Lin Wang, Kimon Ntotsis, James Manifield, Katerina Tziannou, Tatiana Plekhanova, Alex Rowlands, Charlotte L Edwardson, Molly Baldwin, Nicolette C Bishop, Charlotte Bolton, James D Chalmers, Enya Daynes, Annemarie Docherty, Claudia Efstathiou, Omer Elneima, Lucy Gardiner, William Greenhalf, Neil Greening, Beatriz Guillen-Guio, Victoria Harris, Ewen Harrison, Ling-Pei Ho, Alexander Horsley, Linzy Houchen-Wolloff, Joseph Jacob, Olivia Leavy, Nazir Lone, Daniel Lozano-Rojas, Michael Marks, Hamish McAuley, Melitta McNarry, Claire Nolan, Jennifer K Quint, Betty Raman, Matthew Richardson, Jack Sargeant, Ruth M Saunders, Marco Sereno, Aarti Shikotra, Amisha Singapuri, David J Stensel, Maxime Taquet, Mathew Thorpe, Ioannis Vogiatzis, William Man, Louise V Wain, Iain Stewart, Sally Singh, Christopher E Brightling, Claire Lawson, Tom Yates, Rachael A Evans
GPPAQ demonstrates internal consistency, with construct and convergent validity in adults 1 year after COVID-19 hospitalisation. GPPAQ effectively identifies inactive individuals to support clinical care; however, its sensitivity suggests under-estimation of activity relative to accelerometry. Future pathways should combine GPPAQ with device-based assessment to optimise evaluation of physical activity to guide pulmonary rehabilitation and targeted interventions.
BACKGROUND: Physical inactivity is a risk factor for severe COVID-19 and often worsens after hospitalisation. Clinicians need quick, accurate assessments to target interventions. We aimed to assess the validity of the General Practice Physical Activity Questionnaire (GPPAQ) in adults recovering 1 year after COVID-19 hospitalisation.
METHODS: Post-hospitalisation for COVID-19, adults attended a 1-year-visit and completed the GPPAQ Physical Activity Index (PAI-4 active), 14-day wrist-worn accelerometry (moderate-vigorous physical activity (MVPA) active) and other health outcomes. Validity was examined via 1) internal consistency (factor analysis); 2) measurement invariance across sex, age and ethnicity using differential item functioning (DIF); 3) convergent validity (GPPAQ sensitivity/specificity versus accelerometry); and 4) construct validity (correlations with health outcomes).
RESULTS: 752 participants had GPPAQ and accelerometry (265 female, mean±sd age 60.9±11.6 years), MVPA 18.75 min·day-1 (interquartile range 7.55-36.11), PAI-1 46.8%, PAI-2 15.6%, PAI-3 19.4% and PAI-4 18.2%. Factor analyses supported good internal consistency with two factors (daily activities and physical exercise). Comparative factor index (CFI) showed an excellent fit (CFI 0.965). DIF indicated moderate variability by sex and age. GPPAQ-PAI showed limited sensitivity (26.3%) for correctly classifying physically active individuals, but higher specificity (88.4%) for classifying physical inactivity, and weak-to-moderate correlations with health outcomes.
CONCLUSIONS: GPPAQ demonstrates internal consistency, with construct and convergent validity in adults 1 year after COVID-19 hospitalisation. GPPAQ effectively identifies inactive individuals to support clinical care; however, its sensitivity suggests under-estimation of activity relative to accelerometry. Future pathways should combine GPPAQ with device-based assessment to optimise evaluation of physical activity to guide pulmonary rehabilitation and targeted interventions.