Short Orals: Physical activity measurement development

Purpose: To inform studies developing and validating psychometric properties of physical activity (PA) questionnaires for children and adolescents. Methods: We used the results of our two recent reviews to identify gaps and areas of improvement. The first review systematically apprised and summarized 85 papers evaluating the psychometric properties of PA questionnaires to identify the best available questionnaires in childhood. The second review included 197 papers from three systematic reviews to evaluate the inclusion and exclusion of children with disabilities and chronic diseases (CDACDs) within these studies. Results: Our first systematic review did not identify any PA questionnaire with concurrent acceptable validity and reliability due to a lack of high methodological quality and/or satisfactory validity (correlation ≥0.4) and reliability (≥0.7). Most low methodological quality studies used a comparator tool with unknown or unacceptable measurement properties or inappropriate test and retest time intervals. The second review indicated that 22% of studies reported whether they included or excluded CDACDs. The percentage of studies that included CDACDs remained low despite recent increases. Few included papers were from low-income countries (1%). Conclusion: We recommend employing standardized guidelines (e.g., COSMIN) to improve the methodological quality of studies, specifically by reporting detailed information about the content validity, selecting appropriate test and retest time intervals, choosing comparator tools with adequate validity, and evaluating measurement error of PA questionnaires using appropriate statistical methods. We suggest including all age-appropriate domains of PA (school, leisure, sports, chores, transportation, and work) when developing questionnaires to measure total PA, developing and validating more PA questionnaires for preschoolers and including diverse regions/countries, especially low-income countries and rural communities. Further research is needed to develop and validate PA questionnaires or evaluate the psychometric properties of existing PA questionnaires among CDACDs. Funding: Canadian Institutes of Health Research and Prentice Institute for Global Population and Economy.

Background: Sedentary behavior (SB) is an umbrella term for activities that collectively represent an energy expenditure approximating resting levels when in a sitting, reclining, or lying posture. Excessive time spent in SB is linked to deleterious health outcomes; however, recent evidence shows this relationship may differ according to the type of SB. The Questionnaire for Screen Time of Adolescents (QueST), which measures different types of activities, is useful for SB research but has received limited attention in Canada. Purpose: To evaluate a modified QueST (m-QueST) assessing SB activities for Canadian adolescents. Methods: Participants between the ages of 14 and 21 completed an online questionnaire including sociodemographic information, the m-QueST items, and questions probing clarity and comprehension. The m-QueST items were reworded to include a postural component. The original QueST items are study time, working, watching videos, playing games, using social media and chat applications. Two new behaviours, passive transportation and doing crafts, were included in the m-QueST. Clarity for each m-QueST item was assessed with a Likert response scale, participants reported if they understood how to answer the questions (yes/no) and if they had any suggestions (open-ended). Results: 25 individuals participated (64% females, mean age 18.4 years, 88% white, 64% college/university student). The SB definition was rated as highly/quite clear by 92% of participants. Clarity for m-QueST items varied between 84% (work) and 92% (video games, social media, passive transportation, crafts) being highly/quite clear. No comprehension issues of m-QueST items were reported, and no additional SB activities were suggested by participants. Conclusions: This version of the QueST for SB appears clear and comprehensible, and research is warranted to evaluate its reliability. Practical implications: The m-QueST is a new questionnaire that may be used to measure SB in research with Canadian adolescents. Funding: Social Sciences and Humanities Research Council of Canada.  

Background Outdoor active play (OAP) contributes to children’s well-being and fosters their physical and mental health. There is a variety of definitions and measurement approaches reporting on OAP across research projects and policy frameworks. This diversity leads to discrepancies in reporting. Inconsistent findings across studies make it challenging to draw definitive conclusions about trends in OAP. The aim of the current project was to develop a standardized questionnaire on OAP for the Dutch research and policy context. Methods & Results First, we conducted an international literature review on OAP assessments. We compared various sources, revealing consistencies which led to our focus on parental reporting on frequency and duration of OAP in a select number of settings. Second, during a session with colleagues, we discussed findings and outlined the questionnaire’s structure. We drafted first version of the questionnaire. Based on existing research, we chose additional modules. In the third phase, we invited experts to participate in a survey during which they could provide online feedback, shaping questionnaire revisions. The final OAP questionnaire (OAPQ) comprises a core questionnaire of 10 questions and five additional modules, totaling 35 questions. The OAPQ takes into account OAP in three different contexts, ie school, after school care and free time. The five additional modules focus on (1) what else does your child do in their free time?, (2) what does your neighborhood look like?, (3) what is your opinion on OAP?, (4) what does your child do during OAP?, and (5) characteristics child and family. Completion time is approximately 15 minutes. Researchers can tailor the OAPQ by selecting modules. Conclusions & Practical Implications If used consistently by all organizations working in the research and policy context, the OAPQ contributes to a standardized monitoring of OAP levels. Funding: Dutch Ministry of Health, Welfare & Sports

Introduction: Monitoring secular trends in physical activity (PA) is essential to track progress towards the UNs 2030 sustainable development physical activity goal. PA questionnaires (PAQs) are standard in tracking physical activity worldwide. Albeit their practicality and feasibility for surveillance purposes, PAQs have inherent limitations. In contrast, devices are not prone to the same biases as PAQs. An important and complex topic relates to the comparison between PAQ- and device-measured PA and whether they provide similar prevalence data on trends in PA. Purpose: To explore whether PAQs and devices provide the same results when assessing secular trends in physical activity. Methods: The Norwegian Physical Activity Surveillance Study (NPASS) is a repeated cross-sectional study of population level PA in Norwegian 20-85-year old’s. Adherence to PA guidelines was assessed concurrently by the IPAQ-short and 7-day actigraphy. Results: We observed a negative secular change in self-reported adherence to PA guidelines, from 67% in 2008-09 to 65% in 2014-15, with individuals in 2014-15 having 15% lower odds of meeting the PA guidelines compared to individuals in 2008-09 (OR: 0.85, 95% CI: 0.76 – 0.96). In contrast, we observed a positive secular change in device-measured adherence to PA guidelines, from 68% in 2008-09 to 73% in 2014-15, corresponding to 26% higher odds of meeting the PA guidelines in 2014-15 compared with 2008-09 (OR: 1.26. 95% CI: 1.12 – 1.42). Conclusion: Secular trends in PA differ across measurement methods. Self-reported physical activity indicates no or a negative change across the two timepoints, while device-measured physical activity indicates a positive trend. Practical implications: Trends in adherence to PA guidelines differ between assessment methods which needs to be considered when interpreting such data. Funding: NPASS is funded by the Norwegian Directorate of Health, the Public Health Institute of Norway, and the Norwegian School of Sport Sciences.

Background
The Australian Urban Observatory is a digital liveability planning platform measuring and monitoring liveability down to the local neighbourhood level across the 21 largest cities of Australia. Established in 2020, liveability indicators included in the Australian Urban Observatory (auo.org.au) are guided by a comprehensive definition of liveability with the digital portal developed and co-designed overtime in response to the needs to policymakers. The walkability of cities is commonly included in assessments of city liveability and is included in the Australian Urban Observatory but bikeability or infrastructure that supports safe cycling remains unmeasured in Australian cities due to a lack of available and appropriate data sources and the need for new technical and methodologocal approaches.

Purpose
This presentation will present a recently established and ambitious urban spatial research project that is testing new and innovative artificial intelligence and machine learning methods to comprehensively measure infrastructure supporting safe cycling in 40,000 neighbourhoods across Australian cities.

Methods
The transdisciplinary project draws on computer science, artificial intelligence, public health and urban policy to expertise to produce new objective and subjective indicators of cycling. This presentation will provide a review of the most viable machine learning techniques that can be applied to photogrammetry and sources of large scale data collection across Australian cities for the assessment of cycling infrastructure.

Practical implications
This research has been designed to assess an existing urban transport and policy gaps in cycling across Australia. It will be used to develop safe cycling indicators for local neighbourhoods to encourage the use of cycling as a sustainable and health promoting mode of transport. Results will identify neighbourhoods requiring cycling infrastructure intervention and the project will engage with an established national network of policymakers who are already engaged in the use of liveability indicators to inform planning.

Funding
Ian Potter Foundation, Australia.

Background: Self-report questionnaires are frequently used to monitor physical activity (PA) behavior. There is a considerable amount of research on the measurement properties of questionnaires. However, these studies rarely consider the linguistic characteristics of questionnaires and how those might influence their measurement properties. In other words, not all respondents might understand questionnaires that are difficult to read. Purpose: The aim of this study is to analyze the readability level of PA questionnaires and assess their adherence to the level recommended by the American Medical Association. Methods: The analysis included 45 commonly used PA questionnaires. The level of readability was calculated using the Flesch-Kincaid Reading Ease (FKRE) formula. Results: The calculation of readability levels showed that only five questionnaires reached a mean readability score within the recommended level. 22 questionnaires were classified as ‘fairly easy to read’ or ‘plain English’, and can be understood by those with basic education. The remaining questionnaires contained sentences or words rated as being long and complicated. 18 questionnaires were rated as being ‘fairly difficult to read’ to ‘very difficult to read’ and may require a high-school or college-level education to understand. Conclusions: The results suggest that most of the PA questionnaires analyzed do not meet recommended readability standards. As a consequence, individuals with lower levels of education might not always understand them. To increase the accuracy of data and improve the measurement properties of PA questionnaires, readability should be considered when constructing questionnaires. Practical implications: It is recommended to consider linguistic characteristics when developing, testing and translating questionnaires. Strategies for improving questionnaire readability and understandability may include cognitive testing and/or the collaboration with linguists. Funding: no outside funding

Background: Light is an important physiological input into children’s health and wellbeing, with a particular role in the timing of their circadian activity and sleep-wake processes. Increased use of artificial lighting suggests a broader range in the timing, intensity, spectra, and duration of everyday light exposure. However, the consequences of this exposure for children are not well understood. Purpose: The potential impacts of light from digital devices, specifically on children’s sleep and 24-hour activity, is controversial. In this study we explored the habitual light environments experienced by children. Methods: This feasibility and acceptability study included children aged 5- 8 years. Children wore a novel digital spectrometer on their chests during the day, and on their bedside during the night, together with an actigraph for 48-hours. Their carers completed a time-use diary (including physical and digital activities). The study examined the feasibility of this new form of ‘wearable’ via a brief evaluation interview with the children. Results: 20 children were recruited into the study (12 boys, 8 girls). Feasibility testing supported high acceptability and adherence to the study protocol. Well defined 24-hour spectral profiles (340-1010nm) were determined for each child. The light profiles were characterized with respect to estimated non-visual photoreception and circadian responses, in parallel with 24-hour activity and sleep-wake estimates. Conclusions: This study presents data that supports the feasibility of novel light exposure metrics. Understanding the impact of increased use of digital devices by children, the control of their light environments, and the interactions between light and activity requires new measures and protocols. Practical implications: These new tools are well accepted by children and provide unique data that will help improve understanding of the effects of light on children’s 24-hour activity and health. Funding: This project was supported by the ARC Centre of Excellence for the Digital Child (CE200100022).

Background: To adequately assess 24-hour movement behaviours, including physical activity (PA), sedentary behaviour (SB) and sleep in 0-4-year-old children, valid measurement instruments adapted to the child’s developmental stage are essential. However, the lack of a gold standard and valid accelerometer cut-points in 0-4-year-olds pose challenges for evaluating convergent validity of accelerometers and proxy-report tools. Purpose: We aimed to compare data from accelerometers and a proxy-report tool, the My Little Moves (MLM) app, by evaluating the consistency of accelerometer-derived accelerations across activity categories from the MLM app and corresponding 24-hour movement behaviours. We hypothesized that accelerometer-derived acceleration is lowest during sleep, followed by SB and PA. Methods: Hip- and wrist-worn accelerometer data from the MLM study, including 72 children aged 0-4 years (45.8% girls, 20.0 ± 11.5 months) were analysed. Parents (85.9% female, 35.8 ± 5.2 years) simultaneously recorded their children’s 24-hour movement activities (sleeping, sitting/lying, personal care, eating/drinking, passive and active screen use, passive and active transport, calm and active play) using the MLM app. General linear mixed-effects models were fitted to examine differences in acceleration across MLM app activity categories and corresponding 24-hour movement behaviours. Results: Acceleration differed significantly across all behaviours, confirming our hypothesis with the lowest values during sleep, then SB, and highest during PA. Acceleration was higher during the activity categories active play and active transport compared to sedentary activity categories, except for passive transport. Conclusions: The MLM app is promising for accurately assessing 24-hour movement behaviours in young children. Practical implications: Future studies should focus on combining data from both proxy-report tools and accelerometers and improving pattern recognition algorithms, particularly for passive transport. Exploring novel accelerometer data processing approaches will contribute to a more nuanced understanding of young children’s 24-hour movement patterns. Funding: ZonMw (546003008) and the Bernard van Leer foundation.

Background: Many researchers conduct a process evaluation alongside an effectiveness trial of a public health intervention. Yet, there are no standardized, scientifically accepted guidelines for reporting such process evaluations, which impedes interpretation and comparison of study results. Purpose: The aim of this project was to develop a consensus- and expert-based guideline for reporting process evaluations of public health interventions conducted alongside an effectiveness trial. Methods: We conducted an e-Delphi study with 73 international experts. Based on a scoping review, we proposed 32 items to panel members. In the subsequent two rounds, adjustments to the items and descriptions were proposed to the e-Delphi panel until consensus of ≥67% for each individual item was reached. Results: Based on purposive sampling, we invited 137 experts of which 73 (53%) participated in at least one round of the e-Delphi study. In total, 64 (88% of 73) completed round 2 and 55 (76% of 73) completed round 3. This resulted in 19 items that are included in the CONPHES guideline. Items include the description of the intervention, as well as the implementation strategies applied, and the role of the delivery and support team. The guideline also includes describing what framework was used for evaluation and how evaluation outcomes were assessed. The guideline is accompanied with an Explanation and Elaboration (E&E) document. Conclusions: The CONPHES reporting guideline can improve the reporting of process evaluations of interventions aimed at promoting public health. Practical implications: CONPHES can assist authors who are writing a process evaluation manuscript, or those writing the protocol for their planned process evaluation, and can be used by editors and reviewers to check if all necessary process evaluation information is described. Funding: This project was funded by the HBCD Project and Working Group Grant 2019 of the Amsterdam Public Health research institute.

Background: Assessing park use is challenging and time-consuming, especially if data is to be gathered over a longer period of time. Digital tools designed to collect long-term data might help us assess activity patterns in less resource-intensive ways. Purpose: We used an AI-based tool to explore long-term use of a neighborhood activity park. Methods: In this longitudinal monitoring study, a camera was installed at a new activity park in a Norwegian municipality. The camera has a built-in computer with an embedded AI-solution that quantifies moving objects directly from a video image flow. The AI-solution reads 30 images/second, and moving objects are continuously detected in each image. After detection is completed, images are deleted. Data is automatically transferred to Power BI that manage data using AI-capabilities. Through Power BI, we obtained information on monthly, daily, and hourly passings, and movement tracks. We analyzed data over 98 consecutive days, from 2 October 2023 to 7 January 2024, and linked it to daily weather data from the Norwegian Centre for Climate Services. Descriptive statistics were computed, and heat maps were used to visualize movement patters. Results: There were 804 daily passings in the park over the period. Highest frequency of activity was observed in the month after opening. In December, we witnessed a substantial drop in daily passing (mean=93), compared to October (mean=1516) and November (mean=1102). There was most activity in the park Fridays to Sundays between 9am-1pm. We observed a change in users’ movement pattern from mid-November, when the weather data indicated a snow depth of 15 cm. Conclusion: Although a longer observation period is needed, our results indicate that the park is largely used on weekends and that season characteristics influence park use. Practical implications: An AI-based monitoring tool can be useful for assessing activity in parks. Funding: Norwegian Research Council.