Purpose: The objective of this work is to present findings from the available global surveillance systems of physical activity (PA) for children and youth (0-17 year-old), and to highlight research gaps and needs for their improvement.

Methods: This is a narrative review of the international surveillance systems, studies, and global initiatives collecting or compiling evidence on the PA among children and youth. It was informed by a literature search in PubMed, the screening of references from relevant publications, and grey literature. The location, target population, frequency, description, and main findings in terms of the prevalence of physical (in)activity of the identified surveillance systems, international studies, and global initiatives were extracted.

Results: PA surveillance systems lack standardization and are not conducted regularly. Surveillance data are mostly focused on older children and adolescents and are especially scarce for PA trends over time, vulnerable populations, and in resource-limited low-middle-income countries. The PA levels of 5-17year-old children and youth varies depending on the data source but is consistently low across international data. Variation in the geographic patterns of PA levels are observed between studies, which could be attributed to the diversity of approaches used to measure PA. Currently, several international studies and initiatives including both objectively and subjectively measured PA data indicate that children and youth are more active in countries from Africa and Northern/Eastern Europe, while they are less active in North America, the United Kingdom, China, and India; however the opposite was observed in the Health Behaviour in School-aged Children survey.

Conclusions: Systematic surveillance of the PA of children and youth is still emerging or at very early stages in many countries. There is a need for globally accepted and standardized measurement, reporting, and accountability protocols that countries can universally follow in order to have meaningful progress in the global surveillance and promotion of PA among children and youth. The lack of standardized and validated globally-relevant surveillance tools for PA creates a predicament for accurate comparisons but also presents an opportunity for the development, trial, and implementation of universal surveillance mechanisms.

Purpose: Retrospective individual participant meta-analyses and prospective meta-analyses have various advantages over traditional meta-analysis and providing invaluable opportunities to advance scientific knowledge. Such pooled analyses require large consortia of harmonised data. The Prospective Physical Activity, Sitting, and Sleep consortium (ProPASS) is a new platform for retrospective and prospective pooling of cohorts that use thigh-worn accelerometry to measure physical activity, sedentary behaviour, posture, and sleep. The ultimate scientific objective of ProPASS is to explore the associations of physical activity, posture, and sleep patterns with long-term health outcomes. As of October 2019, ProPASS is supported by 12 international cohorts comprising over 70,000 participants.  A key innovation in ProPASS is methodological flexibility that allows for inclusion of data from any accelerometry device that outputs raw acceleration and are is worn on the thigh.

Methods & Results: This talk will describe early outcomes and plans of ProPASS with focal point on the principles and methods that will underpin data harmonisation across 4 axes: retrospective harmonisation of non-accelerometry data, prospective harmonisation of non accelerometry data, and (retrospective and prospective) harmonisation of accelerometry data. The talk will also provide data on the most recent developments (including the empirically validated comparison of different accelerometers) and will describe future ProPASS plans. The challenges and collaborative work opportunities that have arisen during the harmonisation processes will be given particular emphasis.

Conclusions: Consortia of accelerometry data are the future of the evidence on physical activity - related behaviours and health. Carefully conducted and efficient harmonisation methods are a key to their long-term success. It provides an opportunity to increase study power through maximizing sample size and heterogeneity in exposures/outcomes, and permits more sophisticated statistical analyses to  answer novel research questions that could not be addressed using a single study.

Purpose: The current global prevalence of physical inactivity in children calls for prompt political and practical action. Children’s physical activity (PA) can be accumulated through different forms of daily activities such as active travel, physical education (PE) and organised PA. Understanding activity patterns and their contribution to overall PA can inform future behavioural change interventions. Pooling individual data from existing studies enables enhanced exploration of variation in behaviour and increases the statistical power for analyses. Due to the heterogeneity in methodology across contributing studies, this strategy also requires harmonising the data to be comparable across the studies. This presentation will describe the procedure of harmonising data from the International Children’s Accelerometry Database (ICAD) particularly focusing on PE and organised PA variables.

Methods & Results: At present ICAD includes longitudinal and cross-sectional data from 23 studies conducted in Europe, United States, UK, Australia and Brazil. Out of 23 studies, three studies included both PE and organised PA variables, and 13 studies had only organised PA data. Across studies, PE participants were measured across 1-3 waves, with a range of 1-10 items per assessment. Organised PA was assessed at 1-6 waves, with a median of 3 (ranging from 1-457) items per study wave. Questionnaires were completed by children and/or parents, and centred on the involvement, frequency and duration of weekly PE classes and/or specific sports and physical activities. The level of comparability of items across waves in eight studies (with two or more waves) was low, where only two studies utilised the same questionnaire across all waves. All data have been collated into a single repository for cataloguing and harmonisation. This presentation will outline the harmonisation methods applied to this data, and discuss strengths and limitations of the harmonised constructs.

Conclusions: Data pooling across studies can provide large and heterogeneous samples and statistical power to explore differential contribution of activity patterns, but requires complex harmonisation procedures. The development of a rigorous, transparent and systematic process of retrospective data harmonisation is a vital to perform pooled analyses.