{
  "abstract": "Background/Objectives: Large language model (LLM) pipelines increasingly convert free-text meal descriptions into nutrient values, yet reference-standard validation is rarely feasible in production. We asked whether the controls such pipelines conventionally pass suffice and built a reference-standard-free battery any pipeline can run on its output. Methods: We audited 13,073 meals logged by 513 adults through a WhatsApp conversational agent in rural Colombia (9 February 2026 to 6 August 2026). The pipeline is a two-stage cascade of commercial LLMs from different providers: a vision model renders photographs as text, while a text model maps these to nutrients. Six pre-specified tests were applied, including test–retest ICC(1,1) over recurring strings, stratified by modality because 48.3% of records carried a photograph. Results: The pipeline passed every conventional control (Atwater-reconstructed energy r = 0.9951; median relative error 0.0000) but failed reproducibility. Across 191 recurring strings (944 meals), energy ICC point estimates were 0.7061 (95% CI 0.5419–0.8217) text only and 0.5258 (0.3778–0.6513) with photographs; none reached 0.75. Modality is a design control, not an established effect, and the coefficient depends on which inputs recur, so it is not intrinsic. Under the strictest identity available—byte-identical text, no photograph—ICC was 0.6944. Intermediate-text inspection exposed a failure invisible at the output: 1.44% of photograph-accompanied records carried a vision-stage artefact string versus 0.01% of text-only records (ICC 0.0911). Conclusions: Internal consistency does not validate an LLM-assisted nutrition pipeline. This design measures instability under recurring inputs without isolating its cause. Prospective test–retest reproducibility is a cheap, necessary acceptance criterion, and in a cascade, the intermediate representation must also be inspected.",
  "abstract_shown": true,
  "affiliations": [
    "Neptura, Libano 731040, Tolima, Colombia",
    "Facultad de Ciencias Humanas, Corporación Universitaria Minuto de Dios—UNIMINUTO, Bogotá D.C. 111021, Colombia",
    "Facultad de Ingeniería, Universidad Distrital Francisco José de Caldas, Bogotá D.C. 111711, Colombia"
  ],
  "article_type": "research article",
  "arxiv": null,
  "author_total": 3,
  "authors": [
    {
      "affs": [
        0
      ],
      "name": "Jhony Fernando López Torres",
      "orcid": "0009-0009-7202-2915"
    },
    {
      "affs": [
        1
      ],
      "name": "Juan Gabriel Castañeda",
      "orcid": "0000-0001-7632-7526"
    },
    {
      "affs": [
        2
      ],
      "name": "Edwin Rivas",
      "orcid": "0000-0003-2372-8056"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/4j5z9dc25p43p.bib",
  "cited_by": 1,
  "doi": "10.3390/nu18193233",
  "fulltext_url": "https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13667244/?tool=EBI",
  "id": "4j5z9dc25p43p",
  "issns": [
    "2072-6643"
  ],
  "landing_url": "https://europepmc.org/article/PMC/PMC13667244",
  "language": "eng",
  "licence": "CC-BY",
  "licence_label": "CC BY",
  "licence_source": "article",
  "licence_url": "https://creativecommons.org/licenses/by/4.0/",
  "metadata_licence": null,
  "native_subjects": [
    [
      "openalex",
      "2718|Artificial Intelligence in Healthcare and Education"
    ]
  ],
  "openalex": "W7214885811",
  "pdf_url": "https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13667244/pdf/?tool=EBI",
  "pmcid": "PMC13667244",
  "pmid": null,
  "published": "2026-09-30",
  "publisher": "MDPI AG",
  "ris": "https://research.draconitic.com/paper/4j5z9dc25p43p.ris",
  "sources": [
    {
      "first_seen": "2026-10-30T12:17:35Z",
      "native_id": "PMC:PMC13667244",
      "source": "europepmc",
      "url": "https://europepmc.org/article/PMC/PMC13667244"
    },
    {
      "first_seen": "2026-10-30T12:19:56Z",
      "native_id": "10.3390/nu18193233",
      "source": "crossref",
      "url": "https://www.mdpi.com/2072-6643/18/19/3233"
    }
  ],
  "subjects": [
    {
      "field": "medicine",
      "origin": "openalex:2718|Artificial Intelligence in Healthcare and Education",
      "subfield": "epidemiology-public-health"
    }
  ],
  "title": "Internal Consistency Does Not Validate an LLM-Assisted Nutrition Pipeline: A Reference-Standard-Free Forensic Reliability Audit of 13,073 Free-Text Meal Records",
  "title_text": "Internal Consistency Does Not Validate an LLM-Assisted Nutrition Pipeline: A Reference-Standard-Free Forensic Reliability Audit of 13,073 Free-Text Meal Records",
  "type": "peer-reviewed",
  "updated_at": "2026-10-30T12:19:56Z",
  "url": "https://research.draconitic.com/paper/4j5z9dc25p43p",
  "venue": "Nutrients",
  "versions": []
}