Stability and change in leisure-time physical inactivity and its predictors in mid-adulthood: findings from a prospective British birth cohort study

Abstract Background Changes in inactivity occur within individuals as they age, but knowledge of influences on these variations, is lacking. Most studies ignore changes in potential influences, instead focusing on one timepoint. We aimed to identify whether changes in health and social factors are associated with concurrent changes in inactivity. Methods Health, social factors, and leisure-time inactivity (activity frequency less than once a week) were self-reported at 33 (baseline) and 50 years in the 1958 British birth cohort (n=12 271). Using multinomial logistic regression, we assessed the association of baseline health and social factors, and stability and change in these factors from age 33 to 50 years, with inactivity patterns from ages 33 to 50 years (never inactive, persistently inactive, deteriorating, or improving). Missing values were imputed via multiple imputation; results were broadly similar to those using observed values. Findings Approximately 3804 individuals (31%) were inactive at both ages; 4295 (35%) had changed inactivity from age 33 to 50 (2086 [17%] deteriorating, 2209 [18%] improving). Baseline poor-rated health and obesity were associated with subsequent (33–50 years) inactivity; for example, for poor-rated health, relative risk ratios (RRRs) for persistent inactivity and deterioration versus never inactive were 2·01 (95% CI 1·70–2·38) and 1·38 (1·16–1·64), respectively; and for improvement versus persistently inactive was 0·77 (0·63–0·94). Stability and change in health factors were associated with concurrent inactivity patterns; for example, compared with always good or excellent health, worsening and always poor health were associated with lower RRRs for inactivity improvement (0·64 [0·50–0·80] and 0·58 [0·44–0·76], respectively). Low self-efficacy at baseline was associated with persistent inactivity and deterioration; worsening and always low self-efficacy from 33 to 50 years were associated with detrimental inactivity patterns. Although manual socioeconomic position at baseline was associated with persistent inactivity and deterioration, associations for changes in socioeconomic position were less consistent. Baseline depression was associated with persistent inactivity; depression at 33–50 years was not associated with inactivity patterns. No associations were observed for employment, marital or cohabitation status, or parenthood status. Interpretation Self-reported data could potentially bias results because of misclassification, whereas the prospective repeated data spanning decades is a study strength. Associated changes in mid-life health factors with deleterious changes in inactivity shown in our study, highlight the maintenance of health, weight, and self-efficacy across adulthood to deter inactivity. Several of these factors are relatively common, underscoring their potential importance for population levels of inactivity in mid-adulthood. Funding This work was supported by the Department of Health Policy Research Programme through the Public Health Research Consortium. The views expressed in this abstract are those of the authors and not necessarily those of the Department of Health.