Physical activity self-report and accelerometry measures from the Girls health Enrichment Multi-site Studies.

Background: Valid and reliable physical activity checklists are needed to assess effectiveness of interventions. This study tested the validity and reliability of the Girls health Enrichment Multi-site Studies (GEMS) Activity Questionnaire.
Methods: Two-hundred and ten African-American girls completed the GEMS Activity Questionnaire (GAQ), a checklist of 28 physical and 7 sedentary activities, including TV viewing, inquiring whether performed on the previous day ("yesterday"), and whether usually performed ("usual"). The girls wore an accelerometer (used as the criterion for validity) for three consecutive days at baseline and after a 12-week pilot intervention. Data from 172 girls at baseline and follow-up were usable for this report.
Results: Girls were (mean +/- SD) 8.8 +/- 0.8 years old with a BMI of 22.3 +/- 5.9 kg/m2. The intraclass correlation (ICC) examining reliability for the accelerometer across 3 days at baseline for the combined group was fair (ICC = 0.33, P < 0.21). The test-retest reliability coefficient for the 18-item MET-weighted GAQ yesterday scores for the comparison group of girls was 0.57 (P < 0.001). At baseline and follow-up, nonsignificant correlations were observed between 3-day accelerometer counts/minute and GAQ 18-item usual score for both comparison and intervention groups. A significant correlation was found between change in accelerometer minutes of moderate-to-vigorous activity (MVPA) between 12 noon and 6 PM and change in GAQ physical activities in the comparison girls (R = 0.35, P < 0.01). The TV-usual score was correlated with 3-day accelerometer counts/minute (R = -0.19, P = 0.02) at baseline for the total sample.
Conclusions: Correlations between the GAQ and accelerometer were low, indicating low validity. Although the GAQ may be helpful in describing types of physical activities performed, it needs further development to improve its psychometric properties.