Adolescent mental health is affected by neighborhood mental health conditions at the beginning of adolescence

doi:10.21203/rs.3.rs-2954690/v1

Download PDF

Article

Adolescent mental health is affected by neighborhood mental health conditions at the beginning of adolescence

https://doi.org/10.21203/rs.3.rs-2954690/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Neighborhood characteristics are thought to be associated with adolescent health and behavior. However, there have been very few population-based surveys using data from individual responses living in the same school district area. Tokyo TEEN Cohort data, collected from three Tokyo municipalities, were used to examine the association between psychological symptoms in school districts during early adolescence (age 10) and later psycho-behavioral conditions in adolescents (ages 12 and 14). Hierarchical linear models showed that school district-level depressive symptoms, psychological well-being, and prosocial behaviors at age 10 were associated with the corresponding individual-level indices at age 12, after controlling for demographic characteristics and baseline individual scores as covariates. Furthermore, School district-level behavioral problems at age 10 were associated with individual-level indices at ages 12 and 14. These results suggest that school district environments in which young adolescents grow up influence adolescents’ development and well-being.

Scientific community and society/Social sciences/Psychology

Health sciences/Health care/Public health/Epidemiology

school district

mental health conditions

population-based cohort study

hierarchical linear modeling

Japanese adolescents

The adolescent period is thought to last from age 10 to 19¹. The environment during childhood and early adolescence significantly influences adolescent psycho-behavioral problems^2–4. Previous studies have focused on the influence of neighborhood characteristics, such as socioeconomic status, social cohesion, neighborhood stressors, on adolescent mental health^5–7. For instance, King et al. found that adolescents who lived in persistently poor areas at age 15 have higher levels of depressive symptoms and anxiety based on a 3-year longitudinal survey⁸. However, despite growing interest in this topic, most studies use a cross-sectional design and lack population-representative longitudinal verification⁹. Thus, large-scale longitudinal surveys are needed to investigate the persistent impacts of neighborhood characteristics on adolescent mental health conditions.

Considering neighborhood effects during childhood and early adolescence, it must be noted that adolescents have limited physical mobility and tend to engage in daily activities within closer neighborhoods than adults. Elementary schools and their surrounding geographical areas, called school districts, are highly influential in the lives of adolescents. For example, the average daily activities of Japanese elementary students mapped using a global positioning system were 603 meters (1.14 km²) for second graders and 690 meters (1.49 km²) for fifth graders¹⁰, which mostly correspond to their school district area^11,12. Many previous studies have reported a relationship between school-level characteristics such as school features and size, class size, academic achievement, and the health conditions of adolescents^13–17. However, to our knowledge, the effects of regional and neighborhood characteristics at school-district level on psycho-behavioral problems in individuals have been rarely studied.

Although neighborhood effects at the school district level have not been directly examined, many studies have reported that close friends and family can significantly impact individuals’ mental health. For instance, a study on depression similarity indicated that individuals with similar levels of depressive symptoms have a greater chance of interacting with one another¹⁸. A best friend’s psychopathology, including internalizing and externalizing symptoms and personality pathology, has been reported to predict one’s own depression symptoms two to five years later¹⁹. Moreover, the average household income at the school district and individual levels, can also predict adolescent depressive symptoms, with differences in depressive symptoms observed between students who attend poor and wealthy schools²⁰. Therefore, evaluating health status from the peers in a district could impact on adolescents. Assessing the effects of psycho-behavioral characteristics within school districts on individuals requires the availability of neighborhood responses on residency^20–24, allowing us to examine impacts of group-level measures on individuals.

The Tokyo TEEN Cohort (TTC) study offers a valuable opportunity to examine the relationship between neighborhood characteristics at the school district level and adolescent mental health conditions. The TTC is a population-based cohort study that randomly sampled 3,171 children born between 2002 and 2004 living in 99 school district areas (with an average area of 0.97 km²; shown in Fig. 1) in three municipalities in Tokyo²⁵. The study collected data on various psychological symptoms and behavioral problems every two years. Each school district’s psychological health and behavioral scores were compiled from the data of 33 participants on average as previous studies^20–24. It is possible to create characteristic information about school districts using these data. By taking a longitudinal perspective, researchers can also better understand how these characteristics affect individual depressive symptoms, psychological well-being, and behavioral problems over time.

This study examined the effects of early adolescence psycho-behavioral status in school districts on psycho-behavioral conditions during adolescence in the TTC study. Individual measures of participants belonging to a specific school district were used to determine school district status. We intend to examine the extent and duration of the school district’s influence on each health indicator (depressive symptoms, psychological well-being, and self-regulation) to determine which factors predict sustained individual health status over an extended period.

We used TTC data for three waves (10 years old: n = 3,171; 12 years old: n = 2,982; n school district = 97; 14 years old: n = 2,543; n school district = 91; Table 1; Fig. 2). All participants’ correlation matrix and the mean scale scores in the schools are shown in Figs. 3 and 4.

Table 1

Demographic characteristics.
Variable	Baseline (age 10)			Wave 2 (age 12)			Wave 2 vs. Baseline		Wave 3 (age 14)			Wave 3 vs.Baseline
Variable	n (%)	M	SD	n (%)	M	SD	p-value	Hedge's g	n (%)	M	SD	p-value	Hedge's g
n	3159			2982					2543
Boys	1676 (53.0)			1575 (52.8)					1345 (52.8)
Age	3159	10.17	0.28	2982	12.17	0.31			2543	14.37	0.29
IQ	3159	108	14	2982	108	14			2543	108	14
Parental education level
Less than high school	6 (0.2)			6 (0.2)					5 (0.2)
High school	192 (6.1)			184 (6.2)					142 (5.6)
Two-year college	588 (18.6)			553 (18.5)					463 (18.2)
Four-year university	1828 (57.9)			1727 (57.9)					1488 (58.5)
Graduate university or higher	419 (13.3)			400 (13.4)					355 (14.0)
Not available	126 (4.0)			112 (3.8)					90 (3.5)
Maternal depressive symptoms	3159	0.49	0.55	2982	0.49	0.56			2543	0.49	0.56
School district
n	99			97					91
Number of children		512	201		517	200				534	192
Questionnaire measures
Depressive symptoms	3112	0.37	0.35	2500	0.31	0.35	< 0.001	0.16	1996	0.24	0.36	< 0.001	0.34
Psychological well-being	3156	3.95	0.82	2856	3.76	0.94	< 0.001	0.23	2354	3.50	1.01	< 0.001	0.50
Emotional symptoms	3158	0.32	0.34	2982	0.27	0.33	< 0.001	0.16	2542	0.26	0.36	< 0.001	0.07
Conduct problems	3158	0.37	0.31	2982	0.35	0.30	< 0.001	0.07	2542	0.35	0.30	0.025	0.04
Hyperactivity/Inattention	3157	0.61	0.43	2981	0.54	0.42	< 0.001	0.15	2542	0.52	0.32	< 0.001	0.20
Peer problems	3158	0.31	0.32	2982	0.31	0.32	< 0.001	0.20	2542	0.33	0.33	< 0.001	0.08
Prosocial behaviors	3157	1.34	0.41	2980	1.30	0.43	< 0.001	0.09	2542	1.19	0.46	< 0.001	0.38

The summary of hierarchical linear models (HLMs) are shown in Table 2, Supplementary Table S1, and Fig. S1 indicates fixed effects of independent and confounding variables and random slopes. The model incorporates depressive symptoms at age 12 or 14 as the dependent variable, school district means of the depressive symptom scores at age 10 as the fixed effect of the independent variable, individual depressive symptoms at age 10, confounding factors as the fixed effect of the confounding variables, and a random effect of slope of depressive symptoms in school as the mixed effect. The HLM showed that higher school district-level depressive symptoms at age 10 were associated with higher individual depressive symptoms at age 12 (γ = 0.510, SE = 0.096 [95%CI: 0.322, 0.698], false discovery rate [FDR]-corrected p < 0.001), but not at age 14 after multiple corrections (γ = 0.265, SE = 0.110 [95%CI: 0.050, 0.480], uncorrected p = 0.016). For psychological well-being, the HLMs showed that school district-level psychological well-being at age 10 were associated with individual depressive symptoms at age 12 (γ = 0.374, SE = 0.103 [95%CI: 0.171, 0.576], FDR-corrected p < 0.001), but not at age 14 (γ = 0.216, SE = 0.126 [95%CI: -0.030, 0.462], uncorrected p = 0.086).

Table 2

Fixed effect of individual and school district variables.
Measure (Baseline: age 10)	Variable	Wave 2 (age 12)				Wave 3 (age 14)
Measure (Baseline: age 10)	Variable	γ	SE	95% CI	p-value	γ	SE	95% CI	p-value
Depressive symptoms	Individual	0.433	0.023	[0.388, 0.478]	< 0.001	0.279	0.027	[0.226, 0.332]	< 0.001
	School district	0.510	0.096	[0.322, 0.698]	< 0.001	0.265	0.110	[0.050, 0.480]	0.016
Psychological well-being	Individual	0.433	0.023	[0.387, 0.478]	< 0.001	0.306	0.026	[0.254, 0.358]	< 0.001
	School district	0.374	0.103	[0.171, 0.576]	< 0.001	0.216	0.126	[-0.030, 0.462]	0.086
Emotional symptoms	Individual	0.494	0.019	[0.457, 0.531]	< 0.001	0.458	0.026	[0.407, 0.509]	< 0.001
	School district	0.614	0.071	[0.475, 0.754]	< 0.001	0.646	0.077	[0.494, 0.797]	< 0.001
Conduct problems	Individual	0.594	0.018	[0.558, 0.630]	< 0.001	0.539	0.021	[0.499, 0.579]	< 0.001
	School district	0.605	0.076	[0.457, 0.753]	< 0.001	0.364	0.088	[0.192, 0.537]	< 0.001
Hyperactivity/Inattention	Individual	0.651	0.016	[0.620, 0.681]	< 0.001	0.359	0.016	[0.329, 0.390]	< 0.001
	School district	0.757	0.078	[0.605, 0.909]	< 0.001	0.437	0.070	[0.299, 0.574]	< 0.001
Peer problems	Individual	0.590	0.020	[0.550, 0.630]	< 0.001	0.512	0.028	[0.457, 0.567]	< 0.001
	School district	0.699	0.073	[0.555, 0.843]	< 0.001	0.574	0.081	[0.416, 0.733]	< 0.001
Prosocial behaviors	Individual	0.626	0.018	[0.592, 0.660]	< 0.001	NA	NA	NA	NA
	School district	0.606	0.087	[0.435, 0.777]	< 0.001	NA	NA	NA	NA
The model for prosocial behaviors at age 14 did not fit well, even when using the random effect of the intercept of prosocial behaviors or the random effects of the intercept and slope of prosocial behaviors.

A correlation was found between school district-level emotional symptoms at age 10 and individual emotional symptoms at age 12 (γ = 0.614, SE = 0.071 [95%CI: 0.475, 0.754], FDR-corrected p < 0.001), and at age 14 (γ = 0.646, SE = 0.077 [95%CI: 0.494, 0.797], FDR-corrected p < 0.001). School district-level conduct problems at age 10 were associated with individual conduct problems at age 12 (γ = 0.605, SE = 0.076 [95%CI: 0.457, 0.753], FDR-corrected p < 0.001), and at age 14 (γ = 0.364, SE = 0.088 [95%CI: 0.192, 0.537], FDR-corrected p < 0.001). School district-level hyperactivity/inattention at age 10 were associated with individual hyperactivity/inattention at age 12 (γ = 0.757, SE = 0.078 [95%CI: 0.605, 0.909], FDR-corrected p < 0.001) and at age 14 (γ = 0.437, SE = 0.070 [95%CI: 0.299, 0.574], FDR-corrected p < 0.001). School district-level peer problems at age 10 were associated with individual peer problems at age 12 (γ = 0.699, SE = 0.073 [95%CI: 0.555, 0.843], FDR-corrected p < 0.001) and at age 14 (γ = 0.574, SE = 0.081 [95%CI: 0.416, 0.733], FDR-corrected p < 0.001). School district-level prosocial behaviors at age 10 were associated with individual prosocial behaviors at age 12 (γ = 0.606, SE = 0.087 [95%CI: 0.435, 0.777], FDR-corrected p < 0.001). However, the intercept and slope of prosocial behaviors did not fit well in the model of prosocial behaviors at age 14.

We used a prospective adolescent cohort²⁵ to examine the association between adolescent psycho-behavioral conditions and the school district environment. We can calculate school district health status based on individual-level responses since the TTC has a relatively narrow catchment area compared with other national cohorts. We found that school district-level depressive symptoms, psychological well-being, and prosocial behaviors at age 10 were associated with the corresponding individual-level indices at age 12 after considering for baseline individual scores as covariates. School district-level behavioral problems at age 10 were associated with individual-level indices at ages 12 and 14. These results suggest that the school districts’ psychological environment has an impact on individual-level health conditions in adolescence.

This study applied a hierarchical linear model that used both individual and school district data, enabling us to investigate the time-based effects of health status at both levels. After controlling for baseline factors such as age, gender, IQ, SES, main caregivers’ depressive symptoms, and early adolescent health conditions, the results revealed that adolescents’ health status influences health status during adolescence and in school district level at the beginning of adolescence. Consequently, the school district environment influences an individual’s psycho-behavioral problems. For instance, studies investigated the impact of economic disparities in school settings have revealed that adolescent depressive symptoms are influenced by their household income and also by the average household income in the school district²⁰. Additionally, previous research has demonstrated that peer relationships and peer health status can contribute to depressive symptoms^18,19, and that socioeconomic status at the community level can have an impact on adolescents psycho-behavioral problems^{5–8,20−24}. The present study offers additional evidence that the mental health status of the school district can play a role in the transmission of individuals’ psycho-behavioral problems, which is in line with previous findings on the impact of schools and communities.

The impact of depressive symptoms and well-being at the school district level lasted for two years, while behavioral problems lasted four years. Variations in impact duration may be attributed to individual-level time-series correlations. Previous research has indicated that depressive symptoms and psychological well-being are not stable over time during adolescence^26,27, while behavioral problems have been shown to be relatively stable^28,29. This study showed that the correlation between depressive symptoms, well-being, and age was weak, and persistence was not as strong as that of behavioral problems. Furthermore, prosocial behavior showed a boundary fit and could not be tested using the random-effects model. Prosocial behavior in children can increase or decrease from childhood to adolescence, with inconsistent results^30–34. These results suggest that depressive symptoms, psychological well-being, and prosocial behavior may change over time, and that the period in which the school district’s health may have an impact is shorter than that of behavioral problems.

Neighborhood characteristics in the daily activity area of school commuting have an impact on adolescent health. Furthermore, the duration of the impact of school district health conditions on individual health conditions may vary depending on depressive symptoms, psychological wellbeing, and self-regulation and the uniqueness of these concepts. However, we also found commonalities in that district-level variables were associated with individual-level variables up to approximately two years later for indicators of health and prosocial behavior and up to approximately four years later for indicators of problematic behavior. The results of this study suggest that consistency in areal difference studies involving adolescents could be improved by restricting the analysis to areas where they are habitually active. We hypothesized that the differences in school district health status effects would be related to time-series correlations at the individual level. However, it should be noted that while depression symptoms and mental health were self-reported by adolescents, problem behaviors and social behaviors were evaluated by parents, which may have affected the results. However, given that no previous studies have investigated the health status of school districts, cautious interpretation and accumulation of research findings are necessary. Further studies are required to ascertain whether similar concepts, not examined in this study, can be categorized using the methodology described above.

This study has several limitations. First, the analyzed school district area could not be fully characterized by specific neighborhood characteristics. Therefore, it is necessary to extract neighborhood characteristic factors for each school district and examine which parts factors affect adolescent health. Second, this study used the total number of children, indicating school size as a confounding factor. However, not all school factors could be excluded. Future studies should examine the factors that affect school characteristics, size, and class size. Third, the limited scope of the analysis and reproducibility of the results need to be checked by expanding the area of analysis. The area analyzed in this study was part of Tokyo, the capital of Japan, and it is undeniable that the results may change if the analysis area differs. More detailed research is needed, including studies on school district areas in other regions and international comparisons.

These findings indicate that depressive symptoms, psychological well-being, and self-regulation in a school district are associated with depressive symptoms, psychological well-being, and self-regulation in their adolescents. Unlike most previous studies that used cross-sectional analyses, this study employed a population-representative longitudinal approach, allowing for a more comprehensive examination of the effects on adolescent health, including various indicators and temporal causality.

It is important to note that the duration for which health status at the school district level affects the health status of individuals depends on various factors. However, it was also found that there were commonalities where depressive symptoms, psychological well-being, and prosocial behavior were predictors of school district variables for roughly two years, and problem behaviors (emotional symptoms, conduct problems, hyperactivity/inattention, and peer problem) for about four years. Uncovering the relationship between neighborhood and individual levels from the perspective of school districts, an area that has not been previously examined, is crucial for obtaining a holistic understanding of adolescent psychopathological issues.

Participants and procedures

We used TTC data for three waves (10 years old: 2012–2015, n = 3, 171; 12 years old: 2014–2017, n = 2,982, n school district = 97; 14 years old: 2016–2019, n = 2,543, n school district = 91. Shown in Table 1 and Fig. 2). This study was approved by the Institutional Review Board of the Faculty of Medicine, University of Tokyo (approval numbers: 3150, 10057, and 10069), Graduate University for Advanced Studies (SOKENDAI; 2012002), and Tokyo Metropolitan Institute of Medical Science (12–35). Informed consent was obtained from the children’s primary caregivers (usually their mothers) before participation.

To obtain information about the school district, we used the websites of the three municipalities (Setagaya: https://www.city.setagaya.lg.jp/; Chofu: https://www.city.chofu.tokyo.jp/; and Mitaka: http://www.city.mitaka.tokyo.jp/). As the school district was determined by the year, we set the area definition based on information from 2014, when the survey in the first wave was conducted. In addition, we contacted government offices for further information. Finally, we used the school districts of 99 elementary schools: 64 schools in Setagaya Ward, 20 in Chofu City, and 15 in Mitaka City (Fig. 1). To define the cohort members’ school districts using their addresses, we used a coding program to obtain their school districts and ran the program on an offline PC, which stored their personal information in the Tokyo Metropolitan Institute of Medical Sciences. Data sorted in school districts (1–99) were extracted and used for further analysis to maintain anonymity.

Measures

For every group, we obtained self-reported anonymous questionnaires from the children, including the Japanese versions of the Short Mood and Feelings Questionnaire (SMFQ) and the World Health Organization Well-Being Index (WHO-5). In the interview phase, at age 10, we conducted a short version of the Wechsler Intelligence Scale for Children-Third Edition, the Japanese version (WISC-III). Using anonymous questionnaires for the main caregivers, we obtained the Japanese version of The Strengths and Difficulties Questionnaire (SDQ) for their children and confounding variables. For every subscale and scale of the questionnaire comprising five or more items, the average value of the remaining item scores was substituted for analysis if the missing value was within 20% of the items; otherwise, it was coded as missing.

Depressive symptoms

The SMFQ was used to assess children’s depressive symptoms and consisted of 13 self-reported questions (e.g., I felt so tired I just sat around and did nothing)^25,35,36. Respondents chose their answers using a 3-point Likert scale (2 = true; 0 = not true). Higher scores indicate more severe depressive symptoms.

Psychological well-being

The WHO-5 assessed children’s psychological well-being^37–39. It is a 5-item questionnaire (e.g., I have felt cheerful and in good spirits). Respondents were asked to choose answers using a 6-point Likert scale (5 = all of the time; 0 = at no time). Higher scores indicate better psychological well-being.

The Strength and Difficulties questionnaire

The SDQ has been used to assess children’s self-regulation^40–43. The scale consists of 25 items in five subscales: emotional symptoms (e.g., often complains of headaches, stomach-aches, or sickness), conduct problems (e.g., often loses temper), hyperactivity/inattention (e.g., restless, overactive, cannot stay still for long), peer problems (e.g., rather solitary, prefers to play alone), and prosocial behavior (e.g., considerate of other people’s feelings). The scale was assessed using parental questionnaires, in which parents were asked to choose answers using a 3-point Likert scale (2 = certainly true; 0 = not true). The higher the prosocial scores, the lower the children’s strength. For the remaining four subscales, the higher the score, the greater the child’s difficulty.

Confounding variables

Confounding variables included children’s age, gender, intelligence quotient (IQ), parental educational attainment as socioeconomic status (SES), depressive symptoms, and the total population of elementary school students in 2014. For children’s IQ at age 10 years, the 2-battery short version of the WISC-III were used (Information and Picture Completion)^44,45. SES was estimated using the higher academic attainment of their care givers (“1 = junior high school graduate (or under),” “2 = high school dropout,” “3 = high school graduate,” “4 = technical school or 2-year college graduate,” “5 = four-year college graduate,” and “6 = graduate university or six-year college graduate”) (Kanata et al., 2016) ⁴⁶. For the caregivers’ main depressive symptoms, the Japanese version of the Kessler 6 was assessed through an anonymous questionnaire^47,48. The average value was substituted for any missing child’s IQ (n = 3, 0.0009%) and parental SES (n_father = 154, 0.05%n_mother = 25, 0.08%). The total population of elementary school students in 2014, as stated on Gaccom, a school education website in Japan (https://www.gaccom.jp/), was used to measure the school scale (Fig. 1).

Statistical analyses

To test the effects of school district health conditions on each individual variable, we applied an HLM that allowed for the assumption of random effects^20–24. The model included individual and school district psychological conditions as independent variables, corresponding individual psychological conditions at Wave 2 or 3 as dependent variables, and confounding factors (child’s age at baseline survey, sex, child’s IQ, parental educational attainment, main caregiver’s depressive symptoms, and total number of children in the school) using a restricted maximum likelihood. To control for multicollinearity between group- and individual-level scores, both individual- and school-district-level scores were centered on the group mean^49,50.

Previous studies have found that the community influences a person’s health factors^5–8, and experience and years of residence within a community may influence an individual’s health conditions. In this study, the intercept baseline is the same across school districts. Still, the slope differs across school districts since this study aims to determine whether health status at the school district level can be predicted from health status at the individual level. Therefore, the slope of each health indicator may vary according to the school district. Since this study focuses on the slope of whether health conditions at the school district level can be predicted from health conditions at the individual level, it was tested using a random slope model in which the intercept baseline was equal across school districts and the slope differed across districts (see Supplementary materials and Fig. S1).

Statistical analysis packages included summarytools⁵¹, mice⁵², dplyr⁵³, psych⁵⁴, effsize⁵⁵, corrplot⁵⁶, ggplot2⁵⁷, leaflet⁵⁸, and lmerTest (optimizer = "bobyqa")⁵⁹ packages in R version 4.1.0⁶⁰. The significance level was corrected using the Bonferroni method, as the analysis was performed using seven predictive health measures (p < 0.0071 = 0.05/7).

World Health Organization. Adolescent health. https://www.who.int/health-topics/adolescent-health (2023).
Dalton, E. D., Hammen, C. L., Brennan, P. A. & Najman, J. M. Pathways maintaining physical health problems from childhood to young adulthood: the role of stress and mood. Psychol Health 31, 1255–1271 (2016).
Im, Y., Oh, W. O. & Suk, M. Risk factors for suicide ideation among adolescents: five-year national data analysis. Arch Psychiatr Nurs 31, 282–286 (2017).
Patton, G. C. & Viner, R. Pubertal transitions in health. Lancet 369, 1130–1139 (2007).
Coley, R. L., Sims, J., Dearing, E. & Spielvogel, B. Locating economic risks for adolescent mental and behavioral health: poverty and affluence in families, neighborhoods, and schools. Child Dev 89, 360–369 (2018).
Dawson, C. T. et al. Perceived neighborhood social cohesion moderates the relationship between neighborhood structural disadvantage and adolescent depressive symptoms. Health Place 56, 88–98 (2019).
Seiter, N. S., Lucas-Thompson, R. G. & Graham, D. J. Interparental conflict moderates the association between neighborhood stress and adolescent health. J Fam Psychol 33, 253–258 (2019).
King, C., Huang, X. & Dewan, N. A. Continuity and change in neighborhood disadvantage and adolescent depression and anxiety. Health Place 73, 102724 (2022).
Fyfe-Johnson, A. L. et al. Nature and children’s health: a systematic review. Pediatrics 148, e2020049155 (2021).
Shimada, T. et al. Measuring routine activities of elementary school children using GPS: an example at an public elementary school in a Kobe, Japan. Theor. Appl. GIS 18, 85–91 (2010).
Amemiya, M. The area difference of the relationship between children’s independent mobility measured by parental license and children’s outdoor play. J. Jpn. Inst. Landsc Archit. 75, 477–482 (2012).
Seto, D., Taniguchi, A. & Ishida, H. (2019). Influence on situation of commuting to school on children's physical and mental health. J. Jpn. Soc. Civ. Eng. Ser. D3 (Infrastructure Plan Manag.) 75, I_1069–I_1079 (2019).
Aldridge, J. M. & McChesney, K. The relationships between school climate and adolescent mental health and wellbeing: a systematic literature review. Int J Educ Res 88, 121–145 (2018).
Browning, M. H. & Rigolon, A. School green space and its impact on academic performance: a systematic literature review. Int. J. Environ. Res. Public Health 16, 429 (2019).
Byrd, C. M. & Chavous, T. M. Racial identity and academic achievement in the neighborhood context: a multilevel analysis. J Youth Adolesc 38, 544–559 (2009).
Filges, T., Sonne-Schmidt, C. S. & Nielsen, B. C. V. Small class sizes for improving student achievement in primary and secondary schools: a systematic review. Campbell System. Rev. 14, 1–107 (2018).
Konold, T., Cornell, D., Jia, Y. & Malone, M. School climate, student engagement, and academic achievement: a latent variable, multilevel multi-informant examination. AERA Open 4, 1–17 (2018).
Elmer, T. & Stadtfeld, C. Depressive symptoms are associated with social isolation in face-to-face interaction networks. Sci. Rep. 10, 1444 (2020).
Raposa, E. B., Hammen, C. L. & Brennan, P. A. Close friends' psychopathology as a pathway from early adversity to young adulthood depressive symptoms. J Clin Child Adolesc Psychol 44, 742–750 (2015).
Goodman, E., Huang, B., Wade, T. J. & Kahn, R. S. A multilevel analysis of the relation of socioeconomic status to adolescent depressive symptoms: does school context matter? J. Pediatr. 143, 451–456 (2003).
Dykxhoorn, J. et al. Objective and subjective neighbourhood characteristics and suicidality: a multilevel analysis. Psychol Med 53, 1166–1175 (2023).
Kristjansson, A. L., Sigfusdottir, I. D., & Allegrante, J. P. Adolescent substance use and peer use: a multilevel analysis of cross-sectional population data. Subst Abuse Treat Prev Policy 8, 1-10 (2013).
Liu, F., Gai, X., Xu, L., Wu, X. & Wang, H. School engagement and context: A multilevel analysis of adolescents in 31 provincial-level regions in China. Front. Psychol. 12, 724819 (2021).
Maes, L. & Lievens, J. Can the school make a difference? A multilevel analysis of adolescent risk and health behaviour. Soc. Sci. Med. 56, 517–529 (2003).
Ando, S. et al. Cohort Profile: The Tokyo Teen Cohort study (TTC). Int J Epidemiol 48, 1414–1414g (2019).
López-López, J. A. et al. Depressive symptoms and academic achievement in UK adolescents: a cross-lagged analysis with genetic covariates. J. Affect. Disord. 284, 104–113 (2021).
Winzer, R., Vaez, M., Lindberg, L. & Sorjonen, K. Exploring associations between subjective well-being and personality over a time span of 15–18 months: a cohort study of adolescents in Sweden. BMC Psychol. 9, 1–10 (2021).
Beeres, D. T., Andersson, F., Vossen, H. G. & Galanti, M. R. Social media and mental health among early adolescents in Sweden: a longitudinal study with 2-year follow-up (KUPOL Study). J Adolesc Health 68, 953–960 (2021).
Donati, G., Meaburn, E. & Dumontheil, I. Internalising and externalising in early adolescence predict later executive function, not the other way around: a cross-lagged panel analysis. Cogn Emot 35, 986–998 (2021).
Eisenberg, N., Carlo, G., Murphy, B. & Court, P. Prosocial development in late adolescence: a longitudinal study. Child Dev 66, 1179–1197 (1995).
Fabes, R. A., Carlo, G., Kupanoff, K. & Laible, D. Early adolescence and prosocial/moral behavior I: the role of individual processes. J Early Adolesc 19, 5–16 (1999).
Carlo, G. & Randall, B. A. The development of a measure of prosocial behaviors for late adolescents. J Youth Adolesc 31, 31–44 (2002).
Kokko, K., Tremblay, R. E., Lacourse, E., Nagin, D. S., & Vitaro, F. Trajectories of prosocial behavior and physical aggression in middle childhood: links to adolescent school dropout and physical violence. J Res Adolesc 16, 403–428 (2006).
Malti, T. et al. Children’s trust and the development of prosocial behavior. Int J Behav Dev 40, 262–270 (2015).
Angold, A., Costello, E. J., Messer, S. C. & Pickles, A. Development of a short questionnaire for use in epidemiological studies of depression in children and adolescents. Int J Methods Psychiatr Res 5, 237–249 (1995).
Ando, S. et al. Help-seeking intention for depression in early adolescents: Associated factors and sex differences. J. Affect. Disord. 238, 359–365 (2018).
Awata, S. et al. Reliability and validity of the Japanese version of the World Health Organization-Five Well-Being Index in the context of detecting depression in diabetic patients. Int J Methods Psychiatr Res 61, 112–119 (2007).
Awata, S. et al. Validity and utility of the Japanese version of the WHO-Five Well-Being Index in the context of detecting suicidal ideation in elderly community residents. Int Psychogeriatr. 19, 77–88 (2007).
Bech, P. Measuring the dimension of psychological general well-being by the WHO-5. Quality of Life Newsletter 32, 15–16 (2004).
Noda, W. et al. Examining the reliability and validity of the Japanese version of the Strengths and Difficulties Questionnaire Self Rating Form using the entire cohort data in one suburban city in Japan. Jpn J Clin Psychiatry 42, 119–127 (2013).
Matsuishi, T. et al. Scale properties of the Japanese version of the Strengths and Difficulties Questionnaire (SDQ): a study of infant and school children in community samples. Brain Dev. 30, 410–415 (2008).
Moriwaki, A. & Kamio, Y. Normative data and psychometric properties of the Strengths and Difficulties Questionnaire among Japanese school-aged children. Child Adolesc. Psychiatry Ment. Health 8, 1–12 (2014).
Goodman, R. The strength and difficulties questionnaire: a research note. J Child Psychol Psychiatry 38, 581–586 (1997).
Inada, N. & Kamio, Y. Short forms of the Japanese version WISC-III for assessment of children with autism spectrum disorders. Jpn J Child Adolesc Psychiatr 51, 11–19 (2010).
Wechsler, D. The Wechsler intelligence scale for children-third edition (The Psychological Corporation, 1991).
Kanata, S. et al. Enuresis and hyperactivity-inattention in early adolescence: findings from a population-based survey in Tokyo (Tokyo Early Adolescence Survey). PLoS One 11, e0158786 (2016).
Furukawa, T. A. et al. The performance of the Japanese version of the K6 and K10 in the World Mental Health Survey Japan. Int J Methods Psychiatr Res. 17, 152–158 (2008).
Kessler, R. C. et al. Short screening scales to monitor population prevalences and trends in non-specific psychological distress. Psychol Med 32, 959–976 (2002).
Kawabata, K., Iwama, T. & Suzuki, M. (2018). Introduction to multivariate analysis with R: Practice and theory of data analysis (Ohmsha, 2018).
Shimizu, H. Multilevel modelings for individual and group data (Nakanishiya Shuppan. 2014).
Comtois, D. Package ‘summarytools’: tools to quickly and neatly summarize data. https://cran.r-project.org/web/packages/summarytools/summarytools.pdf (2022).
Van Buuren, S. & Groothuis-Oudshoorn, K. mice: multivariate imputation by chained equations in R. J. Stat. Softw. 45, 1–67 (2011).
Wickham, H., François, R., Henry, L., Müller, K. & Vaughan, D. Package ‘dplyr’: a grammar of data manipulation. https://cran.r-project.org/web/packages/dplyr/index.html (2023).
Revelle, W. Package ‘Psych’: procedures for personality and psychological research. https://cran.r-project.org/web/packages/psych/psych.pdf (2023).
Torchiano, M. Package ‘effsize’: efficient effect size computation. https://cran.r-project.org/web/packages/effsize/effsize.pdf (2020).
Wei, T. & Simko V. Package ‘corrplot’: visualization of a correlation matrix. https://cran.r-project.org/web/packages/corrplot/corrplot.pdf (2021).
Wickham, H. et al. Package ‘ggplot2’: create elegant data visualisations using the grammar of graphics. https://cran.r-project.org/web/packages/ggplot2/ggplot2.pdf (2021).
Cheng, J., Karambelkar, B. & Xie, Y. Package ‘leaflet’: create interactive web maps with the JavaScript ‘Leaflet’ library. https://cran.r-project.org/web/packages/leaflet/leaflet.pdf (2022).
Kuznetsova, A., Brockhoff, P. B. & Christensen, R. H. B. lmerTest package: tests in linear mixed effects models. J. Stat. Softw. 82, 1–26 (2017).
R Development Core Team. R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. https://www.r-project.org (2021).

Supplementary Table S1 is not available with this version.

There is NO Competing Interest.

SuuplFigure1180.pdf
SupplFigure 1

Download PDF

Version 1

posted

You are reading this latest preprint version

Adolescent mental health is affected by neighborhood mental health conditions at the beginning of adolescence

Status:

Version 1

Abstract

Figures

Introduction

Results

Discussion

Conclusions

Methods

Participants and procedures

Measures

Depressive symptoms

Psychological well-being

The Strength and Difficulties questionnaire

Confounding variables

Statistical analyses

References

Supplementary Table

Additional Declarations

Supplementary Files

Status:

Version 1