Accuracy of recall of occupational physical activity by questionnaire

The validity of the Tecumseh Occupational Activity Questionnaire (TOQ) and a modified Seven Day Recall (SDR) was measured in 46 healthy women, ages 20-60 years (mean +/- SD, 39.4 +/- 11.8). Occupational data were compared with direct (7-day occupational activity records and Caltrac accelerometer) and indirect (maximal oxygen uptake and percent body fat) measures of physical activity status. Physical activity survey and occupational record data were log transformed to normalize skewed distributions. Intraclass correlations ranged from r = 0.26 to 0.73 (P < 0.05) between similar activities on the TOQ and activity records for light intensity sitting, light-to-moderate intensity standing, and walking. Correlations between similar activities from the SDR and activity records ranged from r = 0.32 to 0.35 (P < 0.05) for light and moderate intensity activities. Correlations for total occupational activity between the surveys and activity records were r = 0.46 (P < 0.001) for the TOQ and r = 0.45 (P < 0.01) for the SDR. A modest correlation (r = 0.34, P < 0.10) was observed between the Caltrac scores and TOQ moderate intensity standing. Only the TOQ light-to-moderate intensity standing was related to maximal oxygen uptake (r = 0.32, P < 0.05). None of the measures were related to percent body fat (P > 0.05). These results suggest that the TOQ and SDR both provide reasonably accurate measurements of light and moderate intensity occupational physical activity. I CLIN EPIDEMIOL 52;3:219-227, 1999. (C) Elsevier Science Inc.