---
description: Nearly every AI application in healthcare optimizes the 10-20% clinical side while 80-90% of outcomes are driven by non-clinical factors so making sick care more efficient produces more sick care not better health
type: claim
domain: health
created: 2026-02-23
source: "Devoted Health AI Overview Memo, 2026"
confidence: likely
---

# healthcare AI creates a Jevons paradox because adding capacity to sick care induces more demand for sick care

The entire healthcare system was built for infectious disease -- designed to give you something or do something to you. But the modern burden is chronic disease, lifestyle, and behavior. Since [[medical care explains only 10-20 percent of health outcomes because behavioral social and genetic factors dominate as four independent methodologies confirm]], 80-90% of what determines health happens outside the clinical encounter: adherence, exercise, food, sleep, coordination.

Yet nearly every AI application in healthcare today optimizes the 10-20% clinical side -- a better diagnostic model, a faster scribe, a smarter claims tool. Even perfected, these cannot solve the fundamental problem. This is the Jevons paradox applied to medicine: adding capacity to the sick care system induces more demand for sick care. A faster diagnostic tool finds more conditions to treat. A better scribe enables more patient visits. A smarter claims processor approves more procedures. Each makes the existing system more efficient at doing what it already does -- treating sickness -- rather than changing what the system does.

The scale of investment flowing into this paradox is enormous. OpenAI reports 230 million users asking health questions weekly and committed $25 billion to a health foundation. Microsoft spent $19.7 billion acquiring Nuance for clinical AI. Google's Med-Gemini scores 91.1% on medical licensing exams. But these companies are building AI engines -- better models, better clinical NLP, better benchmarks. They are not building the integrated delivery system that turns AI capability into health outcomes. The J.P. Morgan 2026 Healthcare Conference warned about the "ChatGPT wrapper" problem -- AI tools layered onto broken workflows that fail to change outcomes.

The structural insight: you cannot solve a system problem with a component optimization. Healthcare needs system-level change -- rebuilding the entire workflow around coordinated care that addresses the 80-90% non-clinical determinants. Since [[healthcare is a complex adaptive system requiring simple enabling rules not complicated management because standardized processes erode the clinical autonomy needed for value creation]], AI must be embedded in the care delivery system, not bolted onto it. The digital health venture funding collapse tells this story: down 65% from the 2021 peak, with over $150 billion in unicorn valuation destroyed (Babylon, Olive AI, Pear Therapeutics) -- all point solutions that created new choke points rather than solving the system problem.

The exception proves the rule: companies that control the full stack -- from insurance through care delivery through technology -- can direct AI at the 80-90% because they have the data, the incentives, and the operational reach to change behavior, not just treat symptoms. Since [[the atoms-to-bits spectrum positions industries between defensible-but-linear and scalable-but-commoditizable with the sweet spot where physical data generation feeds software that scales independently]], the defensible position in healthcare AI is the full-stack operating system, not the AI engine.

---

Relevant Notes:
- [[medical care explains only 10-20 percent of health outcomes because behavioral social and genetic factors dominate as four independent methodologies confirm]] -- the foundational evidence that clinical care is only 10-20% of outcomes
- [[healthcare is a complex adaptive system requiring simple enabling rules not complicated management because standardized processes erode the clinical autonomy needed for value creation]] -- healthcare requires system change, not component optimization
- [[prescription digital therapeutics failed as a business model because FDA clearance creates regulatory cost without the pricing power that justifies it for near-zero marginal cost software]] -- point solutions fail in healthcare because regulatory cost exceeds pricing power
- [[healthcares defensible layer is where atoms become bits because physical-to-digital conversion generates the data that powers AI care while building patient trust that software alone cannot create]] -- the defensible position is at the atoms-to-bits conversion, not in AI engines alone
- [[performance overshooting creates a vacuum for good-enough alternatives when products exceed what mainstream customers need]] -- AI diagnostic accuracy already exceeds physician performance on benchmarks, yet outcomes barely improve, suggesting the bottleneck is not accuracy but system integration

Topics:
- health and wellness