Reliability and validity of two multidimensional self-reported physical activity questionnaires in people with
Flávia A Carvalho1, Priscila K Morelhão1, Marcia R Franco1
1Departamento de Fisioterapia, Faculdade de Ciências e Tecnologia, UNESP - Univ Estadual Paulista, Presidente Prudente, Brazil.
Background:
Although there is some evidence for reliability and validity of self-report physical activity (PA) questionnaires in the general adult population, it is unclear whether we can assume similar measurement properties in people with chronic low back pain (LBP).
Objective:
To determine the test-retest reliability of the International Physical Activity Questionnaire (IPAQ) long-version and the Baecke Physical Activity Questionnaire (BPAQ) and their criterion-related validity against data derived from accelerometers in patients with chronic LBP.
Design:
Cross-sectional study.
Methods:
Patients with non-specific chronic LBP were recruited. Each participant attended the clinic twice (one week interval) and completed self-report PA. Accelerometer measures >7 days included time spent in moderate-and-vigorous physical activity, steps/day, counts/minute, and vector magnitude counts/minute. Intraclass Correlation Coefficients (ICC) and Bland and Altman method were used to determine reliability and spearman rho correlation were used for criterion-related validity.
Results:
A total of 73 patients were included in our analyses. The reliability analyses revealed that the BPAQ and its subscales have moderate to excellent reliability (ICC2,1: 0.61 to 0.81), whereas IPAQ and most IPAQ domains (except walking) showed poor reliability (ICC2,1: 0.20 to 0.40). The Bland and Altman method revealed larger discrepancies for the IPAQ. For the validity analysis, questionnaire and accelerometer measures showed at best fair correlation (rho < 0.37).
Conclusions:
Although the BPAQ showed better reliability than the IPAQ long-version, both questionnaires did not demonstrate acceptable validity against accelerometer data. These findings suggest that questionnaire and accelerometer PA measures should not be used interchangeably in this population.
