CClinicalTrials.gg
CompletedNCT07797777Updated Sep 2, 2026

AI-Assisted Evaluation of Dental Anxiety in Children: A Machine Learning Approach

An observational study in Dental Anxiety in Children, sponsored by Marmara University. Completed at 1 site in Turkey (Türkiye). Open to participants aged 8 Years to 12 Years, including healthy volunteers. Per ClinicalTrials.gov, last updated 2026-09-02.

Sponsored by Marmara University · Observational

Study type
Observational
Model
Case-only
Time perspective
Cross-sectional
Enrollment
262
Ages
8 Years to 12 Years
Sex
All
01

Study summary

This observational study evaluated whether children's dental anxiety could be identified from their speech using artificial intelligence and machine learning methods. Children aged 8-12 years attending a pediatric dentistry clinic answered a set of short, standardized questions before receiving dental treatment. Their speech was recorded, and their dental anxiety was assessed during the same session using three established measures: the Children's Fear Survey Schedule-Dental Subscale, the Modified Child Dental Anxiety Scale, and the Face Image Scale.

Acoustic characteristics of the children's voices and linguistic characteristics of their spoken responses were analyzed together. Four machine learning algorithms were developed and evaluated to determine how accurately they could distinguish between children with lower and higher levels of dental anxiety. No treatment was assigned or modified as part of the study.

Read the detailed description

Dental anxiety can adversely affect a child's cooperation, treatment experience, and willingness to attend future dental appointments. Existing assessment methods primarily rely on self-report questionnaires, visual scales, and clinical observation. Artificial intelligence-based analysis of speech may provide an additional objective and non-invasive method for recognizing dental anxiety before treatment.

This prospective, cross-sectional observational study included children aged 8-12 years attending the Department of Pediatric Dentistry at Marmara University Faculty of Dentistry between September 2025 and February 2026. Data were collected within the natural workflow of an active pediatric dentistry clinic. The study did not assign participants to any treatment or alter their planned dental care.

Before dental treatment, each participant took part in a standardized 1-3-minute conversation conducted by the same researcher. The questions addressed everyday life, school, oral hygiene habits, previous dental experiences, and expectations regarding the planned dental visit. Speech was recorded using a standardized microphone position and recording protocol.

Immediately after the speech recording and before treatment, dental anxiety was assessed using the Children's Fear Survey Schedule-Dental Subscale, the Modified Child Dental Anxiety Scale, and the Face Image Scale. Scores from each measure were categorized into lower and higher anxiety levels using established thresholds and served as reference labels for machine learning model development.

All data were anonymized before analysis. The researchers' speech and prolonged silent segments were removed from the recordings, and the audio files were converted to a standardized format. Acoustic characteristics were extracted from the recordings using OpenSMILE. The spoken responses were transcribed and analyzed using natural language processing methods based on Whisper and a Turkish-language BERT model. Selected acoustic and linguistic features were combined into a multimodal dataset.

Random Forest, XGBoost, LightGBM, and CatBoost algorithms were trained and evaluated using five-fold cross-validation. Model performance was assessed using accuracy, sensitivity, specificity, precision, F1 score, the area under the receiver operating characteristic curve, predictive values, and confusion matrices. Agreement between model predictions was also evaluated. The primary objective was to identify the machine learning approach that most accurately and sensitively distinguished children with lower and higher dental anxiety levels.

02

Conditions studied

  • Dental Anxiety in Children

Keywords

  • Artificial Intelligence
  • Machine Learning
  • Speech Analysis
  • Speech Emotion Recognition
  • Natural Language Processing
  • Pediatric Dentistry
  • Dental Fear
  • Multimodal Analysis
03

Who can participate

Ages eligible
8 Years to 12 Years
Sexes eligible
All
Accepts healthy volunteers
Yes
Sampling method
Non-probability sample

Study population

he study population consisted of children aged 8-12 years who attended the Department of Pediatric Dentistry at Marmara University Faculty of Dentistry for routine dental care between September 2025 and February 2026. Eligible children completed a standardized speech recording and three dental anxiety assessments during a single visit before dental treatment. Children with both lower and higher levels of dental anxiety were included. Participation did not alter or delay the dental care planned for any child.

Eligibility criteria

IInclusion Criteria:

  • Children aged 8-12 years
  • Attendance at the Department of Pediatric Dentistry, Marmara University -Faculty of Dentistry
  • Native Turkish speaker
  • Age-appropriate neuropsychological development
  • Ability to communicate, cooperate in the clinical setting, and understand and follow the study instructions
  • No chronic or systemic condition affecting speech production, respiratory function, or cognitive processes
  • Written informed consent provided by a parent or legal guardian

Exclusion Criteria:

  • Neurological, hearing, speech, or language disorder that could affect clinical communication or study assessments
  • Suspected or diagnosed neurodevelopmental disorder, including autism spectrum disorder or attention-deficit/hyperactivity disorder
  • Native language other than Turkish
  • Active pathology or history of surgery that could affect voice or speech quality, including resonance or phonation
  • Inability to provide sufficient verbal data, such as consistently giving single-word responses or leaving multiple questions unanswered
04

Study design

Observational model
Case-only
Time perspective
Cross-sectional
Enrollment
262 participants (actual)
Patient registry
No

Groups and cohorts

  • Pediatric Dental Patients

    Children aged 8-12 years attending a pediatric dentistry clinic who underwent a standardized pre-treatment speech recording and dental anxiety assessment during a single study visit. No dental treatment was assigned, changed, or delayed as part of the study.

    Diagnostic Test: Multimodal Speech-Based Dental Anxiety Assessment

Interventions

  • Diagnostic testMultimodal Speech-Based Dental Anxiety Assessment

    Participants completed a standardized 1-3-minute speech recording before dental treatment. Acoustic and linguistic characteristics of their speech were analyzed using artificial intelligence methods. During the same session, dental anxiety was assessed using the Children's Fear Survey Schedule-Dental Subscale, the Modified Child Dental Anxiety Scale, and the Face Image Scale. The assessment was conducted for research purposes and did not alter the participants' planned dental care.

05

What researchers measure

Primary outcomes

  1. Classification Accuracy of the Multimodal Machine Learning Models

    Accuracy was defined as the proportion of participants correctly classified as having lower or higher dental anxiety. Random Forest, XGBoost, LightGBM, and CatBoost models were evaluated separately using reference classifications derived from the CFSS-DS, MCDAS, and FIS. Performance was assessed using five-fold cross-validation. Accuracy values range from 0 to 1, with higher values indicating better classification performance.

    Time frame: Day 1, during the single pre-treatment assessment

Secondary outcomes

  1. Sensitivity of the Multimodal Machine Learning Models

    Sensitivity was defined as the proportion of participants with higher dental anxiety who were correctly identified by each machine learning model. Sensitivity was calculated separately using the CFSS-DS, MCDAS, and FIS classifications as reference standards. Values range from 0 to 1, with higher values indicating better detection of children with higher dental anxiety.

    Time frame: Day 1, during the single pre-treatment assessment

Other outcomes

  1. F1 Score of the Multimodal Machine Learning Models

    The F1 score was calculated as the harmonic mean of precision and sensitivity to evaluate the balance between correctly identifying children with higher dental anxiety and limiting incorrect classifications. Macro and weighted F1 scores were calculated for each machine learning model using CFSS-DS, MCDAS, and FIS classifications as reference standards. Values range from 0 to 1, with higher values indicating better performance.

    Time frame: Day 1, during the single pre-treatment assessment

06

Study locations

1 site
  • Marmara University Faculty of Dentistry
    Istanbul, Maltepe 34854, Turkey (Türkiye)
07

References and documents

Individual participant data

Plan to share: No

No publications or documents are linked to this record.

08

Registry details

Key details

Study ID
NCT07797777
Lead sponsor
Marmara University
Responsible party
Gizem Akova (Principal Investigator, Marmara University) — Principal investigator
First posted
Sep 1, 2026
Start date
Sep 10, 2025
Primary completion
Feb 10, 2026
Completion
Apr 10, 2026
Last update
Sep 2, 2026

Study contacts

Gizem Akova
principal investigator · Marmara University Faculty Of Dentistry

Oversight

Data monitoring committee
No
FDA-regulated drug
No
FDA-regulated device
No
View the source record on ClinicalTrials.gov ↗

Not currently enrolling

This study is completed, as verified in Sep 2026. You cannot join it, but the record below documents what was studied.

Follow this study

Get an email when the registry record changes — status, dates, results — or when someone posts here.

Sign in to follow

Discussion

Questions and observations about this study, from anyone following it. Not medical advice, and not a channel to the study team — their contact details are on the registry record.

Sign in to join the discussion. Reading takes no account; posting does. You choose a display name, and a pseudonym is the default.

Nothing here yet. If you are running this trial, taking part in it, or weighing whether to, this is the place to say so.

Start the discussion