This exploratory randomized controlled trial (RCT) evaluates the effectiveness of the AI-based application "HiDongDong" in improving articulation in children aged 2 to under 14 years with suspected articulation delays.
1\. Study Methods
1. Screening and Enrollment Children suspected of articulation developmental delay will be assessed for eligibility according to predefined inclusion and exclusion criteria. Eligible participants will be enrolled, randomized, and scheduled for study visits.
2. Visit 1 - Baseline Assessment
At the first visit, a blinded assessor will collect demographic data (child's age, sex, and underlying medical conditions) and conduct a single baseline assessment, including the following:
* Language assessments: U-TAP (Urimal Test of Articulation and Phonology)
* SELSI (Sequenced Language Scale for Infants) or PRES (Preschool Receptive- Expressive Language Scale)
* Functional independence: WeeFIM (Functional Independence Measure for Children)
* Caregiver emotional status: Satisfaction (5-point Likert scale), depression (BDI; Beck Depression Inventory), anxiety (BAI; Beck Anxiety Inventory)
* Clinician satisfaction: 5-point Likert scale
3. Post-Assessment Procedures Following baseline assessment, the control group will be discharged and return home. For the intervention group, a personalized language training program will be generated based on the results of the HiDongDong assessment.
4. Training and Orientation Participants in the intervention group will receive instruction on how to use the HiDongDong language training program, along with guidance on the study procedures.
5. Intervention Period The intervention group will complete the HiDongDong language training program at home under caregiver supervision for 4 weeks, 5 sessions per week, 20 minutes per session. Adherence will be monitored throughout the intervention period. The control group will receive language training provided by a speech-language pathologist.
6. Visit 2 - Post-Intervention Assessment and Study Completion Four weeks after the intervention, both the control and intervention groups will undergo a single post-intervention assessment conducted by a blinded assessor. Upon completion of the assessment, follow-up will be concluded.
2\. Comparator and Randomization Participants will be randomized by an independent statistician affiliated with the Medical Research Collaborating Center of the Biomedical Research Institute at Seoul National University Bundang Hospital (SNUBH), using SAS version 9.4.
Assessor blinding will be applied. A blinded assessor (e.g., a hospital-based speech-language pathologist) will conduct all assessments without knowledge of each participant's group allocation, in an independent time and setting.
The sole difference between the two groups is the provider of the intervention:
* Control group: Language training delivered by a speech-language pathologist
* Intervention group: Language training delivered via the HiDongDong application (investigational medical device)
Participants will return after 4 weeks for post-intervention assessment by a blinded assessor, after which follow-up will be concluded.
To prevent interim analysis, assessment results will not be disclosed to any study personnel until the dataset is complete.
