{
  "abstract": "Objectives: To assess whether estimated due dates based on Naegele’s rule can be improved by incorporating routinely collected maternal characteristics into a non-parametric machine learning model, particularly in settings where ultrasound is unavailable.\n\nDesign: Population-based retrospective cohort study; development and internal validation of a machine-learning prediction model.\n\nSetting: Nationwide register data from Sweden.\n\nParticipants: 603 861 singleton pregnancies recorded in the Swedish Medical Birth Register between 2008 and 2017. We restricted the sample to spontaneous singleton births with complete information on last menstrual period, ultrasound dating and maternal characteristics and further limited the primary analysis to pregnancies in which both Naegele’s rule and ultrasound-based gestational age at delivery fell between 37 and 42 completed weeks.\n\nPrimary and secondary outcome measures: The primary outcome was prediction accuracy of the date of delivery. Performance was evaluated using exact-day accuracy, accuracy within ±1 and ±7 days, mean squared error and root mean squared error. Machine-learning predictions were compared with those obtained using Naegele’s rule and ultrasound-based dating under identical inclusion criteria.\n\nResults: The machine-learning model improved prediction accuracy over Naegele’s rule across all evaluated measures, while ultrasound-based dating remained more accurate overall. The model reduced between 25% and 45% of the accuracy gap between Naegele’s rule and ultrasound. Subgroup-specific models provided little additional improvement. Among second-time mothers, however, adding gestational duration from the previous pregnancy yielded markedly greater gains and produced prediction accuracy comparable to, and on some measures slightly better than, ultrasound.\n\nConclusions: Incorporating maternal characteristics into a regression-forest model improved due date prediction compared with Naegele’s rule. These improvements may be relevant in settings where ultrasound is unavailable but the routinely collected information and basic digital infrastructure required to apply the model are available. The approach is not intended to replace ultrasound-based dating in routine clinical care. Further research is needed to assess performance in other populations and to evaluate potential implications for clinical decision-making. The model was developed and internally validated in Swedish register data and requires external validation before use as a decision support tool elsewhere.",
  "abstract_shown": true,
  "affiliations": [
    "Department of Economics, Uppsala University, Uppsala, Sweden",
    "Institute for Housing and Urban Research, Uppsala University, Uppsala, Sweden"
  ],
  "article_type": "research article",
  "arxiv": null,
  "author_total": 3,
  "authors": [
    {
      "affs": [
        0
      ],
      "name": "Gilson Dutra"
    },
    {
      "affs": [
        0
      ],
      "name": "Mikael Elinder"
    },
    {
      "affs": [
        1
      ],
      "name": "Oscar Erixson",
      "orcid": "0000-0002-3929-9906"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/8trf7rt8yktg0.bib",
  "cited_by": 0,
  "doi": "10.1136/bmjopen-2026-120970",
  "fulltext_url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC13711609/",
  "id": "8trf7rt8yktg0",
  "issns": [
    "2044-6055",
    "2044-6055"
  ],
  "landing_url": "https://bmjopen.bmj.com/lookup/doi/10.1136/bmjopen-2026-120970",
  "language": "en",
  "licence": "CC-BY-4.0",
  "licence_label": "CC BY 4.0",
  "licence_source": "article",
  "licence_url": "https://creativecommons.org/licenses/by/4.0/",
  "metadata_licence": null,
  "native_subjects": [
    [
      "heading",
      "1506"
    ],
    [
      "heading",
      "1692"
    ],
    [
      "heading",
      "Epidemiology"
    ],
    [
      "heading",
      "Original Research"
    ],
    [
      "openalex",
      "3605|Artificial Intelligence in Healthcare"
    ]
  ],
  "openalex": "W7223376381",
  "pdf_url": "https://bmjopen.bmj.com/content/bmjopen/16/10/e120970.full.pdf",
  "pmcid": "PMC13711609",
  "pmid": "42919588",
  "published": "2026-10-23",
  "publisher": "BMJ",
  "ris": "https://research.draconitic.com/paper/8trf7rt8yktg0.ris",
  "sources": [
    {
      "first_seen": "2026-10-30T12:19:44Z",
      "native_id": "10.1136/bmjopen-2026-120970",
      "source": "crossref",
      "url": "https://bmjopen.bmj.com/lookup/doi/10.1136/bmjopen-2026-120970"
    },
    {
      "first_seen": "2026-10-31T12:22:21Z",
      "native_id": "PMC13711609",
      "source": "pmc",
      "url": "https://doi.org/10.1136/bmjopen-2026-120970"
    }
  ],
  "subjects": [
    {
      "field": "medicine",
      "origin": "heading:Epidemiology",
      "subfield": "epidemiology-public-health"
    }
  ],
  "title": "Can maternal characteristics improve due date prediction without ultrasound? Development and internal validation of a machine-learning prediction model in 600 000 pregnancies in Sweden",
  "title_text": "Can maternal characteristics improve due date prediction without ultrasound? Development and internal validation of a machine-learning prediction model in 600 000 pregnancies in Sweden",
  "type": "peer-reviewed",
  "updated_at": "2026-10-31T12:22:21Z",
  "url": "https://research.draconitic.com/paper/8trf7rt8yktg0",
  "venue": "BMJ Open",
  "versions": []
}