Making Everyone’s Preferences Count: A Prospect Theory Approach to Valuing Health Improvements -- by Karen Mulligan, Drishti Baid, Jason N. Doctor, Darius N. Lakdawalla

TEXT START: Health technology assessment bodies and policymakers have grappled with the issue of whose preferences – the general healthy population or the patient population – should be incorporated into traditional cost-effectiveness analysis.

The Dissection

From the supplied abstract, this is a technocratic legitimacy and measurement project. It modifies cost-effectiveness analysis with prospect theory so adaptation, probability weighting, loss aversion, and patient valuations enter the healthcare calculator. “Everyone’s preferences count” means preferences are represented in the model—not that everyone gains control over healthcare production, funding, or allocation.

The Core Fallacy

It mistakes better representation for preserved economic power. The framework may correct the undervaluation of chronic illness and disability inside healthcare assessment, but it does not address the Discontinuity Thesis’s decisive mechanics: AI’s superiority over cognitive labor (P1), the inability of institutions to preserve human-only economic domains (P2), or the collapse of economically necessary work for the majority (P3).

A more accurate utility function cannot preserve the wage-to-consumption circuit once AI severs it. It is an allocative refinement inside the system, not a defense of the system.

Hidden Assumptions

• Healthcare institutions will retain the authority and budgets to apply these estimates.
• The tax, insurance, employment, and production systems funding healthcare will remain functional.
• Reference points and adapted preferences can be measured reliably enough to guide high-stakes allocation.
• Aggregating heterogeneous preferences produces a legitimate policy rule.
• Human preferences will remain central to resource allocation after AI concentrates productive control among Sovereigns.
• Inclusion in a valuation formula is an adequate substitute for material agency and bargaining power.

Social Function

Primarily partial truth and transition management, with a layer of prestige signaling. The paper addresses a genuine defect in conventional health valuation: static models can undervalue lives shaped by chronic illness or disability. But its sophistication also helps the existing policy apparatus appear more responsive while leaving its economic substrate untouched.

It is not pure copium. It is a better dashboard for a vehicle whose engine is being replaced.

The Verdict

Technically useful, systemically insufficient. The paper can improve healthcare decisions during the institutional lag, but it cannot prevent the post-WWII order from dying when AI breaks mass productive participation. Preferences may be counted with greater precision while the people expressing them lose income, leverage, and control. Under the Discontinuity Thesis, this is calibration of the welfare ledger—not a survival mechanism.