Measurement Properties of Physical Activity Questionnaires Used in Studies Involving Populations with Diabetes: A Systematic Review

초록

Although a large number of questionnaires have been used to assess physical activity in diabetic populations in epidemiological or intervention studies, the measurement properties of physical activity questionnaires (PAQs) used in this population have not yet been reviewed. The purpose of the study was to identify PAQs used in studies involving diabetic populations and to assess the qualitative attributes and the measurement properties of the PAQs. Using MEDLINE, CINAHL, and SPORTDiscus, a systematic literature search was conducted. The Quality Assessment of Physical Activity Questionnaire (QAPAQ) checklist was used to appraise the qualitative attributes and the measurement properties of PAQs included in this review. A total of nine PAQs in nine studies met inclusion criteria. Most of the PAQs had methodological weaknesses with regard to content validity, recall period, ease of use, or interpretability. Of the nine PAQs that provided reliability analyses, none of the PAQs received a positive rating (i.e., the correlation coefficient which is above the specified cut-off point) with the highest level of evidence (i.e., an adequate time interval between test-retest, recommended type of coefficient, and sufficient sample size). Only the long version of International Physical Activity Questionnaire (IPAQ) tested its reliability with an adequate time interval between test-retest and use of appropriate correlation coefficient, but the correlation coefficient values in some categories of physical activity did not reach the specific cut-off points. In the eight PAQs that tested validity, none of the PAQs received a positive rating with the highest level of evidence (i.e., an adequate comparison instrument and correlation coefficient, and sufficient sample size). However, the Community Health Activities Model Program for Seniors (CHAMPS), Physical Activity Scale for the Elderly (PASE), Yale Physical Activity Survey (YPAS), and Pima Indian's Study (PIS) received a positive rating although the correlations coefficients of some categories of physical activity were below the cut-off points. There was no PAQ provided responsiveness analysis. In conclusion, although several PAQs showed a moderate level of results in terms of measurement properties, it is still questionable which PAQ is the most suitable for the use in a diabetic population. Thus, well-designed studies without methodological limitations are needed.

키워드