AI-Enabled Frailty Risk Prediction in Adult Congenital Heart Disease

NCT07479654 | Not Yet Recruiting

• 1 other identifier: 25MMHIS443e
• Study Type: interventional
• Target: 410
• Locations: 0 countries
• Sites: N/A

Brief Summary

The goal of this three-year mixed-methods observational study with an embedded randomized controlled trial is to develop and validate a frailty risk prediction model and evaluate an artificial intelligence-based voice emotion detection-guided counselling intervention in adults with congenital heart disease (ACHD). The main questions it aims to answer are: Are symptom clusters associated with frailty and psychological outcomes in adults with congenital heart disease? Can symptom clusters and psychosocial factors be used to predict frailty risk over time in ACHD patients? Does an AI-based voice emotion detection-guided counselling intervention improve psychological outcomes, fatigue, and quality of life among high-risk ACHD patients? Researchers will compare ACHD patients receiving AI-based voice emotion detection-guided counselling with those receiving usual care to determine whether the intervention reduces depression, anxiety, sleep disturbance, fatigue, and frailty risk, and improves grit and quality of life. Participants will: Complete longitudinal assessments of symptom clusters, frailty, and psychological status at baseline and follow-up time points Participate in qualitative interviews to explore lived experiences related to symptoms and frailty Receive AI-based voice emotion detection-guided counselling (intervention group only in Year 3)

Conditions

Adult Congenital Heart Disease; Symptom Clusters; Frailty; Risk Prediction Model; Artificial Intelligence-Based Voice Emotion Detection

Outcome Measures

Primary Outcomes (1)

• Depression and Anxiety assessed using the Patient-Reported Outcomes Measurement Information System (PROMIS) Short Form Scales

  The PROMIS Depression Short Form evaluates negative mood, feelings of hopelessness, helplessness, and loss of interest. The PROMIS Anxiety Short Form assesses emotional distress related to feelings of fear, worry, tension, and hyperarousal. Each scale consists of multiple items rated on a 5-point Likert scale (1 = Never, 2 = Rarely, 3 = Sometimes, 4 = Often, 5 = Almost Always) based on the participant's experiences during the past 7 days. Raw scores are summed and converted to PROMIS T-scores, standardized to a population mean of 50 with a standard deviation of 10. Possible raw score range: 8-40 (for the 8-item short form). Higher scores indicate greater levels of depressive or anxiety symptoms (worse outcomes).

  Baseline, 2 weeks, 1 month, and 3 months after initiation of the intervention

Secondary Outcomes (3)

• Chinese version of the Satisfaction with Life Scale; SWLS

  Baseline, 2 weeks, 1 month, and 3 months after initiation of the intervention

• Grit-10

  Baseline, 2 weeks, 1 month, and 3 months after initiation of the intervention

• The Chinese Version of Fatigue Assessment Scale; FAS-C

  Baseline, 2 weeks, 1 month, and 3 months after initiation of the intervention

Study Arms (2)

AI-Based Voice Emotion Detection-Guided Counselling (Intervention)

EXPERIMENTAL

Participants assigned to this arm will receive an artificial intelligence-based voice emotion detection-guided counselling intervention in addition to usual care. The intervention uses voice emotion analysis to provide personalized psychological feedback and counselling aimed at improving emotional regulation, reducing psychological distress, and enhancing quality of life.

Behavioral: AI-Based Voice Emotion Detection-Guided Counselling

Usual Care (Control)

ACTIVE COMPARATOR

Participants assigned to this arm will receive usual clinical care for adult congenital heart disease without the AI-based voice emotion detection-guided counselling intervention.

Behavioral: Usual Care (Control)

Interventions

AI-Based Voice Emotion Detection-Guided Counselling BEHAVIORAL

Participants assigned to the intervention arm will receive an artificial intelligence-based voice emotion detection-guided counselling intervention in addition to usual care. The intervention uses voice recordings collected during structured counselling sessions to analyze emotional features, including emotional valence and arousal, through artificial intelligence-based voice emotion detection algorithms. Based on the analyzed emotional profiles, individualized psychological feedback and counselling guidance are provided to support emotional regulation, stress coping, and adaptive self-management. The counselling content is tailored to participants' emotional states and symptom experiences and focuses on reducing psychological distress, improving sleep and fatigue management, enhancing grit, and promoting quality of life. The intervention is delivered by trained healthcare professionals following a standardized protocol, with sessions conducted at predefined intervals during the inter

AI-Based Voice Emotion Detection-Guided Counselling (Intervention)

Usual Care (Control) BEHAVIORAL

Usual Care

Usual Care (Control)

Eligibility Criteria

• Age: 20 Years+
• Sex: all
• Healthy Volunteers: No
• Age Groups: Adult (18-64), Older Adult (65+)

You may qualify if:

• Adults aged 20 years or older.
• Diagnosed with congenital heart disease by a physician.
• Able to communicate in Mandarin or Taiwanese, with normal cognitive function.

You may not qualify if:

• Presence of other congenital anomalies.
• Currently hospitalized for acute decompensation, or having severe cognitive impairment or psychiatric symptoms that preclude completion of questionnaires.
• Pregnant individuals.

Sponsors & Collaborators

• Mackay Memorial Hospital: lead

MeSH Terms

Conditions

Heart Defects, Congenital; Syndrome; Frailty

Condition Hierarchy (Ancestors)

Cardiovascular Abnormalities; Cardiovascular Diseases; Heart Diseases; Congenital Abnormalities; Congenital, Hereditary, and Neonatal Diseases and Abnormalities; Disease; Pathologic Processes; Pathological Conditions, Signs and Symptoms

Study Design

• Study Type: interventional
• Phase: not applicable
• Allocation: RANDOMIZED
• Masking: NONE
• Purpose: TREATMENT
• Intervention Model: PARALLEL
• Sponsor Type: OTHER
• Responsible Party: PRINCIPAL INVESTIGATOR
• PI Title: Assistant Professor

Study Record Dates

• First Submitted: January 18, 2026
• First Posted: March 18, 2026
• Study Start: March 1, 2026
• Primary Completion (Estimated): December 31, 2028
• Study Completion (Estimated): December 31, 2028
• Last Updated: March 18, 2026

Record last verified: 2026-01