Determining known-group validity and test-retest reliability in the PEQ (personalized exercise questionnaire)

Abstract

Background: To determine the known-group validity, a type of construct validity, and the test-retest reliability of a newly developed tool, the Personalized Exercise Questionnaire (PEQ), that assesses the barriers, facilitators, and preferences to exercise in individuals with low bone mass and osteoporosis. Methods: A comparative design was used to assess known-group validity and a test-retest design to examine the reproducibility. Ninety-five participants with low bone mass and osteoporosis were recruited from an outpatient clinic in Hamilton, Ontario. The questionnaire was administered to 95 participants at baseline and a subset of 42 participants completed the survey again one week later. The known-group validity of the PEQ was determined using four hypotheses that compared two known groups based on employment level, age, socioeconomic status, and physical activity level. The reproducibility of individual responses was analyzed using the Kappa Coefficient (κ). Results: There was known-group validity for three of the four hypotheses. Test-retest reliability scores ranged from no agreement to almost perfect agreement; seven items had almost perfect agreement (κ: 0.81-1.00), 12 substantial agreement (κ: 0.68-0.74), six moderate agreement (κ: 0.56-0.60), two fair agreement (κ: 0.36-0.40), one slight agreement (κ = 0.23) and one no agreement (κ = - 0.03). Conclusion: Preliminary support for the usefulness of the PEQ is indicated since the majority of the items had at least substantial agreement and known-group validity was moderately supported for some items. Trial registration: This study was retrospectively registered with ClinicalTrials.gov, NCT03125590, on April 24, 2017.