Abstract
Participation in physical activity is beneficial for multiple health outcomes, however, substantial work remains to improve understanding and promotion of this complex behaviour. Longitudinal studies offer unique insights into physical activity, as they allow examination of how participation, and the factors that are associated with participation, may change during the life-span.
This thesis was based on data from the Dunedin Multidisciplinary Health and Development Study (DMHDS). The DMHDS is a birth cohort of around 1000 individuals, born in Dunedin, New Zealand between April 1972 and March 1973. Included in the three studies presented here was information obtained during assessments at birth and ages 3, 5, 7, 9, 11, 13, 15, 18,21 and 26 years.
Study 1 examined tracking of participation in club sport from childhood to adulthood (7, 9, 15, 18 and 21 years). Correlation coefficients revealed low tracking (r = 0.07-0.28) of club sport participation. Summary statistics, which utilised all the longitudinal data and controlled for the influence of covariates, differed slightly with random effect models finding low tracking (Intraclass correlation coefficient: 0.23) and generalized estimating equations suggesting moderate tracking (Stability coefficient: 0.59). These findings suggest that benefits accrue from early physical activity participation for later participation. However, as tracking is only at low to moderate levels, childhood participation in physical activity should not be considered an 'inoculation' against adult inactivity.
Study 2 examined associations between childhood factors (Family socioeconomic status, Family Active Recreation Orientation, Home activities, Childhood motor ability, Intelligence and Psychiatric disorder), contemporaneous factors (Parental health, Body Mass Index, VO2max, General health, Television watching, Tobacco smoking and Alcohol use) and patterns of Persistent Inactivity, Declining participation and Persistent Activity between ages 15 and 18 years (assessed using the Minnesota Leisure Time Physical Activity Questionnaire). Common childhood influences included Family socioeconomic status, Family Active Recreation Orientation, Home activities and Childhood motor ability. Contemporaneous measures that were associated with patterns of participation included Parental health, General health, VO2max and Television viewing.
Study 3 built on the the findings of Study 2 by investigating if factors associated with adolescent physical activity continued to influence physical activity in adulthood. With the exception of Parental health, all the factors that had been associated with adolescent patterns of physical activity were associated with non-participation in vigorous physical activity during at least one of the two adult assessments (age 21 and 26 years). Common influences for the two ages included participation in Home activities during childhood, Childhood motor ability and Persistent Inactivity during adolescence.
In conclusion, the series of studies presented here provides valuable insights into patterns of physical activity from childhood to adulthood and the factors that are associated with these patterns among the DMHDS cohort. Previous findings are extended through the use of a broad range of covariates and independent variables across childhood, adolescence and adulthood. Key findings include the exploration of different techniques to describe tracking across the lifespan and the identification of factors that can usefully be targeted in early intervention programmes to support later physical activity.