NEJM AI study: LLM-recommended care costs up to 10x more than physician orders
dvir_a · x · 2026-10-21
A NEJM AI study tests whether LLM-recommended clinical care costs more than physician-recommended care, and whether prompting changes that. Across 200 cases split into presentation vs. A&P, 24 AI systems under 3 prompt styles generated plans, whose orders were priced via Medicare CPT fees. An example case shows a physician ordering $0 of tests for bronchiolitis while the default-prompt AI ordered $454, dropping to $292 with a cost-aware prompt and $35 with a parsimonious-but-safe prompt — prompt wording dramatically shifts AI-recommended care costs.
More from Research
- UW × NVIDIA paper: Compositional RL trains a 27B LLM to be the best teammate in multi-model systems — shangbinfeng · 2026-10-21
- Solo RL can make LLMs worse teammates: Shapley-based Compositional RL trains for collaboration — shangbinfeng · 2026-10-21
- Hypothesis Dreamer Agent: A Weekly AI Prompt for Generating Research Ideas — nlarusstone · 2026-10-21
- JAX speedrun PR trains to loss 4.179 in 6.76s on 8×H100 — 2.27× faster than prior best — RobertTLange · 2026-10-21
- Stanford researchers slam GoodfireAI for overhyped interpretability claims as superposition hypothesis faces scrutiny — stanfordnlp · 2026-10-21
- SherpaLabs launches RL environments for finance, coding, quant; top models pass only ~6% on its finance benchmark — ycombinator · 2026-10-21