BACKGROUND

Medication refusal in children is largely driven by aversive taste profiles, which in turn influence adherence and therapeutic outcomes. However, there are no standardized methods for evaluating taste in young children. This study compares facial recognition technology with three hedonic visual scales in this population.

METHODS

Children, 3-7 years of age, were enrolled with informed parental permission into an institutional review board-approved, double-blind, randomized investigation. Each child received three test articles: prednisone (bitter), simple syrup (sweet), and filtered water (neutral), with an appropriate washout. Facial recognition software (Noldus FaceReader 7) recorded facial expression and intensity for 30-60 s after administration. Participants subsequently rated taste using three hedonic scales (5-point Sjövall and 5- and 3-point TASTY) and responded to simple questions on their perception of the test article. Repeated measures analysis of variance and multiple regression analysis were used to explore associations between palatability measures.

RESULTS

Twelve children (seven males: ten white and two black) completed the study without adverse effects. There were no significant differences in participant characteristics by randomization sequence. The three hedonic scales tracked similarly for each test substance, with correlations between the 5-point scales (r = 0.899) comparable to those between the 3- and 5-point scales (r = 0.860-0.903). Hedonic scales appeared more reliable in assessing taste response than facial recognition, which did not effectively discriminate positive and negative responses.

CONCLUSIONS

Our experience suggests that the TASTY scales appear to offer the greatest promise for assessing palatability in future clinical use.