NON-EQUIVALENCE OF SELF-REPORT AND CALTRAC MEASURES OF ENERGY EXPENDITURE

Abstract

62 Caltrac and 7-day Physical Activity Recall (PAR) are used to quantify physical activity in epidemiology research. The purpose of this study was to examine the equivalence of these two instruments. The archival data of 202 women and 195 men came from an observational cohort study of the association between weight fluctuation and cardiovascular disease. The subjects varied in age (Mean = 45.1±13.6) and body mass index (26.4 ± 4.4). The daily caloric expenditure estimated from the PAR was 2790 (±668) kcal. This was not different (F(1,395) = 0.13; p = 0.71) from the Caltrac mean of 2781 (± 542) kcal. The zero-order correlation between the two methods was 0.71. Body weight was highly correlated with both the PAR (0.86) and Caltrac (0.72) suggesting that weight was the source of the PAR-Caltrac correlation. Partial correlation showed that with body weight controlled, only 6.8% of the PAR-Caltrac variance was common. Multiple regression examined the influence of age, weight, gender, and log transformed PAR on Caltrac daily total caloric expenditure. The age, weight, and gender model yielded a R of 0.839. Adding log transformed PAR to the model increased R to 0.841. Although this was a significant (p=0.046) increase, the PAR only accounted for an additional 0.3% of Caltrac variance. While an R of 0.841 is high, nearly all(70.4% vs. 70.7%) of the Caltrac variance is due to age, weight, and gender, not PAR estimates of energy expenditure.