Blog
Modernizing the CMS Innovation Center Through Technology
A policy commentary for the Niskanen Center arguing that the CMS Innovation Center can accelerate learning and improve outcomes by embedding modern data and technology into how payment models are tested and refined.
The CMS Innovation Center was created to test new payment and delivery models, yet its learning cycles remain slow and administratively constrained. In this commentary for the Niskanen Center, I argue that modern data infrastructure and technology-enabled iteration can shorten feedback loops, improve evaluation, and allow payment models to evolve more quickly in response to evidence.
Mitigating Selection in Medicare Advantage Through Reinsurance
A Health Affairs Forefront perspective examining how targeted reinsurance could reduce selection pressures in Medicare Advantage and improve the stability of risk-based payment.
Medicare Advantage markets remain shaped by selection pressures that risk adjustment alone cannot fully offset. This perspective in Health Affairs, co-authored with Zeke Emanuel and Ravi Parikh, examines how a targeted reinsurance mechanism could dampen those incentives, reduce distortions, and strengthen the financial architecture supporting risk-based payment.
The Promise and Challenge of Value-Based Payment
A policy analysis in JAMA Internal Medicine discussing why value-based payment holds promise for better quality and lower costs, and why current implementations struggle without stronger alignment of incentives and systemic architecture.
An opinion piece in JAMA Internal Medicine co-authored with Zeke Emanuel and Amol Navathe that outlines both the potential benefits of value-based payment and the systemic obstacles that have slowed its adoption, arguing for more deliberate design choices in payment reform.
Overcoming Implementation Barriers in Value-Based Payment
A policy perspective in Health Affairs Forefront on the implementation barriers holding back value-based payment adoption, with recommended strategies for overcoming these obstacles through better financial infrastructure and operational alignment.
The second part of a Health Affairs viewpoint with Zeke Emanuel and Amol Navathe that discusses practical and organizational barriers to widespread adoption of value-based payment, including lack of financial visibility and operational readiness, and outlines pragmatic solutions.
Fixing Design Flaws in Value-Based Payment Models
A structural critique published in Health Affairs Forefront of current value-based payment models, examining how design flaws, from risk sharing to financial visibility, undermine effectiveness and what reforms are needed to fix them.
The first part of a two-part Health Affairs viewpoint with Zeke Emanuel and Amol Navathe identifying fundamental design flaws in current value-based payment programs, such as misaligned risk sharing and inadequate financial feedback mechanisms, and proposing structural improvements.
How Primary Care Visit Cadence Affects Hospital Admissions
Evidence from joint work with VillageMD published in AJMC showing that higher primary care visit cadence is linked with lower hospitalization rates among high-risk patients, highlighting how care delivery patterns interact with outcomes in value-based frameworks.
Work with VillageMD published in AJMC showing that higher primary care visit frequency is associated with reduced hospital admissions for high-risk patients, with implications for care delivery models and risk-based payment design.
From Healthcare AI to Payment Design
Two essays from 2023 capturing a transition from developing healthcare AI models to applying analytics to payment design and incentive reform.
In 2023, after years leading healthcare data science teams, I began writing more directly about a constraint that had become unavoidable: models were rarely the constraint. Payment design and incentives were.Payment design and incentives were.
These essays reflect that shift: from building healthcare AI systems to examining the economic architecture that determines whether those systems create durable change.
How My Journey in Health Data Science Led Me to Value-Based Care Economics (2023)
A reflection on how deploying analytics in real healthcare settings revealed that financial structure and incentive alignment often matter more than model performance.Introducing: Risky Judgment (2023)
Using Civil War finance as an analogy, this essay argues that value-based care still lacks the basic financial architecture needed for stability — and that data and analytics can help design and govern that missing infrastructure.
Early Essays on Building AI Products in Healthcare (2018–2020)
Early essays from my time leading healthcare data science teams, focused on turning machine learning work into real products, not just models or blog posts.
Between 2018 and 2020, I was leading data science teams in healthcare and spent a lot of time thinking about a problem that many organizations faced at the time: how do you turn machine learning work into real products, not just models or blog posts?
These essays reflect that period. They focus on integrating data science into product development, building feedback loops, and avoiding the trap of treating AI as a standalone technical function. The domain was healthcare, but the underlying question — how intelligence becomes operational — continues to shape my work today.
• What I’ve Learned Working with 12 Machine Learning Startups (2019)
Lessons from advising early ML companies on why technical talent alone doesn’t create value — and how product alignment and integration determine outcomes.
• How to Build an AI Moat (2019)
An argument that defensibility in AI comes from product-embedded data loops, not from isolated model performance.
• 4 Product-Driven Steps to an AI Roadmap (2020)
A practical framework for ensuring AI efforts start with product value and workflow integration, rather than technical novelty.