Background The health benefits of physical activity in adolescence are well-documented. Currently, UK adolescents are not achieving recommended levels of physical activity.

Due to lack of financial independence and mobility restrictions, adolescents spend a significant amount of time in their neighbourhood making it a key setting for physical activity. Feeling unsafe in their neighbourhood may be a potential barrier to physical activity.

This study aims to examine associations between objective and subjective measures of neighbourhood safety, crime and physical activity.

Methods Participants (n = 10,913) came from the Millennium Cohort Study; a nationally representative UK birth cohort. At age 11 subjective neighbourhood safety was assessed via questionnaire whilst the Index of Multiple Deprivation (IMD) crime domain was linked to participant postcodes and Reported Crime Incidence, measured with Police.Data.UK, was linked to participant Lower Super Output Areas. Only crime categories relevant to the neighbourhood were included in Reported Crime Incidence. At age 14, participants self-reported physical activity and a subsample (n=4,813) wore activity monitors for one weekday and one weekend day.

Associations between subjective safety, IMD crime, Reported Crime Incidence and self-reported physical activity were quantified using linear regression models. Zero Inflated Poisson (ZIP) models were used to examine associations with accelerometer-measured physical activity. We adjusted for parental education, family income, ethnicity and season of accelerometer wear. An interaction term for sex was tested to assess whether associations between crime exposures and physical activity differed by sex. Models are also presented stratified by sex.

Results Participants that described feeling not safe compared to very safe at age 11 reported on average 0.29 (95% CI -0.50, -0.09) fewer days of physical activity at age 14. No association was seen between subjective safety and accelerometer-measured physical activity.

Those living in the highest IMD crime areas reported on average 0.32 (95% CI -0.47, -0.16) fewer days of physical activity compared to those living in the lowest crime areas. Individuals living in the highest IMD crime areas achieved 7.38 (95% CI -13.21, -1.55) fewer minutes of daily accelerometer-measured exercise than those in least crime areas.

No association was seen between Reported Crime Incidence and self-reported or accelerometer-measured physical activity. The Reported Crime Incidence subcategory of violence and sexual offenses was associated with 0.20 (95% CI -0.39, -0.20) fewer days of self-reported physical activity but not accelerometer physical activity.

Conclusion This study demonstrates varying associations between subjective safety and objective crime with physical activity levels in adolescence, highlighting the complexities around subjective and objective measures.