Background: Adequate monitoring of older adults' physical activity (PA) is essential to develop effective health promotion programs. The present study examined criterion validity and test-retest reliability of the long International Physical Activity Questionnaire (IPAQ-L), adapted for Belgian, community-dwelling older adults (65 y and older).
Methods: Participants (n = 434) completed the last seven days version of IPAQ-L, modified for the Belgian population of community-dwelling older adults. This elderly-adapted version of IPAQ-L combined vigorous and moderate activities, and questions on gait speed and recreational cycling were added. Furthermore, participants wore an ActiGraph GT3X(+) accelerometer for at least five days. Criterion validity was determined by comparing self-reported weekly minutes of moderate-to-vigorous physical activity (MVPA) and weekly minutes of total PA with accelerometer data, defined by two different cut points (Freedson vs. Copeland). To examine test-retest reliability, a subsample of 29 participants completed IPAQ-L for a second time within a ten day interval.
Results: IPAQ-L showed moderate criterion validity for measuring weekly minutes of MVPA and total PA (Spearman's ρ range 0.33-0.40). However, plots on agreement between self-reported and accelerometer PA showed a systematic over-reporting of IPAQ-L for MVPA. In contrast, plots indicated that IPAQ-L under-estimated levels of total PA, however, this under-estimation of total PA was substantially lower than the observed over-reporting of MVPA. Test-retest reliability was moderate-to-good for work-related PA, domestic PA, MVPA and total PA (ICC range 0.52-0.81), but poorer for transportation and recreational PA (ICC 0.44 and 0.43, respectively).
Conclusions: Criterion validity results suggest that IPAQ-L is more valid to measure older adults' weekly minutes of total PA than weekly MVPA minutes. Moreover, results might imply that content validity of IPAQ-L can be improved if specific light-intensity PA items are incorporated into IPAQ-L. Test-retest reliability of IPAQ-L was moderate to good, except for weekly minutes of transportation and recreational PA, probably due to week-to-week variability of these behaviors.