Differing associations with childhood outcomes using behavioural patterns derived from three data reduction...

Area of Science:

• Child Health Research
• Behavioural Science
• Data Analysis Methodologies

Background:

• Understanding how multiple health behaviours influence childhood outcomes is crucial.
• Data reduction techniques for deriving behavioural patterns can yield non-identical results.
• These differing patterns may associate differently with childhood outcomes.

Purpose of the Study:

• To compare the associations of behavioural patterns derived from three distinct methods (cluster analysis, latent profile analysis, principal component analysis) with three childhood outcomes.
• To investigate the impact of data reduction technique choice on the observed relationships between health behaviours and child development.

Main Methods:

• Utilized data from the Healthy Active Preschool and Primary Years study (n=432, children aged 6-8).
• Derived behavioural patterns using cluster analysis (CA), latent profile analysis (LPA), and principal component analysis (PCA) from diet, physical activity, sedentary behaviour, and sleep data.
• Assessed associations with child health (height, weight, waist circumference), quality of life, and academic performance using adjusted linear regression.

Main Results:

• CA and LPA each yielded three patterns (healthy, unhealthy, mixed); PCA yielded four patterns.
• Differences in associations were observed between patterns derived from CA/LPA and pattern scores from PCA concerning the three childhood outcomes.
• Each method identified distinct healthy, unhealthy, and mixed behavioural patterns.

Conclusions:

• The choice of data reduction method significantly influences the observed associations between behavioural patterns and childhood outcomes.
• Discrepancies highlight the need for careful consideration of methodology when comparing findings across studies.
• Methodological choices have implications for interpreting and applying research on children's health behaviours.