Questionnaire choice affects the prevalence of recommended physical activity: an online survey comparing four measuring instruments within the same sample

BACKGROUND: Since prevalence estimates of recommended physical activity (PA) considerably vary between different surveys, prevalence might be crucially affected by the choice of measuring instrument. The aim of the present study is to compare the results of four PA questionnaires regarding the current moderate- and vigorous-intensity aerobic PA (MVPA) recommendations of the World Health Organization. METHODS: Within an online survey, participants answered the German Health Interview and Examination Survey for Adults (DEGS), the European Health Interview Survey PA Questionnaire (EHIS), the Eurobarometer (EURO), and a single-item measure (SIM). Weekly volume of MVPA was compared via a Friedman test and the prevalence of participants achieving the WHO’s MVPA recommendation via Cochran’s Q. Agreement between all questionnaire pairs was evaluated via Kappa statistics. RESULTS: One hundred seventy-six participants were included in the analyses (70.5% female, mean age: 33.1 years (SD=12.2)). Between the four questionnaires, the weekly volume of MVPA statistically significant differed (SIM: MED=90.0 (MIN=0.0, MAX=210.0), DEGS: MED=120.0 (MIN=0.0, MAX=420.0), EHIS: MED=24.0 (MIN=0.0, MAX=1395.0), EURO: MED=51.0 (MIN=0.0, MAX=2430.0), p<.001, all pairwise comparisons p<.01), as well as the prevalence of participants achieving the MVPA recommendations (SIM 31.3% (95% CI 24.5–38.7), DEGS 43.2% (95% CI 35.8–50.8), EHIS 67.0% (95% CI 59.6–73.9), EURO 87.5% (95% CI 81.7–92.0), p<.001), except between SIM and DEGS (p=.067). Agreement was weak between all questionnaire pairs (all κ< 0.60). CONCLUSIONS: Questionnaire choice crucially affects the resulting MVPA data and hence the prevalence of achieving recommended PA levels within the same sample. Therefore, for PA surveillance, standardised survey and analysis methods and efforts to harmonise monitoring systems are needed, since whether recommended levels of PA are achieved should not be determined by the choice of one measuring instrument or another.