Navigating the value proposition of artificial intelligence in cardiovascular disease prevention

Abstract

Artificial intelligence (AI) is reshaping healthcare, influencing everything from administrative workflows to direct patient care. It holds promise in addressing the leading cause of death in the United States and a significant driver of costs – cardiovascular disease (CVD). Traditional reimbursement models are slow to incorporate AI. This article explores alternative financial incentives for AI in CVD prevention in fee-for-service (FFS) and value-based care models across four domains – risk prediction, diagnostics, imaging, clinical decision support, and administrative strategies – where AI may provide indirect revenue generation, cost savings, and efficiency gains. Under FFS, AI can enhance revenue by driving appropriate healthcare utilization, improving billing accuracy, and streamlining administrative workflows. In value-based models, AI aligns with incentives to prevent disease progression, reduce hospitalizations, and optimize shared savings. While AI-powered tools offer a compelling financial value proposition in cardiovascular prevention, their real-world adoption and impact will depend on successful clinical validation and seamless integration into existing workflows. The future of AI in cardiovascular care depends on a shift in reimbursement models, regulatory adaptation, and continued evidence generation demonstrating cost-effectiveness and improved outcomes. As healthcare transitions toward value-based care, AI has the potential to be a catalyst for better prevention and long-term cost savings, but only if its business case is strategically developed and implemented.

Introduction

Artificial intelligence (AI) is reshaping healthcare, influencing everything from administrative workflows to direct patient care. It holds promise in addressing the leading cause of death in the United States – cardiovascular disease (CVD). ¹ AI can enhance CVD prevention across four domains – risk prediction, diagnostics, imaging, clinical decision support, and administrative strategies.^2,3 Modifiable risk factors account for 50% of cardiovascular disease and 14% of mortality in North America, highlighting the opportunity for intervention. ⁴ AI tools are already being developed and studied to target major CVD risk factors, though evidence on clinical outcomes remains limited. ⁵

Despite AI's potential, it's impact on CVD outcomes at scale remains limited, as adoption is still in the early stages.^2,6 A key barrier is the underdeveloped business case for AI in CVD prevention, particularly within traditional reimbursement pathways such as current procedural terminology (CPT) codes. There is currently only one reimbursed CPT code covering AI-devices, with 31 additional Category III codes, meaning they are not reimbursed and for data-collection purposes. ⁶ While some AI technologies have secured reimbursement through the new technology add-on payment (NTAP), this pathway covers only seven technologies. ⁷

Traditional reimbursement models are slow to incorporate AI. This article explores alternative financial incentives for AI in CVD prevention in fee-for-service (FFS) and value-based care (VBC) models, where AI may provide indirect revenue generation, cost savings, and efficiency gains.

Risk-prediction

Fee-for-service

AI-driven risk prediction has not been directly reimbursed through CPT or NTAP codes. ⁷ AI offers an opportunity to expand risk-prediction capabilities and indirectly generate revenue in FFS models by driving utilization of covered services. AI-powered tools can identify high-risk patients or those missing validated testing, increasing the uptake of billable services. For instance, lipid control is suboptimal – even among patients with established atherosclerotic heart disease (ASCVD) achieving target levels is challenging.^8,9 An AI-based tool could flag such patients, prompting billable visits focused on cholesterol management and medication optimization. In this way, AI-driven risk prediction can indirectly enhance revenue by facilitating appropriate healthcare utilization.

Value-based care

AI-driven risk-prediction is well aligned with VBC models as financial incentives prioritize disease prevention over reactive treatment. Numerous AI-powered models are currently being tested to predict hospitalizations with the hope that earlier intervention in high-risk patients would prevent costly admissions.^10–12 AI can identify gaps in recommended care – such as uncontrolled blood pressure, cholesterol, or hemoglobin A1c levels – allowing for timely intervention. In a VBC model, this approach improves quality scores, prevents disease progression, and reduces long-term costs. By closing these care gaps, AI-driven prevention strategies can enhance shared savings within value-based contracts.

Diagnostics

Fee-for-service

AI has seen its largest implementation in diagnostic imaging with the majority of FDA-approved AI-enabled medical devices falling under the category of radiology. ¹³ Reimbursed AI-powered cardiovascular solutions are primarily focused on acute care rather than prevention. ⁷ One notable prevention application is AI-driven coronary artery calcium (CAC) scoring, where AI can detect and quantify CAC on routine chest computer tomography (CT) scans. ¹⁴ While this service is not directly reimbursed, it can help identify high-risk patients and trigger appropriate healthcare utilization and preventive interventions. AI can also improve diagnostic efficiency by improving image acquisition, enhancing image interpretation, increasing volume, and reducing patient wait times. A small study on AI inputted measurements in echocardiograms showed promising results to reduce report generation time, though results varied based on image quality and case complexity. ¹⁵

Value-based care

AI-powered tools can support earlier detection and interventions, which aligns well with VBC models that emphasize prevention. Identifying at-risk individuals with subclinical cardiovascular disease, such as through automated CAC scoring, can facilitate timely preventive interventions, reducing the likelihood of costly downstream events. Additionally, improving diagnostic image efficiency can free up clinical resources, reduce costs, and optimize workflows, further enhancing the financial sustainability of VBC models.

Clinical decision support (CDS)

Fee-for-service

AI-powered CDS tools have the potential to enhance care delivery if they are seamlessly integrated into clinical workflows. These tools can improve adherence to guideline-directed medical therapies and quality measures. While CDS tools are unlikely to generate direct revenue in FFS models, they can increase appropriate healthcare utilization. By identifying patients who require medication optimization or follow-up care, AI-driven CDS can prompt billable visits, indirectly contributing to revenue while improving patient management and outcomes.

Value-based care

In VBC models, AI-powered CDS aligns well with incentives to reduce long-term cardiovascular events. By optimizing guideline-directed medical therapy, these tools can prevent complications, enhance adherence, and improve overall patient outcomes. Effective CDS implementation can lead to better quality scores and higher shared savings in value-based contracts, making them a valuable tool for both clinicians and healthcare organizations.

Administrative applications

Fee-for-service

Non-clinical AI applications are an expanding area with benefits extending beyond cardiovascular prevention. These tools provide financial advantages and are often easier to implement since they do not directly influence patient care. AI-driven solutions can streamline administrative processes, enhance clinical efficiency, and improve billing accuracy. AI ambient scribes are gaining traction, with early studies indicating a reduction in administrative workload for providers. ¹⁶ AI-driven scheduling tools can optimize appointment management, reduce patient wait times, and allow for better staff allocation. ¹⁷ There is a financial burden tied to time spent on billing and coding risk, where AI has the potential to enhance revenue cycle management. ¹⁸ By ensuring comorbid conditions are properly documented and identifying undiagnosed risk factors, AI-powered tools can increase appropriate reimbursement.^18,19 AI has the potential to significantly reduce administrative overhead and optimize revenue cycle management.

Value-based care

Similar to its role in FFS models, AI in VBC can enhance financial performance by reducing administrative burdens, optimizing workflow efficiency, and improving billing accuracy. However, in VBC, these tools have an additional role in risk stratification and reimbursement optimization. AI-powered tools can more accurately capture a patient's risk adjustment factor (RAF) score, which is crucial for determining reimbursement and evaluating quality of care in VBC models. The Center for Medicare & Medicaid Services (CMS) utilizes hierarchical condition categories coding, which is based on ICD-10 diagnosis, plus demographic information to calculate RAF scores. ²⁰ AI-driven tools can ensure more precise documentation, leading to more accurate reimbursement and improved quality metrics. By refining risk assessment and optimizing documentation, AI can enhance reimbursement accuracy while supporting high-quality, value-driven care.

Considerations

While AI-powered tools offer a compelling financial value proposition in cardiovascular prevention, their real-world adoption and impact will depend on successful clinical validation and seamless integration into existing workflows. Validation of such tools will need to focus on capturing gains in efficiency and/or improvement in patient outcomes that outweigh financial investments. Many of these technologies require substantial upfront investment, which must be balanced against long-term financial benefits. Additionally, AI applications must be secure, continuously monitored to prevent the introduction and propagation of bias, and designed for interoperability within evolving healthcare systems. Ensuring these tools align with clinical needs and regulatory requirements will be critical to their widespread implementation and success.

Conclusion

AI holds immense promise in transforming cardiovascular prevention, yet its widespread adoption hinges on a viable financial model. While AI applications in diagnostics have seen some traction in reimbursement, the full potential of AI in risk prediction, clinical decision support, and administrative optimization remains untapped. Under FFS, AI can enhance revenue by driving appropriate healthcare utilization, improving billing accuracy, and streamlining administrative workflows. In VBC models, AI aligns with incentives to prevent disease progression, reduce hospitalizations, and optimize shared savings. The future of AI in cardiovascular care will depend on continued evidence generation demonstrating cost-effectiveness and improved outcomes.