{
  "abstract": "BackgroundThe American Society of Anesthesiologists Physical Status (ASA-PS) class underpins perioperative risk communication, but its assignment is subjective. Generative large language models (LLMs) are slow and costly, and LoRA fine-tuning needs site-specific labels and a separately maintained adapter for every site. Non-generative decision models return option probabilities in one forward pass, but their comparison with LoRA in anesthesia, and the number of local labels they need, is unstudied.\n\nObjectiveTo compare, on one open base model, generative zero-shot prompting, a non-generative readout, the readout with cutpoint calibration on a few local labels, and task LoRA for four-class ASA-PS (I, II, III, IV-V) in accuracy, speed and label requirement.\n\nMethodsRetrospective study, not preregistered. MOVER (57,473 operations) was the development source, the local hospitals 2025-2026 operations (n = 2,753) the temporally external validation set, and VitalDB (n = 6,242) a second site. The base model was Qwen3.5-9B; input was preoperative structured variables (T1) or T1 plus diagnosis text (T2). The primary metric was quadratic-weighted {kappa} against each sites recorded ASA class, with patient-level cluster-bootstrap intervals; the non-inferiority margin was 0.05. Cost means tokens and GPU time.\n\nResultsIn the local validation set the readout reached {kappa} = 0.460 versus 0.227 for generative zero-shot prompting with a requested rationale (difference +0.234, 95\n\nConclusionsThe non-generative readout needs no training and ran about 64 times faster than generative prompting with a rationale. A few hundred local labels bring it close to LoRA in {kappa} without training and storing a per-site adapter, although {kappa} of 0.52-0.58 is not evidence of human-level performance. LoRA keeps a higher {kappa} once several hundred or more labels are available, and higher exact agreement. Non-inferiority holds for the prespecified analyses and sits at the margin in stricter sensitivity analyses, so it should be read as conditional.",
  "abstract_shown": true,
  "affiliations": [
    "People's Hospital of Xinjiang Uygur Autonomous Region"
  ],
  "article_type": "PUBLISHAHEADOFPRINT",
  "arxiv": null,
  "author_total": 5,
  "authors": [
    {
      "name": "Y. Shi"
    },
    {
      "name": "Y. Wang"
    },
    {
      "name": "A. Simayi"
    },
    {
      "name": "L. Qu"
    },
    {
      "affs": [
        0
      ],
      "name": "G. Xu"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/798b3aq35e4d8.bib",
  "cited_by": 0,
  "doi": "10.64898/2026.10.20.26366028",
  "fulltext_url": "https://www.medrxiv.org/content/10.64898/2026.10.20.26366028v1.full",
  "id": "798b3aq35e4d8",
  "issns": [],
  "landing_url": "https://www.medrxiv.org/content/10.64898/2026.10.20.26366028v1",
  "language": null,
  "licence": "CC-BY-NC-ND",
  "licence_label": "CC BY-NC-ND",
  "licence_source": "article",
  "licence_url": "https://creativecommons.org/licenses/by-nc-nd/4.0/",
  "metadata_licence": null,
  "native_subjects": [
    [
      "medrxiv",
      "anesthesia"
    ],
    [
      "openalex",
      "1702|Machine Learning in Healthcare"
    ]
  ],
  "openalex": "W7223216249",
  "pdf_url": "https://www.medrxiv.org/content/10.64898/2026.10.20.26366028v1.full.pdf",
  "pmcid": null,
  "pmid": null,
  "published": "2026-10-20",
  "publisher": null,
  "ris": "https://research.draconitic.com/paper/798b3aq35e4d8.ris",
  "sources": [
    {
      "first_seen": "2026-10-30T12:21:30Z",
      "native_id": "10.64898/2026.10.20.26366028",
      "source": "medrxiv",
      "url": "https://www.medrxiv.org/content/10.64898/2026.10.20.26366028v1"
    }
  ],
  "subjects": [
    {
      "field": "medicine",
      "origin": "medrxiv:anesthesia",
      "subfield": "surgery"
    }
  ],
  "title": "Non-generative readout with few-label calibration versus LoRA fine-tuning of a large language model for ASA physical status classification: accuracy, speed, cost and label efficiency in three data sources",
  "title_text": "Non-generative readout with few-label calibration versus LoRA fine-tuning of a large language model for ASA physical status classification: accuracy, speed, cost and label efficiency in three data sources",
  "type": "preprint",
  "updated_at": "2026-10-30T12:21:30Z",
  "url": "https://research.draconitic.com/paper/798b3aq35e4d8",
  "venue": "medRxiv",
  "versions": []
}