Distinguishing associations between neighbourhood features and physical inactivity, sedentary behaviour time, and screen time in boys and girls

Abstract

Objective

Physical inactivity (PI), sedentary behaviour time (SB) and screen time (ST) are related but distinct behaviors for which separate research and environmental intervention frameworks may be warranted. We examined associations between neighbourhood features and PI/SB/ST among boys and girls at risk of obesity at two timepoints, i.e., childhood (8–10 years old) and pre-adolescence (10–12 years old).

Methods

Data were from the QUALITY cohort, an ongoing study of the natural history of obesity in 630 Quebec families. Based on accelerometry, excess PI was defined as accumulating < 60 min/day of moderate to vigorous physical activity and excess SB as recording < 100 counts per minute for > 50% of wear time, and excess ST was based on self report and defined as reporting > 2 h/day of recreational ST. Neighbourhood features including presence of physical activity installations, green space, walkability, traffic indicators, physical disorder and foodscape indicators were measured using direct observation and geographic information systems. Neighbourhood features were measured when children were 8–10 years of age. Separate logistic regression models were estimated at each time point. Models controlled for child’s age, parental BMI, parental education, and area-level material deprivation.

Results

The odds of excess ST were lower in neighbourhoods with a higher number of parks, across all age and sex groups [ORs ranging from 0.70 (95% CI: 0.54–0.91) to 0.81(95% CI: 0.65–1.01)]. Among boys, the odds of excess SB were lower in neighbourhoods with more physical activity structures (OR: 0.44; 95% CI: 0.20–0.99); among girls, the odds of excess SB were lower in neighbourhoods with more sidewalks (OR: 0.67, 95% CI 0.47–0.95) and those that were exclusively residential (OR: 0.13, 95% CI: 0.04–0.45). Few neighbourhood features were associated with PI.

Conclusion

Our findings suggest that PI, SB and ST have both shared and distinct environmental determinants among children with parental obesity. While different patterns are likely to emerge across diverse contexts and populations, it remains relevant to consider that transforming specific features of the built environment may be more effective for some outcomes than others, and may not benefit all groups equally.

Supplementary information

The online version contains supplementary material available at 10.1186/s44167-025-00074-y.

Introduction

Clear concepts and guidelines are crucial for providing useful standards and meaningful threshholds in population health; these allow for adequate surveillance and can guide the evaluation of interventions in a standardized way. A case in point involves physical inactivity and sedentary behaviours, which are often used interchangeably in the literature, despite being distinct constructs [1]. Physical inactivity is usually defined as accumulating less than the recommended minimum minutes of moderate-to-vigorous physical activity (MVPA) (e.g. less than 60 min per day for youth aged 5 to 17 years) [2]. Sedentary behaviours, on the other hand, comprise very low energy expenditure behaviours (≤ 1.5 metabolic equivalent - MET) [3], and are often defined by time spent in pursuits such as sitting, watching television, using the computer and passive travel (i.e. sedentary behaviour time). A related concept is screen time, typically defined as passive leisure time use of screens [4]. Screen time is frequently used as a proxy for sedentary behaviours or physical inactivity [5], despite being conceptually and operationally distinct from both. A recent German study reported that children accumulated similar levels of screen time, regardless of their overall level of sedentary behaviour time [6]. Screen time may thus be largely independent of, and thus a poor indicator of, overall sedentary behaviour time.

Physical inactivity, sedentary behaviour time, and screen time also appear to have health consequences that vary in nature and/or magnitude. For example, sedentary behaviour time is generally associated with cardiometabolic risk in adolescents [7], while sedentary behaviour time is consistently associated with obesity more specifically [8]. Each construct also appears to have a unique set of determinants, which may vary across populations and contexts [9]. For example, age is consistently associated with sedentary behaviour time and screen time [10], while socioeconomic status is consistently associated with physical inactivity and screen time [11]. Because some situations are inherent to all three constructs (e.g. more sitting is associated with more physical inactivity, sedentary behaviour time, and screen time), some degree of overlap in their determinants is expected, albeit with differences in magnitude and associated effect modifiers.

Similarly, some, but not all environmental determinants are shared by these three behavioural outcomes. Reported environmental determinants of sedentary behaviour time in adolescents include less land use mix and connectivity, access to fewer green spaces, and worse traffic-related safety [10], while poor access to parks and to physical activity infrastructure is associated with more screen time [12]. Notably, these associations have primarily been observed in girls. There are several gender differences when it comes to levels of physical activity and engagement within the neighbourhood. Boys generally accumulate more physical activity than girls across all intensity levels, and at all ages, as concluded in a study synthesizing data from 9 countries [13]. Girls may also face additional barriers to engaging in physical activity such as parental safety concerns about the neighbourhood, lower parental encouragement overall, and fear of judgement from peers [14].

Importantly, potentially distinct and evolving causal mechanisms for each construct have implications for prevention and health promotion strategies. As these constructs are often used interchangeably, disentangling their specific contribution to specific outcomes is challenging. Additionally, investigating these associations separately by gender is needed in order to avoid embedding structural gender inequalities when desigining physical activity interventions. To extend this limited knowledge base, we investigated physical inactivity, sedentary behaviour time, and screen time as distinct behavioural outcomes, in a single paediatric population followed over a 2 year time period. Specifically, our objectives were to (1) estimate the associations of a broad range of neighbourhood features with each of physical inactivity, sedentary behaviour time and screen time in boys and girls and (2) compare relations at two different time points, in childhood (i.e. 8–10 years) and as participants approached adolescence (i.e. 10–12 years).

Methods

Study design and participants

This study was a repeated cross-sectional analysis using data from the Quebec Adipose and Lifestyle InvesTigation in Youth (QUALITY) cohort, an ongoing investigation of the natural history of obesity and cardiometabolic risk in youth with a history of parental obesity (n = 630). The sample was identical for all analyses. 1040 primary schools located within a 75-km radius of three major urban centers (Montreal, Quebec city, Sherbrooke) in the province of Quebec, Canada were identified for recruitment. Families self-selected to participate in the study through fliers that were distributed in 89% of these schools. Eligibility criteria comprised participants being 8–10 years at the time of recruitment (2005–2008), having at least one parent living with obesity, based on self-reported weight, height, and waist circumference; and both biological parents being available for the study. At the two year follow up, children were aged 10–12 years (n = 564). Data collection at both time points included questionnaires, accelerometry, and clinical assessments such as anthropometrics and fitness, among others [15]. Residential neighbourhoods were characterized in the same year as the first visit through on-site audits (Montreal only) and using a geographical information system [16]. For the current study, only Montreal participants (n = 512/630) were retained. Analyses were restricted to those with complete neighbourhood data (506/512), and who wore accelerometers at both visits (n = 326/506). Written informed consent was provided by all parents and children and ethical approval was granted by the CHU Sainte-Justine Research Centre Ethics Board and the Institut Universitaire de Cardiologie et Pneumologie du Québec Ethic Boards.

Measures

Neighbourhood environment variables

Neighbourhood features were measured either by audits or through geographic information systems (GIS). Selected variables include walkability measures, traffic density and traffic calming measures, parks and physical activity structures, residential buildings and density, and land use mix. Several studies report a negative association between sedentary behaviour and walkability measures such as residential density, street connectivity, and land use mix in both youth and adult populations [17, 18]. In adults, proximity to parks is associated with more frequent breaks in sedentary time [19], and greater access to parks is associated with higher physical activity in youth [20]. Higher traffic measures can be indicative of more motor vehicle commuting, thus increasing children’s sedentary behaviour [21]. Additionally, greater traffic is associated with reduced pedestrian safety and discourage children and their parents from engaging in active transport or leisurely activities [22]. We used these prior findings to select 17 neighbourhood features drawn from either neighbourhood audits or GIS to retain for analysis. Ten adjacent street segments connected to the participants’ dwelling were audited by trained research assistants using a validated neighbourhood audit tool. Those retained for analysis from the audits include presence of: sidewalks, pedestrian aids (marked crosswalks, pedestrian crossing lights and signs, all-direction stop signs at intersections, widened sidewalk at intersection, and designated ‘school corridor’ signs), traffic calming measures (speed bumps, mid-street segment stop signs, 30 km/h speed limit signs, traffic lights, large obstacles limiting traffic flow), signs of social disorder (graffiti, vandalism, litter, abandoned building/construction), physical activity structures (ex: outdoor playgrounds), and whether all buildings in the area were residential. In most instances, the number of features were summed and averaged across street segments. The presence of sidewalks was categorized as none, one, or on both sides of each street segment. The average of each feature across all audited street segments was then computed.

Other neighbourhood features were obtained through MEGAPHONE, a geographic information system (GIS) that characterizes social, built, and natural environmental factors in the Greater Montreal region using street network, land use and census data. The following indicators were computed for 1 km street-network buffers centered on the participant residences: number of three-way -or more- intersections, a land use mix measure; number of parks; percentage of streets with heavy vehicular rush hour traffic; total length of streets with vehicular traffic at rush hour, density of private dwellings per 10,000 m, and number and proportion of the 1 km street network buffer area covered by parks. Land use mix categories were determined from Leslie et al., and included commercial, open areas, parks and recreational areas, and residential types of land use [16, 23]. Land use mix ranged from 0 to 1, where 0 represents total homogeneity (all land within the area are of a single use), and 1 represents total heterogeneity (even distribution of all land use categories within the area) [23]. Heavy vehicular traffic at rush hour was categorised as < 1%, 1–5%, and ≥ 5% of neighbourhood streets with heavy traffic.

Outcomes

Children’s physical inactivity and sedentary behaviour time were assessed using a 7-day uniaxial accelerometer (Actigraph LS 7164 activity monitor, Actigraph LLC, Pensacola, FL, USA). Data were downloaded as one minute epochs and underwent standardized quality control. A minimum of 4 days of > 10 h of wear was required for data to be considered valid [24]. Nonwear time was defined as 60 min or more of 0 counts, including 1 to 2 consecutive minutes where count values were higher than 0 and lower or equal to 100 [24]. Sedentary behaviour time was calculated using standardized cut-points of < 100 counts per minute (CPM), excluding nonwear time [25]. The proportion of daily time spent in sedentary pursuits was calculated for each child, and excess sedentary behaviour time was defined as averaging ≥ 50% of accelerometer-wear time below 100 CPM (47). MVPA was calculated by summing all daily minutes spent in moderate or vigorous physical activity, and averaging the total over the eligible days of wear time. Children not engaging in the recommended 60 daily minutes of MVPA on average [26] were categorized as physically inactive. Screen time was measured using an interviewer-administered questionnaire, and included items on the time spent watching television, playing videogames, and using the computer during leisure activities, on weekdays and weekend days. A weighted average (5x weekday average + 2x weekend average /7) was computed to define participant screen time. Excessive screen time was defined as reporting an average of > 2 h daily screen time [10]. The same protocol was followed for both data collection time points.

Covariates

Relevant covariates were selected based on previous research and included participant age, maternal and paternal BMI, parental education, school level, and area-level material deprivation. Age and parental BMI were continuous variables. Parental education was categorized as zero vs. at least one parent with a university degree. School level was categorized as primary (grades 3 to 6) and secondary (grades 7 to 10) as per the Quebec schooling system. Area-level material deprivation as defined by Pampalon et al. [27] was centered around the mean score for the Montreal Metropolitan Area and included in models as a continuous variable.

Statistical analyses

Gender-specific associations between the 17 individual neighbourhood environment variables and outcomes of interest were estimated in separate multivariable regressions, adjusting for all covariates. For dichotomous outcomes (i.e.: excessive screen time, excessive sedentary behaviour time, and physical inactivity) the lowest risk group was the reference category. Statistical significance was set at p < 0.05 and illustrated by the 95% CI. All analyses were performed in IBM SPSS v26 and STATA v15 software. Additionally, sensitivity analyses were conducted using less stringent criteria to define risk categories (i.e.: ≥3 h of screen time, and < 30 min/day of MVPA).

Results

There were 326 participants with complete data at the first and second visits, including 183 boys and 143 girls (Table 1). The mean age at V1 was 9.5 years, and 11.6 years at V2. More than half of the sample had at least one parent with a university degree (57.2%). The mean maternal and paternal BMIs at the first visit were 29.4 and 30.6 kg/m² respectively.

Table 1.

Descriptive information for 326 QUALITY participants at first and second visits (Montreal, 2005–2011)

18% of participants did not have access to any parks in their neighbourhood, while half of the sample had access to 1 or 2 parks (Table 2). Three quarters of participants lived in neighbourhoods that were exclusively residential. Similarly, 73% of participants lived in neighbourhoods with low volumes of rush hour traffic. Almost 1 in 5 children (19%) lived in areas with sidewalks on all audited street segments within their neighbourhood, while just over 35% had no sidewalks at all in their neighbourhood.

Table 2.

Built environment features for 326 areas surrounding residential addresses of participants at the first visit

Built environment characteristics
Presence of at least 1 traffic calming feature, (%)	224 (68.7)
Presence of at least 1 traffic calming feature (no 30 km/h signs), (%)	143 (43.9)
Presence of at least 1 pedestrian facilitating feature, (%)	307 (94.2)
Presence of at least 1 pedestrian facilitating feature, (no 4-way stop signs) (%)	156 (47.9)
Presence of exterior playgrounds/recreational facilities, (%)	164 (50.3)
Presence of physical activity structures, (%)	166 (50.9)
Presence of at least 1 fast food restaurant, (%)	34 (10.4)
Presence of at least 1 convenience store, (%)	77 (23.6)
All buildings in area are residential, (%)	244 (74.8)
Presence of sidewalks, n (%)
All segments with sidewalks on both sides	60 (18.8)
High presence of sidewalks	53 (16.6)
Moderate presence of sidewalks	65 (20.3)
Low presence of sidewalks	28 (8.8)
No presence of sidewalks	114 (35.6)
Missing data, n (%)	6 (1.8)
Number of street intersections, n (SD)	77.7 (37.6)
Land use mix, (SD)	0.35 (0.15)
Density of private dwellings per hectare, n (%)
High Density	76 (23.3)
Average-to-high Density	84 (25.8)
Average Density	72 (22.1)
Average-to-low Density	59 (18.1)
Length of streets with normal vehicular traffic at rush hour, in km. (SD)	295.97 (196.91)
Length of streets with heavy vehicular traffic at rush hour, n (%)
0km	99 (30.4)
0.1km to 1km	95 (29.1)
1.1km to 5km	63 (19.3)
>5km	69 (21.2)
% of neighbourhood streets with heavy vehicular traffic at rush hour, n (%)
<1%	239 (73.3)
1–5%	62 (19.0)
>5%	25 (7.7)
Number of parks in neighbourhood, n (%)
No parks	59 (18.1)
1 park	79 (24.2)
2 parks	77 (23.6)
3 parks	44 (13.5)
4 parks	29 (8.9)
5 parks	38 (11.7)
Park Area Ratio, n (%)
<0.01	112 (34.4)
0.01 to 0.05	144 (44.2)
NDVI	0.33 (0.08)
Mean material deprivation index, (SD)	-0.018 (0.998)

Generally, boys were more physically active than girls, accumulating on average 57 min of MVPA per day compared to 41 min for girls at ages 8–10 years. Screen time and sedentary behaviour time did not differ by sex at either time point. Overall, at ages 8–10 years, 68% of participants were physically inactive (i.e. engaged in less than 60 min of MVPA per day on average), and accumulated on average 4.5 h per day of sedentary behaviour time, not correcting for wear time. 58% recorded excessive sedentary behaviour time (i.e., > 50% accelerometer weartime) and 53% reported 2 or more hours of screen time per day. At ages 10–12 years, 79% were categorized as physically inactive, the average daily sedentary behaviour time increased to 5.2 (SD: 0.86) hours per day, and 70% reported an excessive amount of time on screens. Screen time was weakly correlated with sedentary time and physical inactivity (0.09, 0.01), while physical inactivity and sedentary time were moderately correlated (0.3) (Supplementary Table 1).

Determinants of sedentary behaviour time

Neighbourhood features that emerged across the various age-sex groups included some that were shared across outcomes, and some that were unique to each outcome. Associations were generally stronger for girls, with fewer significant associations emerging in boys, in either age group. In the next paragraphs, associations are described for each of the three outcomes individually (Tables 3a and 3b).

Table 3a.

Results from fully adjusted models (BOYS)

Table 3b.

Results from fully adjusted models (GIRLS)

Physical inactivity

Among boys aged 8–10 years, the presence of neighbourhood disorder was associated with a higher likelihood of being physically inactive (OR 2.31, 95% CI 1.05–5.12). No other features were associated with physical inactivity in girls or in boys aged 10–12 years.

Excessive sedentary behaviour time

The presence of physical activity structures was strongly associated with a lower likelihood of excessive sedentary behaviour time in boys aged 8–10 years old (OR 0.44, 95% CI 0.20–0.99). No other determinants were meaningfully associated with this outcome among boys. Among girls, associations only emerged for the 10–12 year age group.

Specifically, a higher proportion of sidewalks (OR 0.67, 95% CI 0.47–0.95) and living in an exclusively residential neighbourhood (OR 0.13, 95% CI 0.04–0.45) were associated with lower odds of accumulating excessive sedentary behaviour time; conversely, the presence of neighbourhood disorder (OR 3.35, 95% CI 1.00-11.23) and greater land use mix (OR 2.10, 95% CI 1.09–4.03) were associated with a greater likelihood of accumulating excessive sedentary behaviour time.

Screentime

Each additional park in the neighbourhood was associated with a lower likelihood to report excessive screen time in boys and girls of both age groups.

Among girls aged 8–10 years, the presence of neighbourhood disorder was strongly associated with a lower likelihood of excessive screen time (OR 0.38, 95% CI 0.14–0.98). In the 10–12 years age group, living in an exclusively residential neighbourhood was associated with an increased likelihood of excessive screen time (OR 2.39, 95% CI 1.00-5.70), while a higher land use mix was associated with a lower likelihood of excessive screen time (OR 0.55, 95% CI 0.32–0.95).

Few meaningful associations were observed among boys. A higher proportion of sidewalks was associated with a higher likelihood of excessive screen time, while the number of street intersections was associated with a lower likelihood of excessive screen time at age 10–12 years.

The presence of traffic calming features, the presence of pedestrian facilitating features, and the density of private dwellings were not meaningfully associated with any outcome of interest. In sensitivity analyses, applying stricter definitions to physical inactivity and screen time did not affect our findings.

Discussion

The aim of this study was to investigate associations between neighbourhood-level features across three related but distinct outcomes: physical inactivity, sedentary behaviour time, and screen time. We compared each association independently, in a single population at two time points, in childhood and pre-adolescence. Fewer associations emerged overall in boys, compared to girls.

Nevertheless, distinct neighbourhood features were uniquely associated with each outcome, with screen time appearing to be the most “responsive” to built environment features. Notably, park access (frequency and ratio) was strongly associated with a lower likelihood of excessive screen time across all age-sex groups, but not with sedentary behaviour time or physical inactivity. While no other neighbourhood features were associated with excessive screen time among boys, several emerged among girls, including greater land use mix and the presence of neighbourhood disorder. This finding is similar to that of Christian et al. [28], who reported that a greater number of neighbourhood destinations was associated with a substantial decrease in screentime in girls, but not boys. In their study, girls who had access to 12 or more youth-related destinations recorded 109 fewer minutes/week of screen time on average compared to girls who had access to 3 or fewer destinations [28]. Additionally, some studies have shown that boys tend to use screens regardless of features of their neighbourhood, whereas neighbourhoods appear to have a stronger effect on screen time among girls [29]. Others have reported that girls who have access to a variety of destinations around the home may be more likely to go places with their parents or friends rather than engage in screen time at home [30].

Like others, our findings were mixed regarding sedentary behaviour time [31]. In our study, park frequency and land use mix were associated with a higher likelihood of excessive sedentary behaviour time in girls aged 10–12 years. Similarly, Marquet et al. reported that the majority of children spend less time in parks as they transition to adolescence, but this divergence is seen in girls more strongly [32]. In their study, energy expenditure decreased along with time spent in parks, indicating a potential increase in sedentary time, screen time, or both. We also reported that the presence of sidewalks and of neighbourhood disorder appeared to decrease the likelihood of sedentary behaviour time in girls of the same age group. Although an increase in neighbourhood disorder intuitively seems to have negative implications, a higher level of social disorder and the presence of sidewalks are inherent to more urban environments. These environments are known to provide greater access to destinations and activities that children may be able to access independently compared to children who do not live in urban environments.

Neighbourhood features assessed in our study appeared to be less meaningful for physical inactivity, across most age and sex groups with fewer associations emerging when compared to sedentary behaviour time and screen time. Only neighbourhood disorder emerged as a potential determinant of physical inactivity, and only among boys aged 8–10 years. Others have proposed that this could be supported by the idea of increased autonomy. This was reported in a systematic review [33] which concluded that children’s levels of physical activity were not determined by built environment features, contrary to expectations, and in contrast to adolescents. This may be due to greater parental involvement in young children’s pastimes, which may divert families beyond neighbourhood boundaries as defined in the current study.

All signficiant associations between our three outcomes and our investigated neighbourhood features were unique, which may be justified by the behavioural trade-offs inherent in these associations. For example, the presence of parks may reduce screen time by offering a competing activity, but may increase sedentary time if the park is used passively, as suggested by our findings. Similarly, features like sidewalks may encourage physical activity through active transportation but could inadvertently increase sedentary time if children walk to destinations such as cafes or gaming centers that promote sitting-based activities. These trade-offs illustrate that the effects of the built environment on movement behaviours are not uniform but instead interact with one another and vary significantly depending on social and contextual factors, such as gender, age, and the specific ways these spaces are used. Moreover, social norms and safety concerns can mediate how individuals engage with their surroundings. For instance, girls may benefit more from access to a variety of destinations, as they are more likely to substitute screen time with social outings or structured activities, whereas boys’ screen time appears less sensitive to environmental features.

As the prevalence of non-communicable diseases is rapidly increasing globally, our findings could have implications for future research. In Canada, 30% of children have overweight or obesity [34], which can lead to the development of other chronic diseases such as Type 2 diabetes, hypertension, and cardiovascular disease later in life [35]. Decreasing sedentary behaviour time through various incentives could have a positive impact on other movement behaviours such as physical activity. Moreover, adhering to recommended physical activity guidelines can help prevent, delay, and manage symptoms of these chronic illnesses [2]. In a recent report produced by the WHO, countries were called to prioritize physical activity as a key strategy for improving health and cardiometabolic disease [36]. As only 2 in 5 Canadian children met physical activity recommendations in 2018 [34], exploring new promotion strategies in this population is a public health priority. One successful promotion measure is modifying neighbourhood features in ways that enhance children and parents’ desire to engage with them. A Montreal-based intervention called “ruelles actives” (active alleys) [37] involves increasing traffic calming measures, greenery, and play facilities on alleyways to encourage children’s outdoor play. This easily implementable intervention is low-cost and could reduce the risk of physical inactivity. In our population, salient neighbourhood determinants differed for each outcome of sedentary behaviour time and this could be true in other populations as well. These differences must be well understood by policymakers to achieve intended targets when designing neighbourhood transformations. Future research should examine these determinants in other groups, and explore the development of targeted programs on marginalized groups to ensure that existing gender and urban inequities are not exacerbated through neighbourhood transformations.

Strengths and limitations

To our knowledge, ours is the first paper to compare the relation between neighbourhood features and physical inactivity, sedentary behaviour time, and screen time in the same population, at two developmental periods in childhood. Additionally, the use of objective data for physical inactivity and sedentary behaviour time as well as for the direct measurement of neighbourhood features is a strength of our work.

Some study limitations warrant consideration. First, the generalizability of our findings may be limited, as the QUALITY sample comprises children of Western European descent with a history of parental obesity; while this mirrors reality for a large proportion of Canadian families, replication in more diverse populations is needed. Second, the QUALITY cohort generally includes more socio-economically advantaged families than the average in the province of Quebec; nevertheless, a broad spectrum of neighbourhoods were represented. Third, while loss to follow-up was greater among families living in more disadvantaged areas, retention was almost 90%, and selection biases are unlikely to have meaningfully affected findings. Fourth, neighbourhood features were only measured at the first visit; while it is possible that some aspects of the environment changed over the two-year period, it is likely negligeable over this short time span. Fifth, while uniaxial accelerometers may be considered outdated, and a limitation of this study, several studies have shown that uniaxial and triaxial accelerometers demonstrate 95% concordance in free-living conditions [38]. Sixth, the large number of models tested could increase the risk of false positive results. However, our findings show clear and consistent trends, particularly regarding sedentary behavior time and physical inactivity, supporting the robustness of our conclusions. Finally, the proportion of children that moved from one neighbourhood to another over the 2 year period is low, and people generally move to similar types of neighbourhoods; any misclassification is likely to be non-differential misclassification and is unlikely to result in spurious findings.

Conclusion

In boys and girls approaching adolescence, physical inactivity, sedentary behaviour time, and screen time may have distinct neighbourhood determinants that evolve as children age and become more autonomous. We identified several neighbourhood features associated with reported screen time and excessive sedentary behaviour time, and comparatively few features with physical inactivity in our particular population. Moreover, we identified fewer neighbourhood determinants for boys in general, with many relations only apparent among girls. While different patterns are likely to emerge in different contexts and populations, it remains relevant to consider that transforming specific features of the built environment may be more effective for some outcomes than others, and may not benefit all groups equally. Decreasing sedentary behaviour through neighbourhood transformations could have positive impacts on other movement behaviour like physical activity. Policy makers should be mindful that these improvements do not exacerbate gender and urban inequalities.

Electronic Supplementary Material

Below is the link to the electronic supplementary material.

Author contributions

TAB and AEG conceptualised the study. MH is the lead of the QUALITY cohort project. MEM developed the algorithms used to obtain the physical activity accelerometry data that are part of our analysis. YK developed MEGAPHONE, the geographical information system used to obtain part of the built environment data analyzed in this project. AVH spearheaded the auditing process of our built environment data. AEG performed all data analysis and prepared all tables. AAL wrote the main manuscript text. All authors reviewed the manuscript and provided input.

Funding

The QUALITY cohort study is funded by the Canadian Institute of Health Research (CIHR), the Fonds de recherche du Quebec (FRQS), and the Heart & Stroke Foundation of Canada.

Data availability

Data described in this manuscript can be made available upon request.

Declarations

Ethics approval and consent to participate

Written consent was provided by all participants and ethical approval was granted by the CHU Sainte-Justine Research Centre Ethics Board and the Institut Universitaire de Cardiologie et Pneumologie du Québec Ethic Boards.

Competing interests

The authors declare no competing interests.

Clinical trial number

Not applicable.