Temporal physical activity patterns are associated with obesity in U.S. adults

Area of Science:

• Public Health
• Epidemiology
• Biostatistics

Background:

• Few studies have integrated multiple aspects of physical activity (PA) to classify patterns associated with health outcomes.
• Understanding temporal PA patterns is crucial for developing targeted health interventions.

Purpose of the Study:

• To develop and validate temporal PA patterns using time and activity counts.
• To investigate the association between these temporal PA patterns and health status indicators, including obesity, type 2 diabetes, and metabolic syndrome.

Main Methods:

• Utilized accelerometry data from 1999 adults in the National Health and Nutrition Examination Survey (2003-2006).
• Employed dynamic time warping and kernel k-means clustering to identify distinct temporal PA patterns (clusters).
• Applied multivariate regression models to assess associations between PA clusters and health indicators (BMI, WC, obesity, diabetes, metabolic syndrome).

Main Results:

• Identified four distinct clusters representing different temporal PA patterns.
• Cluster 1, characterized by the lowest PA counts spread throughout the day (6:00-23:00), showed significantly higher BMI and WC compared to clusters with higher activity in the afternoon/evening (16:00-21:00) or morning (8:00-11:00).
• The lowest PA pattern was associated with increased odds of obesity.

Conclusions:

• Temporal PA patterns, defined by time of day and activity intensity, are meaningfully associated with key health status indicators.
• Individuals with consistently low PA throughout the day exhibit poorer health metrics compared to those with higher PA during specific periods.
• These findings highlight the importance of considering the temporal dimension of PA in health research and public health strategies.