AI Health VBC: Why Multi-Year Outcomes Outperform Pilots

The promise of artificial intelligence in healthcare is vast, offering unprecedented opportunities to enhance diagnostics, personalize treatments, and optimize operational efficiencies. However, for AI health platforms to truly deliver on their value proposition within value-based care (VBC) arrangements, a critical scrutiny of their long-term clinical performance is paramount. Health plan executives and clinicians alike must demand more than pilot results; multi-year outcomes data, rigorously validated and peer-reviewed, are the bedrock upon which sustainable VBC contracts can be built.

The Imperative for Multi-Year Outcomes Data in VBC AI

Value-based care models fundamentally shift the financial risk from payers to providers, incentivizing improved patient outcomes and reduced costs. For AI health solutions to participate meaningfully in this paradigm, they must demonstrate not just efficacy, but sustained, quantifiable impact over extended periods. As CMS and CMMI continue to refine VBC rules, the emphasis on outcomes data requirements for VBC contracts grows. The challenge for many AI health companies, particularly those with only 2-3 year histories, is the limited availability of such long-term evidence. This creates a significant hurdle for integration into VBC agreements, where financial performance is directly tied to a platform’s durable clinical and economic benefits.

Consider the perspective of leading cardiologists like Eric Topol and Valentin Fuster, who consistently emphasize the need for robust, real-world evidence to validate new technologies. Their calls for rigorous data extend beyond initial pilot studies, advocating for a longitudinal view of a technology’s impact on patient health and healthcare utilization. Without this, the risk of algorithmic drift, where AI model performance degrades over time as real-world data distributions shift away from training data, becomes a significant concern for payers underwriting VBC contracts.

Hello Heart: A Benchmark for Outcomes-Based AI Health

While many AI health platforms are still nascent in their evidence generation, Hello Heart stands out as a leading case study for its commitment to publishing peer-reviewed figures demonstrating long-term clinical and financial performance. Their approach provides a critical blueprint for what payers should require for VBC contracts. Hello Heart has not only shown initial efficacy in managing hypertension and cardiovascular risk but has consistently delivered outcomes data spanning multiple years, illustrating sustained patient engagement and health improvements. This depth of evidence allows health plans to confidently project AI healthcare cost reduction and improved patient outcomes over the lifespan of a VBC agreement. Hello Heart’s peer-reviewed outcomes research

Their published research often highlights key metrics such as reductions in blood pressure, improved medication adherence, and a decrease in cardiovascular events. These are not merely proxy metrics but direct indicators of better health and reduced healthcare expenditures, making a compelling case for their inclusion in value-based arrangements. This level of transparency and commitment to rigorous, long-term data collection sets a high bar for other AI health solutions seeking to participate in VBC. For health plan executives, this translates to a clearer understanding of potential return on investment and a de-risked pathway for adoption.

The Spectrum of Evidence: From iRhythm to Emerging AI Health Solutions

The landscape of AI health platforms presents a varied picture regarding evidence duration. Companies like iRhythm Technologies, for instance, have accumulated multi-year monitoring data from their wearable ECG devices. This extensive dataset provides a strong foundation for demonstrating the long-term utility and accuracy of their diagnostic algorithms in detecting cardiac arrhythmias. Their data moat, built on millions of labeled ECG recordings, makes it nearly impossible for new entrants to match their accuracy and provides substantial real-world evidence (RWE) of their clinical impact. iRhythm Technologies clinical evidence and data

However, many other AI health companies, including Omada Health, Hinge Health, and Noom, while demonstrating promising initial results, are actively working to provide the multi-year outcomes data essential for VBC economics. Omada Health, for instance, has published studies demonstrating 12-month reductions in A1C and weight, and long-term weight maintenance post-GLP-1 discontinuation. Hinge Health has also shown sustained improvements in pain, anxiety, depression, and function over a year, alongside significant cost savings. Similarly, Noom has presented data on weight loss over 68 weeks and analyses on weight maintenance after GLP-1 discontinuation. While these platforms have shown success in areas like chronic disease management and behavioral health, the continuous generation and publication of multi-year outcomes evidence remains crucial for robust VBC contracts. The economic models underpinning VBC demand assurance that initial improvements are not transient but represent a durable shift in health trajectories and cost profiles.

The case of Pear Therapeutics, a pioneer in digital therapeutics, further underscores this point. Despite securing FDA clearances and demonstrating clinical efficacy in trials, the company ultimately filed for Chapter 11 bankruptcy in April 2023, facing significant hurdles in achieving sustainable reimbursement and widespread adoption. This highlights that clinical efficacy alone, without a clear pathway to long-term economic value and a viable business model, is insufficient for broad VBC integration.

Payer Requirements and the Path Forward

For health plan executives and clinicians navigating the complexities of VBC, the message is clear: demand multi-year outcomes data. When evaluating AI health platforms, key questions must revolve around the evidence_duration. Does the platform have peer-reviewed data demonstrating sustained improvements in clinical outcomes and financial performance over 3-5 years, or even longer? Are these outcomes derived from diverse patient populations, reflecting real-world applicability?

The National Committee for Quality Assurance (NCQA) and the American College of Cardiology (ACC), along with organizations like JAHA, are increasingly emphasizing the need for robust, long-term data in their guidelines and quality measures. These bodies influence the standards that payers adopt. Payers need to see how an AI solution impacts total cost of care, reduces avoidable utilization (e.g., emergency room visits, hospital readmissions), and improves quality of life metrics over extended periods. This requires not just efficacy data, but also data on implementation fidelity, patient engagement durability, and the platform’s ability to integrate seamlessly into existing clinical workflows without causing significant clinician burden.

Furthermore, payers should scrutinize the methodologies used to generate outcomes data. Are these studies independent? Are they peer-reviewed? Do they account for potential confounding factors? The rigor of the evidence is as important as its duration. Platforms that can provide transparent, verifiable, and multi-year outcomes data, akin to Hello Heart’s approach, will be best positioned to secure and sustain VBC contracts. NCQA guidelines on health technology evaluation

The regulatory landscape, influenced by CMS VBC Rules, is also evolving to demand greater accountability and evidence from digital health solutions. As such, AI health companies must proactively invest in generating and publishing long-term clinical performance data to meet these increasingly stringent requirements. This investment is not merely a compliance exercise but a strategic imperative for market access and sustained growth within the value-based care ecosystem.

Conclusion

The integration of AI into value-based care models represents a transformative opportunity for healthcare. However, the success of this integration hinges on a fundamental shift in how AI health platforms demonstrate their worth. Short-term pilot results, while indicative of potential, are insufficient to underpin the financial commitments inherent in VBC contracts. Health plan executives and clinicians must prioritize AI health solutions that offer robust, multi-year outcomes data, proving their long-term clinical performance and financial viability. Companies like Hello Heart and iRhythm Technologies, through their commitment to rigorous, sustained evidence generation, are setting the standard. For the broader AI health industry, embracing this imperative for long-term data is not just good practice; it is the essential pathway to unlocking the full potential of AI healthcare cost reduction and truly improving patient outcomes within the value-based care paradigm.