Examining the application of the transactional model of sleep to premature infants: A structural equation modelling approach within a population-based dataset

Background: This study utilised structural equation modelling (SEM) to investigate factors influencing sleep patterns of preterm infants, within the Growing Up in Ireland (GUI) population-based dataset. SEM enabled the... [ view full abstract ]

Background:
This study utilised structural equation modelling (SEM) to investigate factors influencing sleep patterns of preterm infants, within the Growing Up in Ireland (GUI) population-based dataset. SEM enabled the examination of multiple independent and dependent variables defined by the transactional model of sleep (Goodlin-Jones et al, 2000). This study identified infant and parent characteristics that promoted optimal sleep in preterm infants and established whether the parent-infant relationship mediated this association. A secondary purpose was to test the transactional model of sleep.
Method:
A full SEM was conducted using the AMOSTM statistical package for use with SPSSTM (Figure 1). A measurement model specified the relationship between observed and latent variables and a path model hypothesised the causal relationship between the latent variables, not directly measured (Tabachnick & Fidell, 2007). The preterm cohort of infants was divided into two groups (n=210 early preterm; n=527 late preterm), with results comparing both groups.
Results:
The father-infant relationship had a mediating impact on the relationship between infant temperament and nighttime waking in the early preterm group only. In contrast, the maternal-infant relationship had a mediating role in both groups.
Conclusion:
These findings add to the body of knowledge on the transactional model of sleep, and are the first to identify infant temperament and the father-infant relationship as important factors.
Application to Practice:
Given the multiplicity of variables considered by occupational therapists, SEM is an ideal research method to capture the complexity of clinical practice.

Figure 1: Full structural equation model