Digital Transformation of Korean Articulation and Phonology Assessment: Reliability and Validity of Hi-DongDong

Background/Objectives: Speech sound disorders (SSDs) are common in childhood and affect literacy and academic achievement. Standardized assessments such as the Urimal Test of Articulation and Phonology (UTAP) rely on time-intensive manual scoring. We evaluated the reliability and validity of Hi-DongDong, a Delphi-developed digital articulation assessment. Methods: This prospective multicenter study across five institutions recruited in total 931 children aged 2–13 years from outpatient clinics and community childcare facilities. Speech samples obtained with Hi-DongDong and UTAP-1 or UTAP-2 were independently evaluated by three experienced speech–language pathologists. Analyses were conducted on participants with valid paired datasets, using intraclass correlation coefficients (ICCs), Pearson correlations, Wilcoxon signed-rank tests, and Bland–Altman analyses. Results: Inter-rater reliability was excellent (ICC ≥ 0.96). Inter-instrument reliability was good to excellent for total percentage of correct consonants (PCC) and total PCC–revised. All outcomes correlated significantly with UTAP (r = 0.347–0.939, p < 0.001). Utterance-level consonant accuracy indices showed minimal mean bias, whereas phonological mean length of utterance, vowel accuracy, and word accuracy showed systematic bias or wide limits of agreement, precluding interchangeable use. Conclusions: Hi-DongDong showed excellent inter-rater reliability and concurrent validity, particularly for consonant accuracy; however, its diagnostic accuracy remains unknown, and it cannot yet replace standardized diagnostic thresholds.