an AI-powered CADe support software for lung cancer, interstitial lung diseases, and other thoracic abnormalities, signed a reimbursement agreement with IKK Südwest in November 2024.
However, Germany is currently implementing a national lung cancer screening program that allows use of low-dose computed tomography (LDCT) for early detection of lung cancer in high-risk individuals. The regulation, effective since July 2024, specifies that radiologists must use computer-aided diagnostic software, such as iterative or deep learning-based reconstruction algorithms and lesion volumetry software that are CE-marked as a medical device. The Federal Joint Committee (G-BA) now has 18 months (until the end of 2025) to develop a directive, outlining criteria and a framework for implementing lung cancer screening in statutory health insurance (GKV) – paving the way for broad coverage via EBM-tariffs.
- France:
In 2021, Sunrise, an AI-based clinical decision support tool assisting in the diagnosis of sleep apnea, received approval for temporary reimbursement via Forfait innovation, limited to 14 centers. This selective funding to generate evidence could serve as an initial step toward long-term reimbursement, provided the outcomes meet HAS requirements for LPPR inclusion.
- United Kingdom:
In 2021, HeartFlow FFRCT, an AI-enabled technology for estimating fractional flow reserve from coronary CT angiography, was added to the MedTech Funding Mandate (MTFM) policy and received continued support in 2024/2025. The policy defines a list of NICE-approved, cost-saving technologies that NHS commissioners and providers are mandated to fund through their existing allocations.
- Japan:
Exact Sciences achieved national health insurance coverage for their Oncotype DX Breast Recurrence Score test in September 2023, a genomic test to support oncologists in decision-making around further breast cancer treatment.
- South Korea:
VUNO Med-DeepBrain, an AI-based 3D brain MRI device for diagnosing neurodegenerative diseases, achieved national health insurance coverage in 2022, allowing radiologists to bill an additional fee for the software-assisted MRI scanning and reading.
Shrinking hurdles to payer coverage
One key reason for the still-limited payer coverage is the lack of clear clinical and economic evidence. Many AI diagnostic tools lack large-scale, peer-reviewed studies proving their clinical efficacy and cost-effectiveness compared to existing methods at the time of launch. Payers may have concerns around AI’s reliability and long-term outcomes. This is especially true for AI tools that depend on new digital biomarkers which require formal clinical validation or regulatory recognition. Generating the evidence necessary to meet payer requirements is a time-intensive process, often taking years. As evidence builds, payer concerns are expected to be gradually alleviated.
Furthermore, regulatory uncertainties and the lack of standardized HTA assessment criteria impose additional hurdles to reimbursement – but are gradually being reduced. Policymakers worldwide are catching up by providing regulatory guardrails for marketing authorization. For example, the EU AI Act, which officially came into force on August 1, 2024, is a significant step forward.
At a national level, some policy makers are taking active steps to provide payer guidance on how to assess AI-based medical devices and underlying algorithms. For instance, Spain’s Ministry of Health, in collaboration with the Agency for Health Quality and Assessment of Catalonia (AQuAS), developed a Digital Health Technology Assessment framework. The framework establishes standardized evaluation criteria and supports national and regional bodies in evidence-based decision-making. It specifically highlights key assessment domains particularly relevant for AI-based technologies, including ethical, legal, regulatory and technical aspects.
Payers are more likely to consider covering AI-enabled medical devices if they are included in clinical guidelines and thereby formally recognized as part of the standard of care. Such guideline inclusion is also crucial for driving adoption: It helps to establish provider trust in AI, raises awareness, and reduces potential reluctance to change current practices. For instance, the UK National Institute for Health and Care Excellence (NICE) recently published new guidance which recommends four AI technologies to detect fractures on X‑rays in urgent care, contingent on generating more evidence and receiving regulatory approval. Another example is the abovementioned national lung cancer screening program established in Germany.
Alternative monetization options without payer coverage
In the absence of payer coverage, tapping into healthcare provider budgets can serve as an alternative (bridging) monetization strategy. With a compelling and tailored value proposition, providers might be willing to invest in AI software without dedicated reimbursement. Convincing arguments could include enhanced workflow efficiencies, better resource allocation, improved employee satisfaction, and differentiated service offerings that attract new patients. In some instances, providers might even pass through partial costs to their customers, such as patient co-pays for AI-assisted skin cancer screening or second opinions on CT scan reading (e.g., FloyRadiology).
Key takeaways
- AI and ML are revolutionizing medical diagnostics by enhancing accuracy and early disease detection, such as cancer and cardiovascular diseases.
- While AI/ML-enabled devices have entered the market rapidly, payer coverage has struggled to keep pace. However, momentum is growing, with early success stories in subnational and temporary funding.
- Regulatory advancements, national policy initiatives, and the increasing recognition of AI, such as its inclusion in clinical guidelines, are paving the path toward broader coverage.
- Recent developments signal a positive shift in payer readiness and acceptance of AI as a cornerstone of the future of medical diagnostics.
At Simon-Kucher, we help medtech and diagnostics companies unlock the full potential of their AI/ML-enabled solutions. Our service spectrum includes product development and launch activities, ranging from early need and feature identification, to evidence generation and access strategy, down to pricing and reimbursement and commercial support like value communication, sales approach, and strategic partnerships. Whether you need to demonstrate clinical and economic value, navigate payer hurdles, or optimize commercialization, our team provides data-driven strategies to accelerate growth and maximize impact.
Reach out to our specialists.
