Presentation (CTI)

OBJECTIVES: To develop an integrated simulation framework that formalizes a previously unaddressed question: under what conditions does AI-enabled cost reduction create viable healthcare partnership corridors under bundled payments?
METHODS: The framework integrates three components.
First, constant elasticity of substitution (CES) production functions model human-AI labor substitution in non-surgical workflows, with elasticity parameter σ governing substitutability.
Second, Monte Carlo simulation estimates cost distributions for specified AI share levels, incorporating labor savings, AI operating costs, supervision requirements, and potential error-related costs, recognizing that AI may increase rather than decrease total costs in some configurations.
Third, risk-adjusted corridor bounds determine feasibility: Floor = E[C(s)] + λ·CVaR_α[C(s)] + π_min must fall below Ceiling = B − π_H − χ, where s is AI share, λ is risk aversion, α is CVaR confidence level, and π terms represent profit requirements.
The framework accepts scenario inputs and outputs corridor feasibility across the AI share spectrum.
RESULTS: Framework demonstration revealed three behavioral patterns.
First, AI share affects both expected cost and variance through competing mechanisms; automation reduces labor costs, while supervision and error risk can increase costs, with the net effect determining corridor viability.
Second, corridor feasibility exhibits threshold behavior: below scenario-specific minimum AI share, no feasible corridor exists; above scenario-specific maximum, oversight costs eliminate corridors.
Third, preliminary sensitivity analysis suggests corridor width is most sensitive to cost variance and risk aversion (λ), moderately sensitive to substitution elasticity (σ). Across tested scenarios, feasibility thresholds ranged from 35% to 60% AI share, with some high-oversight scenarios showing no feasible AI configuration.
CONCLUSIONS: The framework provides a diagnostic methodology for assessing whether AI capabilities create or destroy viable partnership conditions. By revealing scenarios where AI narrows or eliminates corridors, it offers hospitals a rigorous approach for evaluating AI tools against organization-specific feasibility requirements.