Medicare has denied approximately $16 million in radiology artificial intelligence (AI) claims over a five-year period, according to a recent study published in the Journal of the American College of Radiology. This research highlights significant barriers to the integration of AI technologies in clinical practice, despite their potential to enhance diagnostic accuracy and efficiency.
Study Overview
The study analyzed 83,392 AI-related radiology services submitted to Medicare between 2018 and 2023. Of these, 39,535 claims (47%) were reimbursed, totaling $8 million, while 53,857 claims (53%) were denied, amounting to $16.4 million in rejected payments. This indicates a denial rate of approximately 53%, underscoring challenges in the reimbursement process for AI-driven services.
Denial Patterns and Service Utilization
The analysis revealed that certain AI services experienced higher denial rates. For instance, the LiverMultiScan, a noninvasive MRI tool for assessing liver disease, had a denial rate of 98.2%, reflecting $900,000 in potential payments. Similarly, the Low Ejection Fraction AI-ECG faced a 98.2% denial rate, amounting to $2.7 million in denied claims. In contrast, the fractional flow reserve derived from CT, a widely utilized AI service, had a lower denial rate of 18.5%, corresponding to $5.5 million in denied payments.
Implications for Healthcare Providers
The high denial rates for AI services may discourage healthcare providers from adopting these technologies, potentially hindering advancements in patient care. The study suggests that the current reimbursement model, based on Current Procedural Terminology (CPT) codes, may not be well-suited to accommodate the rapid evolution of AI technologies. The proliferation of AI services could necessitate a large number of codes, and the swift pace of technological advancements may render existing codes obsolete, leading to administrative burdens and increased denial rates.
Recommendations for Policy Reform
To address these challenges, the study authors propose several strategies:
- Dynamic Coding Systems: Implementing flexible coding frameworks that can adapt to the evolving landscape of AI technologies.
- Bundled Payment Models: Developing reimbursement models that group similar AI services, reducing the complexity of billing and potentially increasing acceptance rates.
- Value-Based Care Incentives: Aligning reimbursement strategies with value-based care principles, which emphasize patient outcomes and cost-efficiency, to encourage the adoption of AI technologies.
Conclusion
The denial of $16 million in radiology AI claims by Medicare underscores the need for a reevaluation of reimbursement policies to facilitate the integration of AI into clinical practice. By adopting more flexible and outcome-oriented reimbursement models, policymakers can support the widespread adoption of AI technologies, ultimately enhancing patient care and the efficiency of healthcare delivery.
Source: Medicare has denied $16M worth of radiology artificial intelligence claims
