Test‐retest reliability of a self‐reported physical activity environment instrument for use in rural settings. (10th May 2020)

• Record Type: Journal Article
• Authors: Cleland, Verity; Timperio, Anna; Sharman, Melanie J.; Dollman, James
• Is Part Of: Australian journal of rural health. Volume 28:Number 2(2020)
• Page Start: 168
• Page End: 179
• Publication Date: 2020-05-10
• Subjects: epidemiologic methods -- public health -- rural population -- surveys and questionnaires; Rural health -- Periodicals; Rural health -- Australia -- Periodicals; 613
• DOI: 10.1111/ajr.12625
• Languages: English
• ISSNs: 1038-5282

Abstract

Abstract: Objective: Little is known about how the physical environment impacts physical activity behaviour among rural populations, who are typically less active and at higher risk of chronic disease than urban dwellers. The lack of individual‐level instruments to assess the physical environment in rural areas limits advancement of this field. Among rural adults, this study aimed to evaluate (a) the test‐retest reliability of a self‐reported questionnaire of individual‐level perceptions of the physical activity environment, and (b) the stability of a self‐reported physical activity questionnaire. Design: Cross‐sectional questionnaire repeated twice, 2 weeks apart. The questionnaire included 94 items relating to the perceived physical environment (representing nine summary scores), demographic characteristics and physical activity. Setting: Rural Australia. Participants: Rurally residing adults (≥18 years) across three Australian states. Main outcome measures: Test‐retest reliability evaluated by weighted Kappa statistics (individual items) and intra‐class correlations (summary scores). Results: A total of 292 participants (20% men) completed both questionnaires, on average 22 days apart. Test‐retest reliability of individual items ranged from weighted Kappa 0.37‐0.85 (median: 0.59). Internal reliability for five summary scores was good to excellent (Cronbach's alpha: 0.81‐0.97). Test‐retest reliability was good to excellent for six summary scores (intra‐class correlations: 0.67‐0.77). Conclusions: The findings indicated good to excellent test‐retest reliability for most items, particularly "fixed" constructs for this new questionnaire measuring the perceived physical environment in rural populations. This study represents an important step towards improving measurement of physical activity environments in rural populations, potentially leading to better tailored interventions to promote active and healthy living in rural areas.