Associations between the Neighborhood Socio-Spatial Environment and Physical Activity Correlates in Primary School Children: A Cross-Sectional Analysis

Objectives: Physical activity (PA) promotion in primary school children is a public health priority, but social inequalities need to be considered. However, empirical evidence to social inequalities in children´s PA is ambiguous. This study aimed to explore PA and PA correlates of primary school children depending on the neighborhood socio-spatial environment (NSSE) based on a socio-ecological model (SEM).

Methods: 1st grade primary school children (n=556, 48% girls) were examined cross-sectionally. Children’s PA and PA correlates were assessed with a parent questionnaire. The NSSE was classified into low (LD), partial (PD), and high deprivation (HD) operationalized by an urban development index. Data were analyzed using analysis of variance, Chi2, and post-hoc tests. 

Key results: Primary school children from HD neighborhoods were significantly less involved in organized PA, received less instrumental parental support, and were exposed to neighborhood environments perceived less supportive to PA compared to children from LD and PD neighborhoods. Furthermore, mothers from HD neighborhoods reported lower engagement in exercise-related PA.

Conclusions: PA and PA correlates of primary school children differed depending on the NSSE mainly to the disadvantage of children living in HD neighborhoods. Social inequalities are reflected in multiple levels of a SEM ranging from domain-specific PA to parental and environmental correlates. Future health promotion initiatives could use the NSSE as an indicator to address social inequalities in PA among primary school children in urban areas.

Key Words: Health, Social Inequalities, Deprivation, Community Health Promotion

Physical activity (PA) is a resource for the health and development of children. Despite its known benefits, 74 % of the primary school age children in Germany do not meet the WHO recommendations for PA, without noticeable improvements in recent years (9, 16). Consequently, promoting PA in childhood is a priority for public health.

In this context, research has highlighted the necessity to consider social inequalities. Empirical evidence indicates that the lower the children´s socioeconomic circumstances, the higher the need to promote PA (16, 32). This tendency is also evident for further health behaviors, health outcomes, and diseases illustrating the consequences of social inequalities for public health (22, 29).

As a result, socioeconomic indicators are used to address social inequalities. On the one hand, socioeconomic circumstances of children are operationalized using self-reported indicators at the individual level, usually education, occupation, and/or income of parents (32). On the other hand, socioeconomic indicators can also be measured at the environmental level, e.g., using census data on education, income, or unemployment of the neighborhood (1).

Measurements at the environmental level have received attention in recent years because they allow the monitoring of both socioeconomic and spatial circumstances in which children live (1). This can be useful for guiding community-based interventions (17). Such an approach is assumed to be in line with a socio-ecological model (SEM) taking a comprehensive view on health behavior as an interrelation between individuals and their environment (30).

The neighborhood has been identified as important for children’s PA (10) and can affect social inequalities both directly and indirectly (5, 19). The consideration of the neighborhood in a socio-spatial context is based on theories that the residence underlies social (e.g., social status, ethnicity, norms, cohesion) and geographic influences (e.g., costs, built environment, natural environment, infrastructure, land use) determining health and health-related behavior (5, 7). Empirical studies investigating the neighborhood socio-spatial environment and PA in primary school children are rare and ambiguous. One study found no association between neighborhood socioeconomic indicators and PA (39), while other studies observed domain-specific inequalities depending on the neighborhood socioeconomic status (6, 8), or even a higher PA of children living in socio-spatial deprived neighborhoods (27). Analyses shedding more light on the relationship between the neighborhood socio-spatial environment and children’s PA are recommended (6, 8, 27).

However, the sole consideration of PA levels is not sufficient to draw conclusions about social inequalities in PA (37). Following a SEM, PA is affected by individual, interpersonal, and environmental factors requiring knowledge of all these levels to better understand and target individual PA behavior (3, 31). Therefore, this study aimed to explore PA and PA correlates of primary school children depending on the neighborhood socio-spatial environment based on a SEM.

Study Design
Data of 1st grade primary school children participating in the KOMPASS(2) study were analyzed cross-sectionally. The KOMPASS(2) study is a longitudinal, school-based panel conducted in municipal primary schools from the city of Leipzig in Germany (38). Children’s participation was voluntary and written consent was obtained from parents or legal guardians after they received detailed information about the study. The KOMPASS(2) study was approved by the ethics committee of Leipzig University (File number: 253-14-14072014).

Measures
Children’s PA and PA correlates were assessed with a parent questionnaire during the school year 2018/2019. The questionnaire was distributed in 27 primary schools participating in the KOMPASS(2) study and was completed at home by a parent or legal guardian using paper and pencil. Respondents were asked to indicate whether the mother, father, or another legal guardian had completed the questionnaire. The instrument was designed to assess children’s PA, parental, household, and environmental PA correlates using validated items. Information on data collection, data preparation, and measurements used can be found in the KOMPASS(2) study protocol (38). The children’s neighborhood socio-spatial environment was operationalized using an urban development index of the city of Leipzig. The city of Leipzig is located in eastern Germany in the federal state of Saxony and currently has around 630,000 inhabitants. The index categorizes the city of Leipzig into socio-spatial neighborhoods with low (LD), partial (PD), and high deprivation (HD), measured by socioeconomic indicators of the residents (education, employment, income) and spatial concepts of urban development (e.g., housing, mobility, green spaces, social participation, educational infrastructure). Neighborhoods with HD are described by the city of Leipzig as areas with a high need for socioeconomic action in comparison to the rest of the city (36).

Data
Based on a SEM (3, 31), dependent variables were categorized into children’s PA (days per week with at least 90 minutes PA per day, achieving of at least 90 minutes PA per day, active travel to school, outdoor PA, self-organized exercise, membership in organized sports), parental correlates of children’s PA (parental support, parental beliefs in sport for good), parental PA (moderate-to-vigorous intensity PA per week, self-organized exercise, membership in organized sports, meeting WHO recommendation for PA), household resources (family health climate for PA, availability of PA equipment, availability of a garden), neighborhood environment (perception of attractiveness, safety, amenities, and location of the residential environment for PA), and school environment (availability of PA offers at school or in after-school care in addition to physical education, participation in PA offers at school or in after-school care in addition to physical education). The factor was the neighborhood socio-spatial environment categorized into LD, PD, and HD.

Data Analysis
Continuous dependent data were analyzed with univariate analysis of variance (ANOVA). The F-test was interpreted for equal variances and Welch ANOVA for unequal variances (11). Homogeneity of variance (p>0.05) was checked with Levene’s test. Statistically significant ANOVA results were further investigated with post-hoc tests using pairwise comparisons. Equal variances were interpreted wit Bonferroni, unequal variances with Games-Howell post-hoc test (2, 33). Chi2 tests were calculated for categorical dependent data. Data were analyzed with IBM SPSS Statistics Version 29 (SPSS, Chicago, IL, US). Statistical significance was set at α<0.05. 

Participants
The data of n=556 children (Ø 7.44 years, 48 % female, 15 % with migration background) were included in the statistical analyses. The descriptive characteristics of the sample are summarized in table 1. 

Main Results
Statistically significant differences depending on the neighborhood socio-spatial environment were observed for children’s PA, parental, household, and environmental correlates of PA. The main results are summarized in Table 2. 

Children’s PA
PA of children differed significantly in terms of self-organized exercise (p=0.01) and membership in organized sports (p<0.01), with the lowest rates among children living in HD neighborhoods. Children from HD neighborhoods were relatively more physically active for at least 90 minutes per day, without a statistically significant difference (p=0.05). Outdoor PA measured in days per week was significantly lower in children from PD compared to LD neighborhoods (p=0.04). No significant differences were observed for days per week with at least 90 minutes of PA per day (p=0.85), outdoor PA measured in minutes per day (p=0.67), and active travel to school (p=0.26). 

Parental Correlates of Children’s PA
Parental instrumental support was significantly lowest among children from neighborhoods with HD (p<0.01) and PD (p=0.02) compared to LD. No significant differences were observed for parental support through participation (p=0.07), emotional support (p=0.91), informational support (p=0.67), parental support for PA in total (p=0.25), and parental beliefs in sport for good (p=0.49). Mothers from HD neighborhoods reported the lowest moderate-to-vigorous PA (MVPA), with no statistically significant difference for MVPA in minutes per week (p=0.10) and meeting WHO recommendation for MVPA (p=0.31) compared to mothers from LD and PD neighborhoods. Mothers in HD neighborhoods reported significantly lower engagement in self-organized exercise (p<0.01) and were less likely to be members in organized sports with no significant difference (p=0.05). For father’s PA, no significant differences were observed depending on the neighborhood socio-spatial environment.

Household Resources 
Children from HD neighborhoods had less PA equipment available compared to LD (p<0.01) and PD (p<0.01), and a garden was significantly rarer in HD neighborhoods (p<0.01). No significant difference was observed in family health climate for PA (FHC-PA) in total (p=0.82), or the factors value (p=0.23), cohesion (p=0.50), and information (p=0.05).

Neighborhood Environment
Perception of the residential environment for PA differed significantly between LD, PD, and HD, with HD neighborhoods having the lowest ratings for attractiveness (p<0.01), safety (p<0.01), amenities (p<0.01), and location (p<0.01). 

School Environment
Schools in HD neighborhoods had significantly fewer PA offers at school or in after-school care in addition to physical education (p<0.01). Participation of children in school PA offers in addition to physical education did not differ significantly (p=0.06).

This study aimed to explore PA and PA correlates of primary school children living in socio-spatial neighborhoods with LD, PD, and HD. Overall, the results indicated inequalities in both PA and socio-ecological correlates of PA, mainly to the disadvantage of children from HD neighborhoods. 

Children’s PA
Children living in HD neighborhoods engaged significantly less in self-organized exercise and were less likely to be members in organized sports. Inequalities in organized PA were observed for extracurricular facilities, such as sports clubs, and were also evident in the school environment with significantly fewer PA offers available in HD neighborhoods in addition to physical education. This indicates that children from HD neighborhoods possibly have fewer opportunities to benefit from organized PA. As no significant difference was observed for children’s participation in PA offers at school or in after-school care, HD neighborhoods seem to be disadvantaged on a structural level regarding the availability of PA offers in these settings. However, participation in organized PA is considered important for children’s health and development because it is associated with meeting PA recommendations (21), motor skill development (35), and social participation (12). Instead, children’s daily PA did not differ significantly depending on the neighborhood socio-spatial environment. Children from HD neighborhoods were relatively, but not significantly, even more physically active for at least 90 minutes per day. This is in line with a previous study which found that children living in deprived neighborhoods are not necessarily less physically active than their peers in more affluent neighborhoods (27). However, representative data from Germany indicate that children from low socioeconomic backgrounds are less likely to achieve PA recommendations (16, 32). This indicates that children from deprived neighborhoods might not be disadvantaged in terms of overall PA levels but possibly regarding vigorous-intensity PA in organized settings, which has additional beneficial effects on children’s health and development. Similar findings have been reported in other studies (15, 24). The lower opportunities for children from HD neighborhoods to benefit from organized PA underline the need to promote PA offers with low-barrier access in key PA settings, e.g., by subsidized or free offers in schools, sports clubs, or recreational facilities close to the neighborhood (20). 

Parental Support and Role Model Behavior
Furthermore, children living in HD neighborhoods received significantly less instrumental support. Instrumental support for PA is characterized by parental support, e.g., through transportation, payment of fees, or the purchase of equipment (13). Possibly, a connection can be drawn to the findings related to the lower involvement of HD children in organized PA and the lower availability of PA equipment in the households of HD neighborhoods. As these aspects are associated with financial resources, the family’s financial situation possibly plays a crucial role in children’s participation in organized PA. This implication mirrors the emphasis of a SEM indicating an interrelation between children’s PA opportunities and parental resources. For FHC-PA, only the factor information showed a tendency to be higher in families from HD neighborhoods. FHC-PA for information covers “the search, sharing, and use of information related to sports and exercise” (26, p. 5). However, a possible explanation could be an interrelation with digital media use, which is higher among families with a low socioeconomic status (23). Furthermore, mothers from HD neighborhoods reported significantly less self-organized exercise and were relatively, but not significantly, less likely to be members in organized PA. Consequently, children from HD neighborhoods may receive less support through parental role modeling. In contrast, for father’s PA no significant differences were observed depending on the neighborhood socio-spatial environment. Family-based interventions that emphasize parental role model behavior and provide families from HD neighborhoods with instrumental support could possibly counteract social inequalities in children’s PA. A promising strategy may involve the promotion of family co-participation aiming at joint PA of children and their parents. Such an approach is consistent with a SEM by addressing benefits on multiple levels, e.g., children’s PA, parental support, role modeling, parental PA, and family quality time (25, 34).

PA Supportive Neighborhood Environment
Attractiveness, safety, amenities, and location of the residential environment for PA were all rated lowest in HD neighborhoods. This is a noticeable finding as a perceived unsupportive neighborhood environment, on the one hand, seems to restrict child-related PA, e.g., outdoor play, cycling, or independent mobility (4, 14, 18). On the other hand, parental perceptions of neighborhood safety and PA amenities have been identified as predictors of PA correlates in children, e.g., parental support (28) and family co-participation (34). These findings confirm the emphasis of a SEM on the interrelation between children’s PA, parental perceptions, and environmental characteristics. Especially because children from HD neighborhoods were found to be less involved in extracurricular organized PA, the neighborhood possibly is an important PA setting for deprived children. Urban planning efforts should consider these inequalities at the environmental level and promote a more supportive PA environment, particularly in HD neighborhoods. However, household-related characteristics, e.g., having a car and maternal education, were found to have an even greater impact on children’s PA than the neighborhood’s socioeconomic and physical environment (27).

Strengths and Limitations
A strength of this study was the exploration of PA correlates across multiple levels of a SEM, in addition to assessing children’s PA. The use of an urban development index from the city of Leipzig allowed for an objective classification of the children’s neighborhood socio-spatial environment. These objective data were combined with self-reported data on PA and PA correlates, providing a comprehensive view on social inequalities. However, the use of self-report measures is a limitation, as subjective data may be related to recall bias and social desirability bias. Furthermore, the distribution of children across socio-spatial neighborhoods with LD, PD, and HD was uneven. In particular, the number of children from HD neighborhoods was small, which possibly limits the generalizability of the findings. 

PA and PA correlates of primary school children differed significantly depending on the neighborhood socio-spatial environment and were mainly to the disadvantage of children from HD neighborhoods. These findings highlight that social inequalities in PA are reflected in multiple levels of a SEM ranging from domain-specific PA to parental and environmental PA correlates. Ultimately, reducing social inequalities in children’s PA requires multidimensional interventions at the parental, environmental, and institutional level with the need to consider interrelationships according to a SEM. The results can be incorporated into the urban development of the city of Leipzig and used for community health promotion. Future health promotion initiatives could use the neighborhood socio-spatial environment as an indicator to address social inequalities in PA among primary school children in urban areas. 

Conflict of Interest
The authors have no conflict of interest. All authors have completed the ICMJE Uniform Disclosure Form.

Funding
The research was funded by the European Social Fund (ESF) of the European Union and co-financed by the federal state of Saxony, Germany (Grant Number: 100670471).

Ethical Approval
The KOMPASS(2) study was granted by the ethics committee of the Faculty of Medicine, Leipzig University (File number: 253-14-14072014). 

Physical activity (PA) promotion in primary school children is a priority for public health, but social inequalities need to be considered. Empirical evidence on social inequalities in children’s PA is ambiguous, requiring the investigation of social inequalities beyond the sole consideration of PA levels. Therefore, this study aimed to explore PA and PA correlates of primary school children depending on the neighborhood socio-spatial environment classified into low, partial, and high deprivation.

Significant differences were observed in both children’s PA and PA correlates, mainly to the disadvantage of children living in socio-spatial neighborhoods with high deprivation. The findings highlight that social inequalities in PA are reflected in multiple levels of a socio-ecological model ranging from domain-specific PA to parental and environmental correlates. Future health promotion initiatives could use the neighborhood socio-spatial environment as an indicator to address social inequalities in PA among primary school children in urban areas.