Cited by 4Dental Health and Care UtilizationOral microbiology and periodontitis researchReliability and Agreement in Measurement

Calibration and reliability testing of a novel asynchronous photographic plaque scoring system in young children

Open access
Avenetti, David M., Martin, Molly A., Gansky, Stuart A., Ramos‐Gomez, Francisco J., Hyde, Susan, Van Horn, Rebecca, Jue, Bonnie, Rosales, Genesis F. +2 more·2023·Journal of Public Health Dentistry

Abstract

OBJECTIVES: The Simplified Oral Hygiene Index for Maxillary Incisors (OHI-MIS) is a novel plaque scoring system adapted for young children. This study describes calibration training and testing used to establish the inter- and intra-rater reliability for OHI-MIS measured from clinical photographs. METHODS: Two raters from the Coordinated Oral Health Promotion Chicago (CO-OP) and one from the Behavioral EConomics for Oral health iNnovation (BEECON) randomized controlled trials (RCTs) underwent calibration with gold standard raters, followed by annual re-calibration. Raters from CO-OP also completed inter-rater reliability testing; all three raters completed intra-rater reliability testing rounds. Photographs were obtained from children aged 9-39 months. RESULTS: All three raters achieved greater than 0.77 Lin's Concordance Correlation (LCC) versus gold standard consensus during calibration. All three raters had LCC ≥0.83 at recalibration 1 year later. CO-OP trial raters scored 604 photos (151 sets of 4 photographs); mostly both raters were somewhat/very confident in their scoring (≥89%), describing the most photos as "clear" (90% and 81%). The CO-OP inter-rater LCC for total OHI-MIS score was 0.86, changing little when low quality or confidence photos were removed. All three raters demonstrated high intra-rater reliability (≥0.83). CONCLUSIONS: The OHI-MIS plaque scoring system on photos had good reliability within and between trials following protocol training and calibration. OHI-MIS provides a novel asynchronous plaque scoring system for use in young children. Non-clinicians in field or clinical settings can obtain photographs, offering new opportunities for research and clinical care.

AI snapshot

Written from the paper’s full text.

TL;DR

The paper reports calibration and reliability testing for OHI-MIS, a photographic plaque scoring method for primary maxillary incisors in young children. Across calibration, inter-rater, and intra-rater checks, the index showed generally good concordance, supporting asynchronous scoring by trained clinicians from photographs.

Key findings

  • Initial calibration failed to meet the preset LCC threshold of 0.75 for all three raters (range 0.51–0.73); after retraining, all met it (0.77–0.89), and at one-year recalibration all were ≥0.83 (0.83–0.89).
  • In CO-OP, two raters scored 604 photos from 151 children over nine batches; mean inter-rater LCC for total OHI-MIS was 0.86 (range 0.71–0.91), with only batches 1 and 5 marginally below 0.75. Mean simple kappa was 0.66 (0.41–0.76) and mean weighted kappa was 0.73 (0.54–0.85).
  • Photo quality and confidence were mostly high: Rater C rated 89.7% clear, 9.1% blurry, and 1% missing/unreadable; Rater A rated 81.5% clear, 16.6% blurry, and 2.0% missing/unreadable. Both raters were somewhat or very confident in ≥89% of ratings, and excluding low-quality or low-confidence photos changed LCC little (0.86 to 0.88 when low confidence was excluded).
  • Intra-rater reliability was high: Rater C had LCC 0.83 and 0.95; Rater A had 0.88 and 0.99; BEECON Rater B had ≥0.96. Agreement with adjudicated scores was 84.5% for Rater A and 91.1% for Rater C.

Methods

  • Reliability and calibration study nested in the CO-OP and BEECON randomized controlled trials. OHI-MIS adapts the OHI-S Debris Index to the labial surfaces of four primary maxillary incisors; research assistants applied disclosing agent, photographed with iPhone 7 cameras, uploaded images to REDCap, and raters scored asynchronously.
  • Three clinician raters (two CO-OP, one BEECON) calibrated against a gold-standard consensus using 30 BEECON pilot photos, then a second batch of 25 new plus 5 repeated photos; annual recalibration followed. CO-OP inter-rater testing occurred in nine batches of 15–20 children; intra-rater testing rescored 20–30 photos at least 4 weeks later, with a second round 12–18 months later.
  • Analysis used Lin’s Concordance Correlation with 95% CIs, a prespecified 0.75 threshold, and kappa/weighted kappa for tooth-level agreement in SAS 9.4. The abstract states photographs came from children aged 9–39 months; the trials enrolled children aged 6–36 months (CO-OP) and 6 to <48 months (BEECON).

Limitations

  • Paper-stated: plaque is not stable over time, so scores provide a momentary assessment; further validation is needed on plaque surface area versus thickness and on longitudinal links to oral health outcomes. Photo quality, positioning, lighting, blurriness, retraction, dental anomalies, and young-child cooperation posed scoring challenges requiring protocol and feedback.
  • Paper-stated: reliability of photo quality and confidence scoring was not an aim of the study; kappa has known limitations, including influence of agreement and skewed marginal distributions.
  • Evident from the text: reliability was tested with only three clinician raters, and CO-OP inter-rater testing involved only two raters, so broader generalizability to other raters or settings is not established. Calibration photos came from the BEECON pilot, which may limit transferability.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1111/jphd.12557

Cites

32 of 32 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

2 of 12 cited sources checked · 1 supported · 0 flagged · accuracy not yet measured

  • Premisecitation

    OHI-S was developed for children and adolescents aged 10-19 years and for adults.

    “The Simplified Oral Hygiene Index (OHI-S) was developed for use in children and adolescents aged 10–19 years, as well as adults.22”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites [22]

      The available Abstract discusses dental caries, plaque, and oral-hygiene habits in children and teenagers, but does not address the development or intended age range of OHI-S. Searches for “Simplified Oral Hygiene Index,” “OHI-S,” and the stated ages found no relevant passage in our copy. (LLM, conf=0.72)

      Experimental — accuracy not yet measured
  • Findingown data

    Baseline calibration with gold standard raters required two rounds of training and scoring to establish threshold LCC in May/June 2018.

    “Baseline calibration with gold standard raters required two rounds of training and scoring to establish threshold LCC in May/June 2018.”
    Quote in the paperSubstantial claim
  • Findingown data

    After additional training, the three raters scored a second batch of 30 photos with a gold standard consensus mean OHI-MIS score of 1.8 and rater mean scores ranging from 1.9 to 2.0.

    “Following additional training, the three raters scored a second batch of 30 photos with a gold standard consensus mean OHI-MIS score of 1.8 where the raters had mean scores ranging 1.9–2.0.”
    Quote in the paperSubstantial claim
  • Findingown data

    In the second calibration batch, all three raters met the 0.75 threshold with LCCs of 0.77-0.89.

    “In the second batch of photos, all three raters met the threshold of 0.75 with LCCs 0.77–0.89.”
    • Also stated in: § Results: · plain
      “All three raters achieved greater than 0.77 Lin’s Concordance Correlation (LCC) versus gold standard consensus during calibration.”
    Quote in the paperSubstantial claim
  • Findingown data

    Raters maintained a high level of intra-rater reliability.

    “Raters maintained a high level of intra-rater reliability.”
    • Also stated in: § Results: · boosted
      “All three raters demonstrated high intra-rater reliability(≥0.83).”
    Quote in the paperSubstantial claim
  • Classified as probably not a substantial claim. 15
  • May combine two claims. 23
  • The quote only partly states this claim. 1
  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    Disclosing agents provide more effective plaque visualization, aiding evaluation of oral hygiene and improving debridement.

    “Disclosing agents provide more effective plaque visualization which aids evaluating oral hygiene and improves effective debridement.24, 25”
    Quote in the paperNot a substantial claim1 of 2 sources checked
    • cites [24] source
    • cites [25] source
  • May combine two claims.

    Findingown data

    Of the 151 children, Rater A indicated at least one clear photo for 133 children, and Rater C indicated at least one clear photo for 138 children.

    “Of the 151 children, Rater A indicated there was at least one clear photo for 133 children, while Rater C indicated 138 children had at least one clear photo.”
    Quote in the paperNot a substantial claim
  • May combine two claims.

    Findingown data

    In the CO-OP trial, 151 participants were scored over 9 batches, yielding up to 604 scored photos.

    “In the CO-OP trial, 151 participants were scored over 9 batches, yielding up to 604 scored photos.”
    • Also stated in: § Results: · plain
      “CO-OP trial raters scored 604 photos (151 sets of 4 photographs); mostly both raters were somewhat/very confident in their scoring (≥89%), describing the most photos as “clear” (90% and 81%).”
    Quote in the paperNot a substantial claim
  • May combine two claims.

    Findingown data

    In the initial calibration scoring of 30 photos, 29 photos were scorable, with a gold standard mean plaque score of 2.0 (SD 0.52) and rater means ranging from 1.7 (SD 0.68) to 2.2 (SD 0.61).

    “Following the initial scoring of 30 photos, including the two CO-OP and one BEECON raters, 29 photos with gold standard mean plaque score of 2.0 (SD = 0.52) were “scorable” with means ranging from 1.7 (SD = 0.68) to 2.2 (SD = 0.61).”
    Quote in the paperSubstantial claim
  • May combine two claims.

    Findingown data

    There was a slight tendency for raters to score higher than the gold standard consensus, but the difference was not statistically significant.

    “While there was a slight tendency for raters to score higher than the gold standard raters’ consensus, this difference was not statistically significant (Table 1 and Figure 1).”
    Quote in the paperNot a substantial claim
  • May combine two claims.

    Findingown data

    Rater C scored six teeth as missing or unreadable (1%), 55 as blurry (9.1%), 542 as clear (89.7%), and 1 data point as missing (<0.1%).

    “Among these, Rater C scored six teeth as missing or unreadable (1%), 55 as blurry (9.1%), 542 as clear (89.7%), and 1 data point was missing (<0.1%).”
    Quote in the paperNot a substantial claim
  • May combine two claims.

    Findingown data

    In the initial calibration, LCC ranged from 0.51 to 0.73, and no raters met the required threshold.

    “LCC ranged from 0.51 to 0.73, indicating no raters met the required threshold.”
    Quote in the paperNot a substantial claim
  • May combine two claims.

    Findingown data

    At recalibration one year later, scoring 30 photos with a mean plaque score of 2.2 and a gold standard of 2.0, LCC ranged from 0.83 to 0.89, all exceeding the minimum 0.75 threshold.

    “Calibration versus gold standard consensus 1 year later included scoring 30 photos with a mean plaque score of 2.2 with the gold standard being 2.0. LCC ranged from 0.83 to 0.89, all exceeding the minimum 0.75 threshold requirement.”
    • Also stated in: § Results: · plain
      “All three raters had LCC ≥0.83 at recalibration 1 year later.”
    Quote in the paperSubstantial claim
  • May combine two claims.

    Findingown data

    When cases with low confidence were excluded, mean Kappa was 0.69 (range 0.41-0.87) and mean Weighted Kappa was 0.75 (range 0.52-0.90).

    “When cases with low confidence were excluded, the mean Kappa Coefficient was 0.69 (range: 0.41–0.87) and the mean Weighted Kappa Coefficient was 0.75 (range: 0.52–0.90).”
    Quote in the paperNot a substantial claim
  • May combine two claims.

    Findingown data

    Rater C was very or somewhat confident in more than 92% of ratings, and Rater A was very or somewhat confident in more than 89% of ratings.

    “Rater C was “very” (69.5%) or “somewhat confident” (22.4%) in more than 92% of ratings, while Rater A was “very” (56.6%) or “somewhat confident” (32.8%) in more than 89% of ratings.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingnone given

    Participant cooperation posed a frequent challenge, particularly in the young age group included in these studies.

    “Not surprisingly, participant cooperation posed a frequent challenge, particularly with the young age group included in these studies.”
    Quote in the paperNot a substantial claim
  • May combine two claims.

    Findingown data

    Rater C had intra-rater LCCs of 0.83 and 0.95, and Rater A had LCCs of 0.88 and 0.99 for the two batches.

    “Rater C had an LCC of 0.83 and 0.95, while Rater A had LCC of 0.88 and 0.99 for the 2 batches.”
    Quote in the paperSubstantial claim
  • May combine two claims.

    Findingown data

    In the CO-OP trial, Rater A agreed with the adjudicated score 84.5% of the time and Rater C 91.1% of the time.

    “In the CO-OP trial, Raters A and C were in agreement with the adjudicated score 84.5% and 91.1% of the time, respectively.”
    Quote in the paperSubstantial claim
  • Classified as probably not a substantial claim. May combine two claims.

    Findingown data

    Challenges with photo quality included positioning, lighting, blurriness, and adequate retraction, so regular screening of photo quality with timely feedback is needed.

    “Some challenges were also seen with scoring related to photo quality, including positioning, lighting, blurriness, and adequate retraction, so regular screening of photo quality with timely feedback is needed to ensure staff collect adequate quality photos.”
    Quote in the paperSubstantial claim
  • May combine two claims.

    Findingnone given

    In the CO-OP trial, when raters were not concordant, the difference in plaque score was usually one category, and there was no directional error by tooth being scored.

    “In the CO-OP trial, when raters were not concordant, the difference in plaque score was usually one category, and there was no directional error by tooth being scored (central vs. lateral incisor).”
    Quote in the paperSubstantial claim
  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Findingown data

    Although not statistically significant, greater variance was noted among scores for central incisors for Rater C and lateral incisors for Rater A.

    “Although not statistically significant, greater variance was noted among scores for the central incisors for Rater C and lateral incisors for Rater A (Table 3).”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingnone given

    Disclosed plaque varied from early plaque to thick debris, reflecting longstanding plaque accumulation.

    “Disclosed plaque varied from early plaque to a thick debris, reflecting longstanding plaque accumulation.”
    Quote in the paperNot a substantial claim
  • May combine two claims.

    Findingown data

    Pearson correlation coefficients were greater than 0.90 for all teeth, indicating error was not more prevalent for particular teeth.

    “Pearson correlations were calculated to determine if error was more prevalent when scoring particular teeth (central incisors vs. lateral incisors), however correlation coefficients of >0.90 were achieved for all teeth.”
    Quote in the paperNot a substantial claim
  • May combine two claims.

    Findingown data

    Rater A scored 12 teeth as missing or unreadable (2.0%), 100 as blurry (16.6%), 492 as clear (81.5%), and no data as missing.

    “Rater A scored 12 teeth as missing or unreadable (2.0%), 100 as blurry (16.6%), 492 as clear (81.5%), and no data as missing.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Heavy visible plaque places a patient at very high caries risk.

    “heavy visible plaque is consistently agreed to place a patient at very high caries risk.3, 4”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites [3] source
    • cites [4]

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Classified as probably not a substantial claim. May combine two claims.

    Premisecitation

    Dental plaque biofilm and plaque accumulation are risk factors for dental caries and periodontal disease.

    “Dental plaque biofilm and plaque accumulation are risk factors for dental caries and periodontal disease.1, 2”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites [1] source
    • cites [2] source
  • Classified as probably not a substantial claim.

    Interpretationreasoning

    Multiple clinicians can review photographs one or more times without extra participant burden.

    “Multiple clinicians can also review photographs one or more times, without extra participant burden.”
    Quote in the paperSubstantial claim
  • Classified as probably not a substantial claim.

    Interpretationreasoning

    The findings suggest potential pragmatic application in the clinical field through teledentistry.

    “The findings also suggest potential pragmatic application in the clinical field through teledentistry.”
    Quote in the paperSubstantial claim
  • May combine two claims.

    Interpretationown data

    Excluding low quality and/or low confidence photos had minimal effect on overall LCC, so exclusion was not necessary for analysis.

    “Excluding low quality and/or low confidence showed minimal effect on the overall LCC, indicating that exclusion was not necessary for analysis.”
    Quote in the paperSubstantial claim
  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Study design and conductnone given

    A general limitation of using plaque disclosing and measurement to assess oral health is that plaque may not be stable over time.

    “A general limitation of using plaque disclosing and measurement to assess oral health is that plaque may not be stable over time.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Interpretationown data

    Photo quality and raters' confidence can be consistently high with proper quality assurance and training.

    “Reliability with scoring photograph quality and confidence was not an aim of this study, although data suggest photo quality and raters’ confidence can be consistently high with proper quality assurance and training.”
    Quote in the paperNot a substantial claim
  • May combine two claims.

    Interpretationreasoning

    The OHI-MIS protocol can be applied to sites with a single clinician rater as well as sites with more than one clinician rater.

    “The OHI-MIS protocol can be applied to sites with a single clinician rater as well as sites with more than one clinician rater.”
    Quote in the paperNot a substantial claim
  • May combine two claims.

    Interpretationown data

    Excluding low quality and/or low confidence photos had minimal effect on overall Kappa coefficients, so exclusion was not necessary for analysis.

    “Excluding low quality and/or low confidence showed minimal effect on the overall Kappa Coefficients, indicating that exclusion was not necessary for analysis (Figure 2).”
    Quote in the paperSubstantial claim
  • Probably not a substantial claim: may not pass S2 (checkable), S4 (based).

    Interpretationnone given

    The ability for non-clinicians to capture photos could provide a pragmatic and cost-effective means of conducting research and executing oral health screening and promotion strategies.

    “Furthermore, the ability for non-clinicians to capture photos could provide a pragmatic and cost-effective means of conducting research and executing oral health screening and promotion strategies such as interventions aimed at increasing tooth brushing and plaque control.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    Raters completed three rounds of photo review and adjudication before being allowed to score independently.

    “Raters completed three rounds of photo review and adjudication before they were allowed to score independently.”
    Quote in the paperNot a substantial claim
  • Classified as probably not a substantial claim.

    Interpretationreasoning

    Clinicians and researchers can use these methods to evaluate plaque index as a proxy for oral health behaviors such as brushing quality.

    “Clinicians and researchers can use these methods to evaluate plaque index as a proxy for oral health behaviors, such as brushing quality.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: Plaque index is an intermediate variable between brushing behaviors and future caries.
  • Probably not a substantial claim: may not pass S2 (checkable).

    Interpretationreasoning

    Because families may face geographic and other barriers to care, the ability to capture photos for later clinician review may provide pragmatic screening for high-risk oral health behaviors or conditions.

    “Given geographic and other barriers families may face accessing care, the ability to capture photos that clinicians subsequently review may provide pragmatic screening for high-risk oral health behaviors or conditions.”
    Quote in the paperNot a substantial claim
  • May combine two claims.

    Interpretationown data

    Tooth-level variability was modest, so OHI-MIS provided stable inter- and intra-rater reliability evaluation.

    “Our findings suggested tooth level variability was modest, so OHI-MIS provided stable inter- and intra-rater reliability evaluation.”
    Quote in the paperSubstantial claim
  • The quote only partly states this claim.

    Interpretationreasoning

    Non-clinicians can be trained to capture quality photographs in field or clinical settings, including non-clinical environments such as the home.

    “Non-clinicians (e.g., research assistants) can be trained to capture quality photos in non-clinical environments, such as the person’s home.”
    • Also stated in: § Conclusions: · plain
      “Non-clinicians in field or clinical settings can obtain photographs, offering new opportunities for research and clinical care.”
    Quote in the paperSubstantial claim
34 of 34
  • The paper relies on this without support.

    Plaque index is an intermediate variable between brushing behaviors and future caries.

    “Plaque index is an intermediate variable between brushing behaviors and future caries which can be evaluated and intervened upon as primary caries prevention.”
  • The paper relies on this without support.

    Assessing plaque scores in the clinic environment is challenging with young children because of limited cooperation from fear and limited attention span.

    “Assessing plaque scores in the clinic environment can be challenging, particularly with young children due to limited cooperation from fear and limited attention span.”
2 of 2

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

  • Low-confidence exclusion is described as both unchanged and changed

    Minor

    The paper says excluding low-confidence cases left both mean kappa coefficients unchanged, then reports means of 0.69 and 0.75, rather than the previously stated 0.66 and 0.73.

  • Missing-data percentage is understated

    Minor

    One missing data point among 604 photos is approximately 0.17%, not less than 0.1%.

    Reported 0.1 · recomputed 0.165563

2 of 2

Cited works that were retracted

Works this paper cites that were later retracted, from the retraction notices publishers deposit. Major when the notice came out before this paper was published.

  • Cites a retracted work: 10.1002/14651858.cd009857.pub2

    Major

    The cited work was retracted on 2019-04-24 (notice 10.1002/14651858.cd009857.pub3), before this paper was published (2023-02-13). Its authors could have known.

    Cited work 10.1002/14651858.cd009857.pub2 · retracted 2019-04-24 · notice 10.1002/14651858.cd009857.pub3 · this paper published 2023-02-13

1 of 1

Cites

1 retracted

Flagged first, then by citations

1 of the 33 references couldn’t be matched to a known paper.

Research works printed but not linked: 2

  1. 10.14219/jada.archive.1964.0034 · DOI · Paper page
  2. Wei SHY, 1981, Periodontal epidemiological indices for children and adolescents: II. Evaluation of oral hygiene, Pediatr Dent, 4, 73

Cited by

Sorted by citations

Authors
Avenetti, D. M., Martin, M. A., Gansky, S. A., Ramos‐Gomez, F. J., Hyde, S., Van Horn, R., Jue, B., Rosales, G. F., Cheng, N. F., & Shiboski, C. H.
Affiliations
  • University of Illinois Chicago (US)
  • University of California, San Francisco (US)
  • University of California, Los Angeles (US)
  • Department of Pediatric Dentistry, UIC College of Dentistry University of Illinois Chicago Chicago Illinois USA
Venue
Journal of Public Health Dentistry
Publisher
Wiley
Published
2023-02-13
Volume
83
Issue
1
Pages
108-115
ISSN
  • 0022-4006 (ISSN-L)
  • 1752-7325
OpenAlex
W4320480474
Cited by
4
References
33
Citations by year
2025: 1 · 2024: 2 · 2023: 1
FWCI
1.13
Type
journal article
Language
en
Open access
Locations
  • Journal of Public Health Dentistry (publishedVersion)
  • PubMed (publishedVersion)
  • PubMed Central (acceptedVersion)
  • eScholarship (California Digital Library) (submittedVersion)
Funders
  • National Institute of Dental and Craniofacial Research (U01‐DE025507, UH2‐DE025483, UH2‐DE025514, UH3‐DE025483, UH3‐DE025514)
  • University of Illinois
  • University of California, San Francisco
Topics
  • Health Sciences › Dentistry › Periodontics › Dental Health and Care Utilization
  • Health Sciences › Dentistry › Periodontics › Oral microbiology and periodontitis research
  • Social Sciences › Decision Sciences › Statistics, Probability and Uncertainty › Reliability and Agreement in Measurement
Keywords
oral hygiene index, young children, inter-rater reliability, intra-rater reliability, clinical photographs, maxillary incisors
MeSH terms
Calibration, Child, Child, Preschool, Humans, Reproducibility of Results

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Claims read under definition version 3 (extract-v3) on 2026-09-28.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-09-28.

Cited sources are checked by gpt-6-sol.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review