Skip to content

Construction of a Multimodal Fusion AI Model for Pain Scoring Based on Facial Expressions, Voiceprints and Linguistic Text

Construction of a Multimodal Fusion AI Model for Pain Scoring Based on Facial Expressions, Voiceprints and Linguistic Text

Status
Recruiting
Phases
Early Phase 1
Study type
Observational
Source
ChiCTR
Registry ID
ChiCTR2600119004
Enrollment
Unknown
Registered
2026-02-13
Start date
2026-03-01
Completion date
Unknown
Last updated
2026-02-16

For informational purposes only — not medical advice. Sourced from public registries and may not reflect the latest updates. Terms

Conditions

pain

Interventions

Gold Standard:Numeric Rating Scale (NRS) 0-10 pain score jointly determined by 2 independent pain physicians (blinded) based on patient interview information, physical examination, and pain descriptio
Index test:Multi modal AI pain scoring model: integrating facial expressions, patient voiceprints, and language text semantic features to achieve automatic scoring of NRS 0-10 points.

Sponsors

The First Affiliated Hospital of Anhui Medical University
Lead Sponsor

Eligibility

Sex/Gender
All
Age
18 Years to 85 Years

Inclusion criteria

Inclusion criteria: 1. Patients with pain symptoms (Numeric Rating Scale [NRS] score 1-10 points); 2. Patients who are conscious (or have mild cognitive impairment but can cooperate with basic communication) and capable of completing the collection of facial expression, voiceprint, and language text data; 3. Patients who voluntarily sign the informed consent form; 4. Patients aged 18-85 years (covering both the general population and the elderly population).

Exclusion criteria

Exclusion criteria: 1. Patients with severe facial injury or voice disorder, who are unable to undergo facial expression/voiceprint data collection; 2. Patients with severe cognitive impairment or mental illness, who are unable to cooperate with pain assessment; 3. Patients who refuse to participate in data collection or cannot complete follow-up; 4. Patients with severe underlying diseases (e.g., severe cardiac, hepatic, or renal dysfunction) that may affect pain assessment.

Design outcomes

Primary

MeasureTime frame
Consistency between AI model and gold standard NRS score;Gradient differentiation accuracy of moderate-to-severe pain;

Countries

China

Contacts

Public ContactKe Penghui

The First Affiliated Hospital of Anhui Medical University

864040149@qq.com+86 551 6323 6507

Outcome results

None listed

Source: ChiCTR (via WHO ICTRP) · Data processed: Feb 19, 2026