Development and assessment of relative validity and reliability of the Workplace Beverage Intake Questionnaire in UK office workers

Abstract

Background: There are no beverage measurement tools evaluated for use in UK working-age adults. This study aimed to develop and evaluate a novel beverage intake questionnaire. Methods: A 57-item online tool (Workplace Beverage Intake Questionnaire [WBIQ]) was developed through stakeholder consensus. Relative validity was measured against 7-day food records, and reliability was tested across three time points. Evaluation outcomes of interest were total beverage intake and beverage intake during working hours, intake from seven beverage categories (plain water, sugar sweetened, low/zero calorie, tea, coffee, milk based and 100% fruit based) and energy, caffeine and free sugar intake from beverages. Reliability was determined by intraclass correlation coefficients (ICC) and validity via correlation analyses and visual assessment of Bland–Altman plots. Results: The evaluation study population comprised office workers (n = 71, 74.6% women, mean age: 32, standard deviation: 8.5 years). The WBIQ had moderate reliability (ICC: 0.50–0.75) across total fluid intake and all beverage categories except milk-based drinks and 100% fruit-based drinks where it was rated poor. Caffeine, free sugar and energy from beverages had poor-to-moderate reliability. Correlation coefficients were large (r > 0.50, p < 0.001) comparing diet records and WBIQ across all categories of beverage except low-/zero-calorie soft drinks (r = 0.34, p < 0.01). Bland–Altman plots showed a similar trend across all variables, with better agreements at lower intake and the absolute difference increasing proportionally at higher intakes. Over 90% of respondents agreed/strongly agreed that the tool was easy to navigate and understand. Conclusions: The WBIQ is the first stage in the development of a tool for UK-specific beverage intake measurement in working-age adults. Further refinement and testing are required to improve reliability.