Validity and reliability of single-item questions about physical activity

Area of Science:

• Epidemiology
• Public Health
• Biostatistics

Background:

• Large-scale cohort studies are crucial for understanding lifestyle-related diseases.
• Accurate measurement of physical activity is essential for epidemiological research.
• The Japan Collaborative Cohort Study for Evaluation of Cancer Risk Sponsored by Monbusho (JACC Study) utilized self-administered questionnaires for physical activity assessment.

Purpose of the Study:

• To evaluate the validity and reliability of single-item physical activity questions used in the JACC Study.
• To determine if these questions are appropriate for large-scale epidemiological research on lifestyle-related diseases.

Main Methods:

• Validity was assessed by comparing questionnaire responses with interview-elicited data on time spent and energy expenditure (Spearman's rank correlation).
• Reliability was evaluated using test-retest analysis over approximately one year (weighted kappa coefficients).
• Participants included 1,730 adults for validity and 1,075 for reliability testing.

Main Results:

• Spearman's rank correlation coefficients for validity ranged from 0.43 to 0.60, indicating moderate correlations.
• Weighted kappa coefficients for reliability ranged from 0.39 to 0.56 (moderate) for sports/exercise and 0.25 to 0.39 (fair) for walking.
• Both validity and reliability varied by sex and age group.

Conclusions:

• The single-item physical activity questions demonstrate moderate validity and reliability.
• These measures are deemed appropriate for establishing baseline data on long-term physical activity in large cohort studies.
• The findings support the use of these questions for research on lifestyle-related diseases.