Title: A Bayesian smoothing spline ANOVA model to re-examine the effects of an intervention on infant massage
Authors: Chin-I Cheng - Central Michigan University (United States) [presenting]
Abstract: An intervention on infant massage called the M technique, which is a structured manual massage technique to reduce stress, was administered to nine hospitalized, very preterm infants. The outcome variables are repeated measures on physiological (heart rate, RR -respiratory rate-, and SaO2 -oxygen saturation) and behavioral (ABSS -Anderson Behavioral State Scale) status during a period of time. The outcome measurements from the intervention were published based on results from repeated measure ANOVA. The outcome measurements are re-examined with the Bayesian smoothing spline ANOVA model. Smoothing spline ANOVA is an approach to semiparametric function estimation based on an ANOVA type of decomposition. This model allowed testing for effects which can be linear or nonparametric (i.e., smooth or interactions between selected linear and smooth effects). The ability to test for these effects provides insights of the cumulative impact of the M technique intervention. The Bayesian approach was considered, and the Bayes factor was used for variable selection and hypothesis testing. Findings showed that the intervention had a cumulative impact on heart rate, SaO2 and that for ABSS effect is curvature (nonlinear).