Depression and its association with self-esteem, family, peer and school factors in a population of 9586 adolescents in southern Taiwan
*Cheng-Fang Yen, MD, Department of Psychiatry, Kaohsiung Medical University, No. 100, Tzyou 1st Rd, Kaohsiung 807, Taiwan. Email: firstname.lastname@example.org
Aims: The aim of the present study was to gain insight into the prevalence of depression and its association with self-esteem, family, peer and school factors in a large-scale representative Taiwanese adolescent population.
Methods: A total of 12 210 adolescent students were recruited into the present study. Subjects with a score >28 on the Center for Epidemiological Studies' Depression Scale were defined as having significant depression; the Rosenberg Self-Esteem Scale, Adolescent Family and Social Life Questionnaire and Family C-APGAR Index were applied to assess subjects' self-esteem, family, peer and school factors. The association between depression and correlates were examined on t-test and χ2 test. The significant factors were further included in logistic regression analysis.
Results: Among 9586 participants (response rate: 86.3%), the prevalence of depression was 12.3%. The risk factors associated with depression in univariate analysis included female gender, older age, residency in urban areas, lower self-esteem, disruptive parental marriage, low family income, family conflict, poorer family function, less satisfaction with peer relationships, less connectedness to school, and poor academic performance. After adjusting the effects of sex, age and location, only subjects with lower self-esteem, higher family conflict, poorer family function, lower rank and decreased satisfaction in their peer group, and less connectedness to school were prone to depression on logistic regression.
Conclusion: The prevalence of depression is high in Taiwanese adolescents, and the multiple factors of family, peer, school and individuals are associated with adolescent depression. The factors identified in the present study may be helpful when designing and implementing preventive intervention programs.
DEPRESSIVE DISORDERS WITH a steadily increased lifetime risk and earlier onset in successive birth cohorts have been documented in several epidemiologic studies.1,2 Because lifetime prevalence of depressive disorders increases with age, adolescence is the first stage of life associated with a significant rise.3 The prevalence of major depressive disorder in children (below 13 years old) was 2.8% in a meta-analytic study, while the estimated rate rose to 5.6% in adolescents aged 13–18 years old.4
Depression may impair adolescents' academic performance, result in school refusal and truancy,5 delinquency and oppositionality,6 promiscuity and teen pregnancy,7 and may increase the risk for substance abuse.8 Moreover, depression is the major risk for suicide, which is the one of the five leading causes of death in adolescents globally.9–11 Depression in adolescence also creates a substantial burden for both school and family.12 These consequences further support the importance of identifying adolescent depression at as early a stage as possible. Identification of the risk factors for depression is crucial for the development and implementation of effective prevention and treatment programs for adolescents.
Adolescent development is the result of multi-systemic interactions,13 and thus the association between depression and individual as well as social factors in adolescents needs further examination. For example, adolescence is the critical period for the development of self-esteem and self-identity,14 and low self-esteem may endanger adolescents' emotional regulation.15 Among the constellation of social contexts in which adolescents are embedded, family, peer and school contexts are the most critical.16
As adolescents develop, they express a clear preference for spending time with peers,17 and the campus, as a microcosm of society, is a major living environment for adolescents. Both interaction with the peer group and subjective feelings toward school life may have an influence on adolescents' mental health. Nevertheless, family characteristics remain the profound influence on adolescent development.
Family aggregation and secular changes in depression suggest the influence of genetic and environmental interaction.1 Whereas genes and other fixed markers have been considered to be unchangeable, socio-environmental factors, with the related risk mechanisms of depression, are suitable targets for intervention.18
Numerous studies have focused on the correlation of depression with social contexts in adolescents.8,19–22 Few of them, however, were community-based studies that investigated the correlation between depression and individual and social factors. Furthermore, most of them have been conducted in Western societies,23–25 and the sample sizes were relatively small. To our knowledge, no study has examined the correlation between depression and individual, family, peer and school factors in a large Asian adolescent population. The purpose of the present study was to gain insight into the prevalence of depression and its association with self-esteem, family (parental marriage and education, family income, conflict and function), peer (rank and feeling in peer groups) and school factors (connectedness and achievement) in a large representative Taiwanese adolescent population. Previous studies have found that female gender,26 older age27,28 and living in urban areas29 increased the risk of depression in adolescents; the present study examined the association between adolescent depression and self-esteem and social contexts by adjusting the effects of sex, age and location.
The current investigation is based on data from the Project for the Health of Adolescents in Southern Taiwan, a research program studying the mental health status of adolescents living in southern Taiwan, an area of three metropolitan cities and four counties. In 2003 there were 257 873 adolescent students in 209 junior high schools and 202 456 adolescent students in 140 senior high schools and vocational schools in this area. Based on the Taiwan Demographic Fact Book,30 12 junior high schools and 19 senior high and vocational schools were randomly selected from urban areas, as well as 11 junior high schools and 10 senior high and vocational schools from non-urban areas. The classes in these schools were further stratified into three levels based on the grade levels in both the junior high schools and senior high and vocational schools, and then 207 classes, containing 12 210 adolescent students were randomly selected based on the ratio of students in different grades.
Research assistants explained the purpose and procedure of the study to the students in class, and encouraged them to participate. Written informed consent was obtained from the adolescents beforehand, and then the participants were invited to complete the research questionnaires anonymously. The protocol was approved by the Institutional Review Board of Kaohsiung Medical University. We also recruited 76 adolescents (40 junior high school students and 36 senior high school students) into a pilot study to examine the 2-week test–retest reliability of research instruments.
Center for Epidemiological Studies' Depression Scale
The 20-item Mandarin-Chinese version31 of the Center for Epidemiological Studies' Depression Scale (MC-CES-D) is a self-administered 4-point evaluation scale assessing frequency of depressive symptoms in the preceding week, with scores ranging from 0 (none or very few) to 3 (always).32 Higher MC-CES-D scores indicate more severe depression. The Cronbach alpha for the MC-CES-D in the present study was 0.930 and 2-week test–retest reliability (r) was 0.782. A previous study using the MC-CES-D in a two-phase survey for depressive disorders among non-referred adolescents in Taiwan found that the adolescents with total MC-CES-D scores >28 were more likely to have major depressive disorder with or without functional impairment.33 In the present study we defined those adolescents whose MC-CES-D score was higher than 28 as having significant depression.
Rosenberg Self-Esteem Scale
The Chinese version of the Rosenberg Self-Esteem Scale (C-RSES) contains 10 4-point items that assess subjects' self-esteem, with good reliability and construct validity.34 A higher C-RSES score indicates a higher level of self-esteem. The Cronbach's alpha in the present study was 0.692 and 2-week test–retest reliability was 0.745.
Adolescent Family and Social Life Questionnaire
Three subscales on the Adolescent Family and Social Life Questionnaire (AFSLQ) were used to evaluate participants' family conflict, rank and subjective perception with respect to the peer group, and connectedness to school,35 with Cronbach's alpha ranging from 0.722 to 0.835 and 2-week test–retest reliability from 0.642 to 0.718. Higher scores on the three AFSLQ subscales indicate a higher level of family conflict, lower rank within and decreased satisfaction with status in peer groups, and less connectedness to school.
Family APGAR Index
The Chinese-version of the Family APGAR Index (C-APGAR),36 which measures satisfaction with aspects of family life, is based on the original version developed by Smilkstein.37 The 5-point response scales reflect frequency, ranging from never to always. High scores indicate good family support. The Cronbach's alpha in the present study was 0.843 and 2-week test–retest reliability was 0.724.
We also collected participants' family data, including parental marriage status, educational levels and family income. The participants whose academic performance in the recent semester ranked in the lower one-third of their class were considered to have poor academic achievement.
Procedure and statistical analysis
The adolescents were asked to anonymously complete all questionnaires based on the explanations of the research assistants and under their direction. All students received a gift worth $NT30.00 at the end of the assessment.
Data analysis was performed using SPSS 12.0 (SPSS, Chicago, IL, USA). The rate of adolescents with depression was calculated using the MC-CES-D cut-off determined in a previous study.33 The association between depression and demographic characteristics, self-esteem, and family, peer and school characteristics were examined using t-test and χ2 test. The significant factors were further selected and included in logistic regression to examine their association with depression in adolescents. P < 0.01 was considered statistically significant.
Of the 12 210 randomly selected adolescents, 11 111 (91.0%) returned their written informed consent. Of them, 9586 (86.3%) completed the research questionnaires without omission, and their demographic characteristics, level of self-esteem, and family, peer and school characteristics are shown in Table 1. According to the MC-CES-D cut-off for Taiwanese adolescents, 12.3% of the participants were currently in a depressive state.
Table 1. Subject characteristics
| Male||4673 (48.7)|| |
| Female||4913 (51.3)|| |
|Age (years)|| ||14.7 ± 1.8|
| Urban||5637 (58.8)|| |
| Rural||3949 (41.2)|| |
| ≥29||1179 (12.3)|| |
| <29||8407 (87.7)|| |
|Self-esteem on the C-RSES|| ||28.3 ± 5.4|
| Disruptive family|
| Intact||8266 (86.2)|| |
| Disruptive||1320 (13.8)|| |
| Father's education level (years)|
| ≤9||3249 (33.9)|| |
| >9||6337 (66.1)|| |
| Mother's education level (years)|
| ≤9||3798 (39.6)|| |
| >9||5788 (60.4)|| |
| Total family income ($NT)|
| ≤30 000||3630 (37.9)|| |
| >30 000||5956 (62.1)|| |
| Family conflict|| ||1.9 ± 0.5|
| Family function on the C-APGAR|| ||13.7 ± 3.8|
|Peer and school characteristics|
| Rank and feeling in peer group|| ||1.8 ± 0.5|
| Connectedness to school|| ||1.8 ± 0.4|
| Academic performance|
| Good or average||2359 (24.6)|| |
| Poor||7227 (75.4)|| |
The associations between depression and demographic characteristics, self-esteem, and family, peer and school characteristics examined on t-test and χ2 test are shown in Table 2. The results indicated that the adolescents who were female, were older, lived in urban areas, had lower self-esteem on the C-RSES, perceived higher family conflict, poor family function on the C-APGAR and lower rank within and decreased satisfaction with status in their peer group, who had less connectedness to school, had poor academic performance, and whose parental marriage was disruptive and family income low (≤NT$30 000) were more likely to have depression.
Table 2. Major depressive disorder and subject characteristics (n, %)
|Gender,|| || ||22.797***†|
| Male||498 (10.7)||4175 (89.3)|| |
| Female||681 (13.9)||4232 (86.1)|| |
|Age (years), mean ± SD||15.0 ± 1.7||14.6 ± 1.8||−6.094***‡|
|Location|| || ||19.392***†|
| Urban||763 (13.5)||4874 (86.5)|| |
| Rural||416 (10.5)||3533 (89.5)|| |
|Self-esteem on the C-RSES (mean ± SD)||22.8 ± 5.2||29.0 ± 4.9||40.907***‡|
|Parental marriage|| || ||26.140***†|
| Intact||960 (11.6)||7306 (88.4)|| |
| Disruptive||219 (16.6)||1101 (83.4)|| |
|Father's education (years)|| || ||0.029†|
| ≤9||397 (12.2)||2852 (87.8)|| |
| >9||782 (12.3)||5555 (87.7)|| |
|Mother's education (years)|| || ||0.925†|
| ≤9||452 (11.9)||3346 (88.1)|| |
| >9||727 (12.6)||5061 (87.4)|| |
|Total family income ($NT)|| || ||8.525**†|
| ≤30 000||492 (13.6)||3138 (86.4)|| |
| >30 000||687 (11.5)||5269 (88.5)|| |
|Family conflict (mean ± SD)||2.3 ± 0.6||1.8 ± 0.5||−27.835***‡|
|Family function on the C-APGAR (mean ± SD)||11.3 ± 3.8||14.0 ± 3.6||23.452***‡|
|Rank and feeling in peer group (mean ± SD)||2.1 ± 0.6||1.8 ± 0.5||−24.374***‡|
|Connectedness to school (mean ± SD)||2.0 ± 0.4||1.8 ± 0.4||−22.192***‡|
|Academic performance|| || ||37.540***†|
| Good or average||804 (11.1)||6423 (88.9)|| |
| Poor||375 (15.9)||1984 (84.1)|| |
The significant factors were further selected and included in logistic regression (Table 3). The results indicated that the adolescents who had lower self-esteem on the C-RSES, perceived higher family conflict, poor family function on the C-APGAR and lower rank and decreased satisfaction in their peer group, and less connectedness to school were more likely to have depression, after adjusting the effects of sex, age and location (−2Log likelihood = 5131.353, d.f. = 11, P < 0.001). Parental marriage status, family income or academic performance were not associated with depression on logistic regression.
Table 3. Correlates of depression in adolescents: logistic regression model
|Living in urban area||13.790***||1.326||1.143–1.541|
|Disruptive marriage of parents||1.139||1.112||0.915–1.353|
|Low family income||2.465||0.887||0.764–1.030|
|High family conflict||109.649***||2.155||1.867–2.489|
|Low family function||25.997***||1.058||1.036–1.082|
|Lower rank and decreased satisfaction in their peer group||73.158***||1.878||1.625–2.169|
|Less connectedness to school||48.524***||1.940||1.610–2.338|
|Poor academic performance||2.540||0.878||0.747–1.030|
Prevalence of depression in adolescents
Previous studies have found that the prevalence of major depressive disorder in adolescents ranges from 0.4% to 8.3%.38 In Taiwan, epidemiological studies have found that 0.5–4.4% of adolescents aged 13–15 had major depressive disorder.33,39 In the present study the rate of adolescents who had significant depression (12.3%) was higher than that seen in previous studies. The difference may be accounted for by at least two factors. First, we recruited older adolescents. Older age has been found to be one of the risk factors for depression in adolescents.27,28 Second, the definition of depression in the present study was based on the cut-off point of the self-reported questionnaire and did not take functional impairment into account. A previous study applying the self-reported Chinese version of Beck Depression Inventory also found that a high proportion (18.9%) of high school students in Hong Kong had moderate to severe depression.40
Previous studies have found that low self-esteem in patients with an initial depressive episode may prolong the index episode41 and was related to the chronicity of major depressive disorder.42 Low self-esteem may persist, even though the mood disorder remits.43 This study further supported the notion that adolescents with depression have low self-esteem. Further prospective studies are necessary to investigate the causal relationship between self-esteem and depression in adolescents.
Parental marriage, family conflict, function and income
A previous study found that parental marital disruption was one of the significant life events that may precipitate the onset of depression in school-aged children44,45 or predispose them to onset in adulthood.46,47 A longitudinal prospective study also found that family disruption in childhood was independently associated with a twofold lifetime risk of adult depression.19 The effect of family disruption, however, was amplified for those subjects reporting high levels of parental conflict.19 In the present study, the parents' disruptive marriage was associated with depression on univariate analysis only, and not on multivariate analysis, but high family conflict and poor family function were still associated with adolescent depression.
The adolescent must deal with developmental tasks and struggle for their autonomy and privacy, therefore, the relationship with the family must be renegotiated, which leads to an increasing conflict within the parent–child relationship.48 Although many adolescents prefer to associate with peers rather than with family, this does not mean that the family climate will no longer influence the adolescent. Poor family cohesion (family support and connectedness) and high levels of parent–adolescent conflict have been shown to be related to mental health, onset of depression and even suicidal ideation in adolescents.25,49–52 We inferred that the high levels of family conflict and impaired family functioning accompanied by parents' marital discordance have important roles in adolescent depression.
The role of low household income or low socioeconomic status in adolescent depression has been debated in previous studies. While some studies failed to find a significant association between low household income and adolescent depression,23,40 other studies found that perceived family financial difficulty was related to adolescent depression and maladaptive behaviors53 and the onset of depression in adulthood.19 In the present study, low family income was associated with adolescent depression only on univariate analysis, and not in the multivariate analysis model when family conflict and family function were taken into consideration. A previous longitudinal study found that increased family conflict was significantly related to depression in the offspring in low-income families in urban areas.54 Thus, we inferred that low household income may have an indirect rather than direct effect on adolescent depression. Parents experiencing financial hardship may exhibit negative parenting and a poor parent–child relationship,55 which increases the adolescents' vulnerability to depression. Meanwhile, low household income may increase the risk of intimate partner violence and parental depression, which further increase the possibility of child abuse and harsh discipline,56,57 and increase the risk of onset of child and adolescent depression.58,59
Good peer relationships have been proven to not only buffer the adolescent from the overwhelming stress of the developmental task and achievement issues in reaching adjustment,60,61 but also mitigate negative cognition and improve depression.62 Affiliation with a peer group was a protective factor against depression, while peer victimization and poor-quality friendships may induce internalized distress.22 In the present study, lower rank and decreased satisfaction in their peer group were associated with current depression in adolescents, which was consistent with the findings of previous investigations.24,63,64 This result further supported the notion that the effect of peer characteristics on adolescent depression was independent of family characteristics.
School affinity and academic achievement
In the present study, low school affinity was another significant correlate of current depression in adolescents. School refusal and truancy were found to be associated with psychiatric comorbidity, and one of these was depression.5 Such school absenteeism was the extreme manifestation of looseness of school connectedness. School affinity was related to relationship with classmates and teachers, attendance at school and a preference for school life. Likewise, peer companionship and teacher support were also important for adolescents to adjust to life stress, and could alleviate depressed mood and increase adolescent self esteem.27,65 Nonetheless, the emotional overinvolvement of the teacher may elicit adverse effects.21
Previous studies have found that poor academic achievement might induce adolescents to have negative feelings toward themselves, which could precipitate depression.24 In contrast, depression might decrease the adolescent's ability and motivation to learn, which results in poor academic expectations on the part of the adolescent.24 No association, however, between adolescent depression and poor academic achievement was found in the present study. The result may be partially accounted for by the fact that adolescents' subjective satisfaction with their academic achievement might be more meaningful than absolute scores or rank when considering the influence of depression.40 Another possible explanation is that academic achievement is not the only criterion that might influence adolescents' self-evaluation of their own competence. Success in one arena may buffer the frustration of failure in another.66
Some limitations of this study should be kept in mind. First, the present study utilized the self-reported inventory, and the functional impairment caused by depression was not taken into consideration. Second, the cross-sectional nature of the present study limited our ability to draw meaningful conclusions about the causal relationship between adolescent depression and its correlates. Third, the validity of the C-APGAR index has not been systemically examined in adolescent population.
In summary, this large-scale representative study found that the prevalence of depression was high in Taiwanese adolescents, and that multiple factors related to family, peer, school and individuals were associated with adolescent depression. The factors identified in the present study may be helpful when designing and implementing preventive intervention programs. Although it is difficult to change adolescents' family and school characteristics, early monitoring of the possibility of depression should be commenced for adolescents with adverse family and school characteristics. Individual or group intervention should be implemented for adolescents with low self-esteem or poor interaction with peer groups.
This study was supported by grant NSC 93-2413-H- 037-005-SSS awarded by the National Science Council, Taiwan.