The Accuracy of Single-Time-Point Physical Activity Measures Compared to Multi-Year Composite Scores from the English Longitudinal Study of Ageing

This study evaluated the validity of using baseline physical activity as a proxy for habitual activity in longitudinal research. Data were sourced from the English Longitudinal Study of Ageing (ELSA), including older adults aged 50+ in England, observed over a 20-year period with data collected every 2 years from 2002 to 2023. A total of 12,099 participants from Wave 1 of ELSA (2002-2003) (mean age 64.2, 44.1% male) were included in the analysis. Physical activity frequency was measured using self-reported questionnaires at each of the 10 Waves, and a 10-wave composite score was created using the available physical activity data. Health outcomes included depressive symptoms, executive function, memory, quality of life, and self-rated general health. Baseline and 10-wave composite physical activity measures were moderately to strongly correlated (ρ = 0.56-0.72), with comparable associations observed for health outcomes such as moderate activity with depressive symptoms and general health, and vigorous activity with executive function and quality of life. The composite score, however, demonstrated larger effect sizes and better reflected the decline in physical activity over time. To conclude, baseline physical activity measurements can serve as a reliable proxy for habitual activity in longitudinal studies, though multiple measurements over time improve statistical power.