Validation of the physical activity instrument for the Life in New Zealand national survey.

A physical activity recall instrument suitable for self-administration by the adult New Zealand population was devised and validated for 140 subjects selected randomly from urban electoral rolls. Validation used bootstrapping to compare correlations between subjective measures derived from the instrument, the Stanford 7-day recall questionnaire and other questionnaires, and objective measures derived from an exercise test. Subjective measures were grouped as measures of high intensity activity, low intensity activity, or total activity metabolism. The high intensity measures correlated moderately with each other (mean r = 0.39), but poorly with metabolism and low intensity measures (r = 0.16, -0.02, respectively). The mean correlation between the metabolism and low intensity groups (r = 0.26) was similar to those within these groups (r = 0.31, 0.25, respectively). Thus, the high intensity measures formed a group distinct from the metabolism and low intensity groups, which represented similar measures. The objective measures that correlated with high intensity measures (mean r = 0.25) are predominantly recognized risk factors for cardiovascular disease; these did not correlate with the metabolism or low intensity measures (mean r = 0.03, -0.07, respectively). Activity measures from the instrument had mean correlations with subjective and objective measures that equaled or surpassed those of the Stanford measures from the same group. The authors conclude that 1) population health studies that assay physical activity should include well-defined measures of high intensity activity, and 2) valid measures of physical activity in the New Zealand population are achievable with this instrument.