Validity and reliability of the telephone-administered international physical activity questionnaire in Brazil

Purpose: To evaluate the validity and reliability of the telephone-administered long IPAQ version. Methods: The questionnaire was administered by telephone to adults on days 1 and 6. On day 1, the same questionnaire was administered by face-to-face interview, and accelerometers were delivered to subjects. Reliability was measured by comparing data collected using the telephone questionnaire on days 1 and 6. Validity was measured by comparing the telephone questionnaire data with (a) face-to-face questionnaire and (b) accelerometry. Results: Data from all instruments were available for 156 individuals. The Spearman correlation coefficient for telephone interview reliability was 0.92 for the leisure-time section of IPAQ, and 0.87 for the transport-related section of IPAQ. The telephone interview reliability kappa was 0.78. The Spearman correlation between the telephone-administered and the face-to-face questionnaire was 0.94 for the leisure-time and 0.82 for the transport-related section. The kappa was 0.69. There was a positive association between quartiles of accelerometer data and total telephone-administered IPAQ score (P < .001). The Spearman correlation was 0.22. Conclusions: The telephone-administered IPAQ presented almost perfect reliability and very high agreement with the face-to-face version. The agreement with accelerometer data were fair for the continuous score, but moderate for the categorical physical activity variables.