J Med Internet Res. 2026 Aug 7;28:e96800. doi: 10.2196/96800.
ABSTRACT
BACKGROUND: Pain remains a critical issue among hospitalized children and may negatively affect postoperative recovery. In addition to pharmacological pain management, nonpharmacological approaches have been used to support pediatric care. Among these emerging approaches, socially assistive robots (SARs) may offer an opportunity to support children during hospitalization. However, limited evidence exists regarding the use of SARs in pediatric postoperative recovery and their influence on children’s emotional responses during child-robot interaction (CRI).
OBJECTIVE: This study aimed to examine changes in postoperative pain levels following a SAR intervention among hospitalized children. In addition, it aimed to explore emotional responses during CRI using automated facial expression analysis.
METHODS: A single-arm pre-post study was conducted in a pediatric surgical ward. Children recovering from surgery participated in a structured SAR intervention consisting of 3 phases: warm-up, educational video, and interactive engagement. Pain outcomes were assessed using the self-reported Wong-Baker FACES pain rating scale and the observer-rated FLACC (face, legs, activity, cry, and consolability) scale. Emotional responses were evaluated using automated facial expression analysis, which generated continuous emotional valence scores ranging from -1 (negative) to +1 (positive). Wilcoxon signed-rank tests were used to analyze pain outcomes, and Friedman tests were used to examine differences in emotional valence across intervention phases.
RESULTS: A total of 37 children were included in the pain outcome analysis, and 35 (95%) children were included in the emotional valence analysis after excluding participants with insufficient facial expression data. Significant reductions were observed in both self-reported and observed behavioral pain following the intervention. Self-reported pain scores decreased from a median of 6 (IQR 4-6) to 4 (IQR 2-4; P<.001), and FLACC scores decreased from a median of 3 (IQR 2-4) to 1 (IQR 1-2; P<.001). Emotional valence remained negative across all intervention phases. The Friedman test did not reach statistical significance across the 3 phases and showed a small effect size (P=.05).
CONCLUSIONS: The SAR interventions may be associated with lower postoperative pain scores among hospitalized children. Although emotional valence did not significantly change during CRI, automated facial expression analysis was implemented and demonstrated the feasibility of continuous affective assessment in a real-world pediatric clinical setting. These findings support the potential use of the SAR interventions as a complementary strategy in pediatric postoperative care and provide preliminary evidence supporting the integration of real-time affective assessment into pediatric health care.
PMID:42566770 | DOI:10.2196/96800
