Investigating between-group effects of a physical activity intervention across the physical activity intensity...

Area of Science:

• Childhood physical activity research
• Intervention study analysis
• Statistical modeling in public health

Background:

• Novel methods are needed to assess physical activity (PA) intervention effects across the full intensity spectrum.
• This study evaluated multivariate pattern analysis for determining between-group effects in PA interventions.

Purpose of the Study:

• To test the applicability of multivariate pattern analysis for evaluating intervention effects on physical activity across the intensity spectrum.
• To compare multivariate pattern analysis with linear mixed models for analyzing PA intervention outcomes.

Main Methods:

• Utilized data from 698 Norwegian preschool children in a cluster randomized controlled trial.
• Employed multivariate pattern analysis and linear mixed models with varying PA descriptor resolutions (4, 17, 51 variables).
• Assessed PA using ActiGraph GT3X+ over 7- and 18-month follow-ups.

Main Results:

• Higher-resolution PA descriptors showed marginally better model fit in multivariate pattern analysis (1.27-3.59% explained variance).
• Detected significant standardized mean differences ranging from ±0.15 to 0.36.
• Multivariate pattern analysis and linear mixed models yielded comparable results.

Conclusions:

• Multivariate pattern analysis is a suitable method for evaluating between-group effects across the PA intensity spectrum in intervention studies.
• The approach is recommended as a secondary analysis for cluster trials due to complexities in accounting for clustering.
• Findings support the use of advanced statistical methods for nuanced PA intervention research.