Thursday, January 16, 2025

Spunky: Gottlieb and Sharfstein Debate FDA vs CMS vs Money

 How much, if at all, should the government tie regulatory approvals (FDA) to payment (CMS)?

The usual industry answer is, NOT AT ALL.

Former FDA administators or deputy administrators Scott Gottlieb and Josh Sharfstein weigh in.   Audience often chuckles.  It's an entertaining video.

One Gottlieb comment, 'I had the impression that CMS didn't usually have enough expertise in clinical trial design to give thoughtful input..." (3m20s)

Find the video here:

https://www.youtube.com/watch?v=OY1jdCIPPcQ

See other new talks of similar interest at the UCSF-Stanford "CERSI" regulatory science institute web site.   Here's one from CERSI with 4 differerent FDA commissioners.  Here's one from CERSI on FDA AI.


AI Summary:

The FDA debate featuring Scott Gottlieb, former FDA Commissioner, and Josh Sharfstein, former FDA Deputy Commissioner, centered on whether there should be a firewall between FDA regulatory policy and drug costs. Here are the key points from their arguments:


Scott Gottlieb: Pro-Firewall

  • FDA's Role and Expertise: FDA should focus on evaluating drug safety, efficacy, and risk-benefit, not on costs or reimbursement decisions. Adding payer input during the approval process risks making clinical trials longer, more complex, and expensive.
  • Challenges with CMS Involvement: Past efforts to integrate CMS into trial design were largely unsuccessful. CMS lacked the expertise in clinical trial design, leading to impractical expectations and unrealistic coverage evidence development (CED) requirements.
  • Separation of Roles: Coverage decisions are inherently political and subjective, often influenced by policymakers, which could compromise the FDA’s objective regulatory framework.
    • In contrast, specific yes/no FDA decisions landing on the Comissioner's desk "didn't ever happen in my entire experience" (6m30s)
  • Regulatory Uncertainty: Introducing cost considerations into FDA approval could create ambiguity and politicize the process, undermining trust in FDA’s standards.
  • Alternative Mechanisms: Real-world data collection post-market and payer-driven negotiations are better suited to address cost concerns without impacting the regulatory process.

Josh Sharfstein: Anti-Firewall

  • FDA as a Public Health Agency: Sharfstein argued that high drug costs hinder public access to medications, contradicting the FDA’s mission to promote health and prevent illness.
  • Policy Proposals Integrating Cost:
    1. Encouraging Generic Competition: FDA could prioritize approving generics for high-cost drugs, improve standards for complex generics, and act against price-gouging.
    2. Addressing Patent Gaming: More transparency and regulatory action against anti-competitive practices that delay generics could reduce costs.
    3. Reforming Accelerated Approvals: Strengthening post-approval study requirements and capping payments for certain drugs could ensure value without overburdening patients or Medicare.
    4. Transparency with Payers: Sharing FDA data could enable value-based pricing and broader patient access.
  • Partnership with CMS: While acknowledging challenges, Sharfstein advocated for voluntary collaboration between FDA and CMS, particularly in early reimbursement decisions for devices or drugs, to streamline processes and reduce inefficiencies.

Debate Outcome

The audience vote was close, slightly favoring the anti-firewall position (51%). While Gottlieb defended the current separation as essential for preserving FDA’s integrity, Sharfstein argued for innovative approaches to integrate cost considerations while maintaining FDA’s focus on public health.


###

BONUS

Here's an AI Summary of the AI talk cited above.  

Russ Altman moderates; Rahul Adora from OpenAI.  Panelists, Mark Taisey, Amgen, Janet Woodcock, fmr FDA, Liang Zhao, UCSF.

TL;DR Summary

A panel discussion at the UCSF-Stanford CERSI explored the integration of generative AI into regulatory science. Participants included industry leaders, academic researchers, and former FDA officials. The focus was on understanding the benefits, risks, and future applications of AI in regulatory workflows. Key themes included:

  1. Current Uses of AI:

    • Automating regulatory processes, such as creating clinical study reports and quality summaries, reducing timelines significantly (e.g., from 26 weeks to 8-12 weeks for Amgen’s applications).
    • Exploring co-pilot tools for FDA reviewers to enhance knowledge management, consistency, and efficiency.
  2. Challenges and Risks:

    • Concerns over data privacy and security, especially when AI relies on external servers or cloud-based platforms.
    • Risks of AI hallucination and compounded errors in autonomous workflows, requiring rigorous quality checks.
    • Barriers to structured data adoption, including reliance on legacy systems and resistance to change within regulatory agencies.
  3. Opportunities:

    • Using AI to reduce human workload on "busy work" and streamline decision-making.
    • Enhancing global regulatory collaboration via cloud-based platforms like Accumulus.
    • Creating structured data systems to eliminate redundant documentation and expedite drug approvals.
  4. Future Vision:

    • Transitioning from text-heavy documentation to structured data systems with real-time cloud integration.
    • Leveraging AI for strategic innovations while ensuring human oversight for critical judgments.
    • Expanding public data availability to foster innovation while addressing confidentiality issues.

Takeaway:

AI offers transformative potential for regulatory science, but success requires addressing privacy concerns, ensuring accuracy, and fostering collaboration between stakeholders. A balance between automation and human judgment remains essential