Puberty timing and relative age as predictors of physical activity discontinuation during adolescence

Among same-age adolescents, those who enter puberty relatively later and those who are relatively younger (e.g., born later in the year) might be at greater risk of physical activity discontinuation. This study aimed to (1) describe gender-specific discontinuation, re-engagement, and uptake rates in various types of physical activities from the age of 11 to 17 years, and (2) assess puberty timing and relative age as predictors of discontinuation from organized, unorganized, individual, and group-based physical activities. Longitudinal data from 781 (56% girls, age 10-13 years at study baseline) Canadian participants who self-reported puberty status, birthdate, and involvement in 36 physical activities every four months from 2011 to 2018 was analyzed. The incidence of discontinuation, re-engagement, and uptake in organized/unorganized and individual/group activities from grade 6 until grade 12 was described and Cox proportional hazard models were used to estimate associations of puberty timing and relative age with organized/unorganized and individual/group activity discontinuation. Results demonstrate that individual and unorganized activities are maintained longer than group-based and organized activities. Girls who started puberty earlier were more likely to discontinue organized activities than girls with average-puberty timing [Hazard ratio (HR) (95% confidence interval (CI)) 1.68 (1.05-2.69)]. Compared to boys born in the 4th quarter of the year, boys born in the 2nd quarter of the year were less likely to discontinue organized [HR (95% CI) 0.41 (0.23-0.74)], unorganized [HR (95% CI) 0.33 (0.16-0.70)], group [HR (95% CI) 0.58 (0.34-0.98)], and individual activities[HR (95% CI) 0.46 (0.23-0.91)], and boys born in the 3rd quarter were less likely to discontinue unorganized activities[HR (95% CI) 0.41 (0.19-0.88)]. This study illustrates the patterns of physical activity participation throughout adolescence. However, the generalizability of findings may be limited due to participant representation.