Journal of Adolescent Health
Volume 42, Issue 6 , Pages 605-616, June 2008

Protective and Vulnerability Factors Predicting New-Onset Depressive Episode in a Representative of U.S. Adolescents

  • Benjamin W. Van Voorhees, M.D., M.P.H.

      Affiliations

    • Section of General Internal Medicine, Department of Medicine, The University of Chicago, Chicago, Illinois
    • Section on Community Pediatrics, Department of Pediatrics, University of Chicago
    • Section on Child and Adolescent Psychiatry, Department of Psychiatry, University of Chicago
    • Corresponding Author InformationAddress correspondence to: Benjamin W. Van Voorhees, M.D., M.P.H., The University of Chicago, 5841 S. Maryland, MC 2007, Chicago, IL 60637.
    • ,
  • David Paunesku, B.A.

      Affiliations

    • Section of General Internal Medicine, Department of Medicine, The University of Chicago, Chicago, Illinois
    • ,
  • Sachiko A. Kuwabara, M.A.

      Affiliations

    • Department of Psychology, Illinois Institute of Technology, Chicago, Illinois
    • ,
  • Anirban Basu, Ph.D.

      Affiliations

    • Section of General Internal Medicine, Department of Medicine, The University of Chicago, Chicago, Illinois
    • ,
  • Jackie Gollan, Ph.D.

      Affiliations

    • Department of Psychiatry, Feinberg School of Medicine, Northwestern University, Chicago, Illinois
    • ,
  • Benjamin L. Hankin, Ph.D.

      Affiliations

    • Department of Psychology, University of South Carolina, Columbia, South Carolina
    • ,
  • Stephanie Melkonian, B.A.

      Affiliations

    • Section of General Internal Medicine, Department of Medicine, The University of Chicago, Chicago, Illinois
    • ,
  • Mark Reinecke, Ph.D.

      Affiliations

    • Department of Psychiatry, Feinberg School of Medicine, Northwestern University, Chicago, Illinois

Received 21 June 2007; accepted 14 November 2007. published online 20 February 2008.

Article Outline

Abstract 

Purpose

Depressive episodes cause considerable morbidity and mortality in adolescents. We sought to identify factors predicting new onset depressive episode in a representative sample of U.S. adolescents.

Methods

We conducted logistic regression analyses to identify baseline individual, family, school/peer and community factors predicting new-onset depressive episode at a 1-year follow-up in a longitudinal cohort study of 4791 U.S. adolescents. Potential protective and vulnerability factors included individual (sociodemographics, general health and maturity, coping behavior, self-concept, and affect regulation), family (connectedness and conflict), school/peers (acceptance and performance), and community (engagement, delinquency, and adverse events).

Results

African American and Hispanic ethnicity, female gender, and low-income status predicted higher risk of onset of a depressive episode. Active coping and positive self-concept, predicted lower risk, whereas poor affect regulation and greater depressed mood predicted higher risk. Family “connectedness,” parental warmth, peer acceptance, better school performance, and religious activities were protective, whereas parental conflict, delinquent activities, and greater numbers of adverse events increased risk of depressive episodes.

Conclusions

Female gender, nonwhite ethnicity, low-income status, poor health, and parental conflict, increase risk of a depressive episode. Physicians should consider recommending behaviors that enhance perceived fitness, favorable self-concept, family connectedness, peer acceptance, and community engagement to youth as means a of mitigating this risk for developing a depressive episode.

Keywords: Adolescent, Depressive disorder, Prevention

 

Depressive episodes affect 25% of adolescents by age 24, with a significant adverse impact on development and attendant social morbidity [1]. Pediatricians and family physicians are increasingly being called upon to screen for and intervene in adolescent depressive episodes [2]. Psychiatry and primary care medicine have increasingly focused on treating depression in individuals, and have deemphasized social context as a source of vulnerability and protection [3]. Knowledge of potential protective and vulnerability factors and behaviors would be helpful for clinicians addressing the needs of adolescents who may be at risk for developing a depressive episode and for the further refinement of preventive interventions. We chose the broad definition of protection as “attributes with direct ameliorative effects.” Conversely vulnerability was defined as attributes associated with “greater manifestation of maladjustment” [4].

We chose to integrate the models of developmental psychopathology and sociological ecology in understanding vulnerability and protection in depression in adolescents (Figure 1). From the perspective of developmental psychopathology, proximal cognitive, and interpersonal vulnerability factors interact with adverse events to precipitate depressed mood. Vulnerability factors include cognitive/behavioral (e.g., negative interpretations and expectations, and poor coping skills), interpersonal (social skills/problem solving deficits, lack of social support), and parental/ family factors (discord, high levels of criticism, lack of support, and insecure attachment) [5], [6]. In a broader sociological framework, individual adolescents live within family, school, and community contexts [7]. We believe that adolescent behaviors and perceptions of their own individual (internal) experiences and those that pertain to their experiences within the family, school, and community framework are critical to understanding the onset of depression.

Previous studies—primarily of vulnerability—have relied on clinical samples, cross-sectional study designs, or ethnically homogenous U.S. or non-U.S. samples. To our knowledge, there is only one study examining ecological determinants of the onset of depression in young adults [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19]. Clarification of relations between individual, family, school, and community factors, with a particular focus on those that may be protective, will broaden our understanding of the disorder's onset. Confirmation of such factors in a representative sample of U.S. adolescents will be useful in planning preventive interventions and for physicians seeking to better understand onset of depressive episodes.

To address this need, we conducted a secondary data analysis of the public-use data set of the National Longitudinal Study of Adolescent Health (N = 6504). We evaluated the relationship between potential baseline protective and vulnerability factors and the likelihood of depression at a 1-year follow-up. Potential protective and vulnerability factors included individual (sociodemographics, general health and maturity, coping behavior, self-concept, and affect regulation), family (connectedness and conflict), school/peers (connection and performance), community (involvement, delinquency, and adverse events) [5]. We hypothesized that protection and vulnerability are derived from multiple spheres of adolescent experience.

Back to Article Outline

Methods 

Survey design and data collection 

The National Longitudinal Study of Adolescent Health (AddHealth) is a representative sample of U.S. adolescents in grades 7–12 with an additional oversampling of African American youth from high educational backgrounds. The data used in this study was collected over a 1-year period, from 1995–1996 [20]. Surveys were conducted face to face, in home (Waves 1 and 2), and in school (Wave 1 only, not completed by all participants). It also included a parent survey. Each school that participated in the study was systematically chosen to represent urban, regional, and ethnic strata. After being stratified by grade and gender within each school, 17 adolescents were chosen randomly from each age–gender group (17/strata, 200/schools). The study had an overall response rate of 76.8%, with a total sample of N = 6504 (public-use data set, Sociometrics, Inc.) [21]. Nearly three quarters (73.5%) of the Wave 1 participants completed the Wave 2 survey (N = 4791). This survey's methods have been described in previous work [20].

Outcome variable 

We constructed the outcome variable for depressive episodes using depression-specific items from the Center for Epidemiologic Studies Depression (CES-D) scale (20 items, score 0–60, Chronbach's alpha = 0.88) at Wave 2. Adolescents who met the following criteria: (1) reported experiencing at least one core symptom (depressed mood or anhedonia) “most all of the time (5–7 days) for the last week” and (2) endorsed a similar level of severity for at least three other Diagnostic and Statistical Manual (DSM) III criteria symptoms were classified as having a depressive episode. We chose a cutoff of four symptoms based on the literature that suggests those with four or more symptoms may warrant treatment [22]. We excluded those with a depressive episode at baseline (N = 100) in order to focus on new-onset cases.

We chose this method for constructing our depressive episode outcome variable to avoid the problem of low specificity associated with standard CES-D cut offs for depressive episodes (sensitivity ranges of 86%–100% and specificity ranges of 53%–84%) [23]. Cutoff specificity may be even lower in adolescents [24]. Because standard cutoffs for CES-D scores have low specificities for detection of depressive episodes in adolescents, Schoenbach [25], [26] developed this method to derive DSM criteria based on the specific CES-D items relating to the diagnostic criteria (e.g., not including those related to anxiety or general distress). Radloff [27] then validated this approach in junior and senior high, college, and young adult populations. Radoloff reported that Schoenbach's method identified 76.35% of individuals identified as meeting criteria for having major depression. Additionally, we examined the validity of the outcome variable in this dataset by comparing it to results obtained in other studies. A sensitivity analysis was completed to determine the effect of our method of constructing the outcome variable on our results.

Independent variables 

All 5800 variables that had been incorporated into the ADHEALTH dataset were screened, and 128 potential protective and vulnerability factors were identified by two investigators based on models/reviews [5], [6]. Protective and vulnerability factors include variables related to behavior and affect regulation (active problem solving and positive self-esteem and expectancies), interpersonal relationships (social support from peers and family), community activity, and the absence of vulnerability factors such as delinquency and adverse events, socioeconomic characteristics, and general health [5], [6]. The items used n ADHEALTH survey were derived from multiple sources to address a mandate from the Congress to assess broad aspects of health in adolescents. No complete scales were used in the survey, and many items were changed in the course of field testing [28].

Individual factors 

Sociodemographic characteristics and general health 

Socioeconomic and demographic variables included gender, ethnicity, household income, age, and parental education. Household income was reported in quintiles for ease of reporting but was entered into the adjusted model as a continuous variable. Parents' education level was assessed for both mother and father; we distinguished between those who had no high school degree, those who graduated from high school, and parents who had earned a college degree. In addition, variables related to general health were identified and included items related to physical health and biological maturity (e.g., Tanner stage).

Self-concept, coping, expectancies and affect regulation 

In terms of self-concept, adolescents were asked to rate their level of agreement with statements such as, “I do everything just right,” “I have lots of good qualities,” and “I have a lot to be proud of.” Participants were also asked questions about problem-solving orientation (e.g. “I avoid confronting problems”) and problem-solving strategies (e.g. “I think of as many ways as possible to find a solution,” “Get as many facts as possible at first,” “Use a systematic method for comparing alternatives,” and “After carrying out a solution, [I] usually try to analyze what went right and what went wrong”). With regard to expectancies, adolescents were asked to rate the likelihood that they would be “middle class by 30” or “alive at age 35.” Affect regulation was assessed by queries on irritability, anxiety, and mood lability. For irritability, participants' parents were asked if the adolescent had a “bad temper” (yes/no response). Anxiety was assessed with the question, “Over the last twelve months, have you had trouble relaxing?” For mood lability, adolescents were asked, “Over the last twelve months have you been moody?” and, “How often have you had crying spells in the last month?” The total CES-D score was used as a measure of depressed mood.

Family 

Multiple aspects of family life were examined. Adolescents rated the condition of their home and whether they performed chores. Parents reported whether they were married, the quality of their relationship, and whether they were involved in school organizations, such as at the PTA. Adolescents reported the degree which they felt “paid attention to,” “understood,” and that the family “had fun together.” Adolescents reported the extent to which they felt close to each parent or that they were “warm” and whether parents fought or “talked about splitting up.” Family relationship variables addressed family connectedness, support, and the strength of their parents' relationships with one another (e.g., if the parents talked about splitting up or if they fight).

School and peers 

School and peer environment were assessed with questions that gauged the degree to which participants felt accepted, active, successful, and comfortable in their schools. They were asked if they felt “socially accepted,” “close to people at this school,” “safe at school,” “happy at this school,” and “like I am a part of this school.” Participants were also asked if they felt their classmates were “prejudiced,” and whether or not they think their teachers are fair. In terms of school performance, they were asked what their most recent grades were in several subject areas and whether they had had a failing grade. With regard to extracurricular activities, adolescents were queried with regard to exercise, amount of television watched, frequency of unstructured “hanging out,” whether they had a license, and whether they participated in sports or other activities.

Community engagement and delinquency 

Community and school environment and engagement were assessed across a range of behaviors, including neighbor relations, religious activities, delinquency, and adverse events. Neighborhood variables included cleanliness and the degree to which the participant interacted with his or her neighbors. To assess religious activity, participants were asked about youth group participation, frequency of church attendance, frequency of prayer, and the importance of religion to him or her. Delinquency and adverse events were assessed in the areas of criminal activity, substance abuse (e.g., marijuana, cocaine and inhalants), and sexual activity. We constructed variables corresponding to the total number of both violent and nonviolent criminal acts committed. Substance abuse was assessed by asking about frequency of use. Sexual activity was assessed by asking adolescents if they had ever had sex and if they ever had ever been in a relationship. In addition, we evaluated the occurrence of adverse events such as emotional or physical trauma; for example: a family member committing suicide, physical injuries, breaking up with significant others, school suspensions, fights, stabbings, and so forth.

Analysis 

Using logistic regression, we evaluated the relations between baseline independent variables at Wave 1 and the outcome variable at Wave 2. Participants who manifested a depressive episode at Wave 1 were excluded. To ensure an adequate number of observations in the reference group, we combined reference level indicator variables with <100 observations with the nearest neighbor indicator variable. We report results for bivariate and multivariate models for each variable. The number of adolescents with the outcome of depressive episode was not sufficient to support a multivariate model incorporating all 120 variables simultaneously. In the first multivariate model, we controlled for sociodemographic variables (the model included the variable of interest and gender, age, ethnicity, and household income). In the second multivariate model, we controlled for baseline depressed mood (model included variable of interest and CES-D score). We conducted goodness-of-fit tests and sensitivity analyses. To address the potential effects of subject attrition, we compared those who completed the Wave 2 survey with those who did not. All results were adjusted for complex survey design, and sampling weights were employed to reflect a probability sample of U.S. adolescents. We used the STATA SE version 9.0 for this analysis.

Back to Article Outline

Results 

Sample and nonresponder characteristics and validation of the outcome variable 

The characteristics of this r sample of U.S. adolescents have been reported previously [29]. Those who did not respond to the Wave 2 survey were more likely to be African American, lower income, and had lower parental education. However, they were not more likely to have a depressive episode at Wave 1 or to have greater levels of depressed mood. The prevelance of a depressive episode was 2.49% at the Wave 1 baseline and 2.8% at the Wave 2 follow-up. This is similar to the prevelance obtained in Schoenbach's [26] original study, which compares the construction of a major depression variable from the CES-D to actual structured psychiatric interviews (2.9%). The prevalence of a depressive episode increased with age from 0% at age 11 to 1.56% at age 13 to 3.3% at age 18. The mean CES-D score for those with a depressive episode was 34 for Wave 1 and 33 for Wave 2 (vs. 34 in clinically depressed populations) [27]. With regard to the outcome variable used in the sensitivity analysis, those whose CES-D scores were greater than 28 represented 4.1% of the individual surveyed at Wave 2 (N = 201).

Individual 

Sociodemographic characteristics and general health 

Several individual characteristics and health experiences predicted future risk of a depressive episode (Table 1). Female gender, age greater than 14, and African American and Hispanic American ethnicity (compared with whites) were all associated with twofold increase in likelihood of developing depression in bivariate analyses. Low socioeconomic status, lowest household income quartile (significant in all models), and neither parent graduating from high school predicted three- and twofold higher risk of the onset of a depressive episode, respectively. Residing in an urban, suburban, commercial, or rural area did not significantly impact the risk for depression. Having “enough sleep” (all models except adjustment of depressed mood) and “feeling fit” were associated with three- and fivefold lower likelihood of developing a depressive episode, while “having a lot of energy” was not. Reporting headaches every day predicted a sixfold increase in risk of a depressive episode that was only slightly attenuated after adjustment.

Table 1. Sociodemographic factors and general health
Odds ratio (95% confidence interval)
UnivariateDemographic adjustmentaCES-D adjustmentb
Gender
Male11d1
Female2.11(1.32–3.39)c1.92(1.14–3.23)c1.71(1.04–2.8)c
Age
Under 14 years old at time 111d1
Over 14 years old at time 12.52(1.19–5.36)c2.6(1.02–6.64)2.02(0.9–4.53)
Ethnicity
White11d1
Hispanic1.88(1.02–3.49)c1.75(0.74–4.15)1.56(0.82–2.98)
Black1.97(1.16–3.35)c1.81(0.95–3.45)1.62(0.95–2.77)
Asian1.01(0.30–3.38)1.40(0.29–6.84)0.8(0.28–2.63)
Other1.25(0.43–3.65)0.78(0.18–3.46)0.86(1.1–1.16)d
Household income ($k)
66–90011d1
46–651.29(0.5–3.37)1.3(0.5–3.41)1.19(0.46–3.1)
32–451.3(0.49–3.41)1.15(0.42–3.12)1.08(0.4–2.93)
19–311.47(0.58–3.75)1.35(0.53–3.41)1.13(0.44–2.9)
0–183.43(1.61–7.29)c3.02(1.36–6.7)c2.43(1.11–5.31)c
Parents finished high school
One or more parents111
Neither parent2(1.15–3.47)c1.6(0.83–3.07)1.47(0.81–2.68)
Do you usually get enough sleep?
No111
Yes0.37(0.25–0.55)c0.37(0.23–0.62)c0.64(0.41–1.01)
“I am physically fit”
Strongly disagree111
Disagree0.22(0.06–0.88)c0.24(0.05–1.05)0.28(0.07–1.17)
Neither agree nor disagree0.14(0.04–0.45)c0.15(0.04–0.58)c0.25(0.08–0.78)c
agree0.11(0.03–0.35)c0.17(0.05–0.59)c0.27(0.08–0.88)c
Strongly agree0.12(0.03–0.39)c0.19(0.05–0.69)c0.36(0.1–1.31)
In the last 12 months, how often have you had a headache?
Never111
Just a few times2.12(0.58–7.77)2.67(0.53–13.42)2.18(0.56–8.55)
About once a week3.83(1.06–13.85)c4.08(0.76–21.88)3.05(0.8–11.57)
Almost every day4.66(1.25–17.43)c6.58(1.15–37.48)c3.41(0.85–13.72)
Every day6.6(2.15–20.23)c4.81(1.76–13.13)c4.02(1.13–14.28)c
You have a lot of energy.
Strongly disagree111
Disagree1.77(0.3–10.43)1.27(0.19–8.37)1.54(0.27–8.74)
Neither agree nor disagree0.27(0.04–1.78)0.22(0.03–1.56)0.32(0.05–2.06)
agree0.4(0.07–2.34)0.29(0.05–1.8)0.66(0.11–3.86)
Strongly agree0.43(0.08–2.43)0.38(0.07–2.05)0.82(0.14–4.71)

aAdjusted for race, income, age and gender.

bAdjusted for baseline CES–D score.

cIndicated p < .05.

dNot adjusted for covarying items; see Methods.

Coping, self-concept, expectancies, and affect fegulation 

Active coping or problem solving and favorable self-concept were protective against new onset of a depressive episode, whereas poor affect regulation and higher levels of depressed mood increased risk (Table 2). Not strongly agreeing that “you go out of your way having to deal with problems” and “going with your gut without thinking too much” predicted a twofold reduction in risk of depression across all the models. Favorable self-concept (“doing everything just right, “like myself,” and “have a lot to be proud of”) predicted lower likelihood of developing a depressive episode in all models except after adjustment for baseline depressed mood (similar trend, but not significant). Having “trouble relaxing” (twofold), a “bad temper” (sixfold), more frequent “blue spells” (13-fold), “moodiness” (eightfold), and higher quartile of depressed mood (11-fold) predicted greater incidence of a depressive episode at the 1-year follow-up in all models, with substantial attenuation after adjustment for current depressed mood. Future expectations, levels of television watching, exercise, and participation in hobbies and school activities did not affect risk of depression.

Table 2. Coping, self-concept and affect regulation
Odds ratio (95% confidence interval)
UnivariateDemographic adjustmentaCES-D adjustmentb
Coping and self–concept
You usually go out of your way to avoid having to deal with problems in your life.
Strongly agree111
Agree0.49914380.51027530.556031
Neither agree nor disagree0.4108862c0.3755071c0.4646573
disagree0.293801d0.2052794d0.3729087c
Strongly disagree0.45216820.48449150.543535
When making decisions, you usually go with your “gut feeling” without thinking too much about the consequences of each alternative.
Strongly agree111
Agree0.54(0.3–0.97)c0.64(0.32–1.29)0.66(0.34–1.26)
Neither agree nor disagree0.27(0.12–0.62)c0.34(0.12–0.96)c0.36(0.15–0.85)c
Disagree0.5(0.27–0.91)c0.55(0.28–1.11)0.77(0.38–1.53)
Strongly disagree0.58(0.22–1.56)0.53(0.14–2.08)0.82(0.29–2.34)
I feel like I am doing everything just right.
Strongly disagree111
Disagree0.28(0.12–0.66)c0.23(0.08–0.61)c0.77(0.3–2)
Neither agree nor disagree0.11(0.04–0.31)c0.12(0.04–0.37)c0.36(0.11–1.18)
Agree0.23(0.08–0.67)c0.23(0.07–0.78)c0.71(0.21–2.43)
Strongly agree0.24(0.09–0.61)c0.22(0.07–0.64)c0.43(0.15–1.23)
I like myself just the way I am.
Strongly disagree111
Disagree0.43(0.13–1.47)0.28(0.07–1.07)1.12(0.27–4.67)
Neither agree nor disagree0.22(0.06–0.76)c0.13(0.03–0.54)c0.79(0.2–3.13)
Agree0.11(0.03–0.36)c0.08(0.02–0.27)c0.59(0.15–2.29)
Strongly agree0.12(0.03–0.41)c0.1(0.02–0.37)c0.78(0.19–3.16)
I have a lot to be proud of.
Strongly disagree111
Disagree0.49(0.04–5.5)0.23(0.02–2.65)0.33(0.03–3.29)
Neither agree nor disagree0.22(0.02–2.21)0.08(0.01–0.87)c0.24(0.03–2.14)
Agree0.11(0.01–1.05)0.06(0.01–0.52)c0.24(0.03–1.98)
Strongly agree0.09(0.01–0.87)c0.04(0–0.44)c0.25(0.03–2.24)
Affect Regulation
In the past year how often: have you had trouble relaxing?
Never111
Just a few times1.88(1.15–3.09)c2.07(1.16–3.69)c1.4(0.83–2.35)
About once a week2.43(1.34–4.42)c2.9(1.51–5.59)c1.23(0.63–2.4)
Almost every day4.81(2.09–11.04)c3.83(1.31–11.25)c1.79(0.7–4.57)
Every day6.08(2.31–16)c6.73(2.42–18.68)c1.92(0.6–6.1)
[asked parent] Does [adolescent] have a bad temper?
No111
Yes2.21(1.37–3.58)c2.01(1.16–3.5)c1.73(1.07–2.79)c
In the last month, how often: did you feel depressed or blue?
Never111
Just a few times3.49(1.3–9.4)c3.21(1.07–9.61)c2.71(0.97–7.59)
About once a week6.47(2.6–16.08)c6.6(2.24–19.46)c3.56(1.34–9.44)c
Almost every day8.65(3.19–23.5)c7.78(2.61–23.18)c3.82(1.31–11.11)c
Every day13.56(4.21–43.69)c9.91(2.48–39.56)c4.45(1.22–16.21)c
Please tell me how often you have had each of the following conditions in the past 12 months:
Moodiness.
Never111
Just a few times0.86(0.41–1.8)1.08(0.46–2.52)0.85(0.41–1.75)
About once a week1.45(0.69–3.04)2(0.85–4.74)0.95(0.44–2.03)
Almost every day3.15(1.38–7.16)c4.35(1.71–11.08)c1.51(0.63–3.63)
Every day7.81(3.89–15.66)c9.6(3.54–26.06)c3.31(1.56–7.05)c
CESD quartile at wave 1 (CESD score)
1st quartile (up to 5)11
2nd quartile (5–9)1.31(0.43–4.04)6.02(1.34–26.98)c
3rd quartile (10–15)6.14(2.5–15.05)c22.15(4.83–101.48)c
4th quartile (16+)11.01(4.46–27.2)c34.12(7.61–152.86)c

aDemographic adjustment included race, income, age and gender.

bAdjusted for mood at baseline. Mood was measured by CES–D score.

cIndicates p < .05.

dNot adjusted for covarying items; see Methods.

Family 

Factors related to family function also were associated with both protection and vulnerability for the onset of a depressive episode (Table 3). Feeling that your family “had fun together,” “pays attention to you,” “understands you,” and makes you feel “loved and wanted,” predicted a fivefold reduction in future likelihood of developing a depressive episode in all models (except “fun” and “loved and wanted” after mood adjustment). Parents report of not “fighting with your partner” (twofold reduction), “talking about splitting” (fourfold increase), and adolescent report of “wanting to leave home” (fourfold increase) predicted risk of a depressive episode in all models (except leaving home after mood adjustment). Mother and father's “warmth' and “closeness to residential father” demonstrated a four- to fivefold reduction in risk of subsequent depressive episodes in all models except for that which included depressed mood. Better physical condition of the home, parents being married, closeness to mother, doing chores around the house, and parental involvement in school organizations were not protective.

Table 3. Family
Odds ratio (95% confidence interval)
UnivariateDemographic adjustmentaCES-D adjustmentb
Family function
How much do you feel that you and your family have fun together?
Not at all111
A little0.92(0.34–2.52)0.7(0.2–2.42)1.1(0.37–3.24)
Some0.41(0.17–0.97)c0.37(0.14–0.99)0.67(0.26–1.73)
Quite a bit0.21(0.09–0.5)c0.25(0.09–0.68)c0.5(0.2–1.26)
Very much0.26(0.1–0.69)c0.28(0.09–0.83)c0.64(0.22–1.89)
How much do you feel that your family pays attention to you?
Not at all111
A little0.89(0.27–2.96)0.5(0.13–1.94)0.61(0.18–2.09)
Some0.4(0.12–1.31)0.28(0.09–0.95)c0.42(0.13–1.37)
Quite a bit0.18(0.05–0.57)c0.15(0.04–0.48)c0.29(0.09–0.92)c
Very much0.19(0.06–0.62)c0.13(0.04–0.45)c0.35(0.1–1.16)
How much do you feel that people in your family understand you?
Not at all111
A little0.47(0.2–1.13)0.5(0.18–1.36)0.55(0.22–1.38)
Some0.24(0.11–0.53)c0.33(0.14–0.8)c0.43(0.18–0.99)
Quite a bit0.09(0.04–0.19)c0.09(0.03–0.26)c0.22(0.1–0.47)c
Very much0.19(0.08–0.47)c0.25(0.09–0.7)c0.48(0.18–1.26)
You feel loved and wanted.
Strongly disagree111
Disagree0.31(0.08–1.22)0.32(0.07–1.52)0.78(0.18–3.41)
Neither agree nor disagree0.24(0.07–0.8)c0.27(0.07–1.09)0.66(0.18–2.41)
Agree0.2(0.07–0.54)c0.22(0.07–0.73)c0.77(0.25–2.42)
Strongly agree0.18(0.07–0.45)c0.2(0.06–0.61)c0.85(0.28–2.56)
[asked parent] How much do you fight or argue with your current (spouse/partner)?
A lot111
Some111
A little0.47(0.27–0.8)c0.52(0.28–0.97)c0.51(0.3–0.86)c
Not at all0.57(0.27–1.2)0.56(0.27–1.17)0.63(0.3–1.29)
[asked parent] In the past year, have you and your current (spouse/ partner) talked to each other about separating?
No111
Yes3.76(2.16–6.55)c3.15(1.64–6.07)c3.33(1.9–5.84)c
How much do you feel that you want to leave home?
Not at all111
A little1.13(0.62–2.08)1.07(0.54–2.11)0.83(0.45–1.53)
Some1.5(0.77–2.91)1.64(0.74–3.63)0.96(0.47–1.94)
Quite a bit3.24(1.66–6.32)c2.63(1.21–5.7)c1.57(0.72–3.39)
Very much4.47(2.14–9.33)c4.41(1.95–9.94)c1.72(0.75–3.92)
Family relationships
Most of the time, your father is warm and loving toward you.
Strongly disagree111
Disagree0.75(0.19–2.85)0.91(0.17–4.81)0.67(0.17–2.67)
Neither agree nor disagree0.29(0.1–0.85)c0.3(0.07–1.2)0.41(0.14–1.17)
Agree0.16(0.05–0.5)c0.2(0.05–0.83)c0.28(0.09–0.84)c
Strongly agree0.13(0.05–0.39)c0.17(0.05–0.65)c0.28(0.1–0.81)c
How close do you feel to your [residential father]?
Not at all111
A little0.26(0.08–0.9)c0.36(0.08–1.72)0.37(0.1–1.34)
Some0.61(0.17–2.2)0.63(0.11–3.71)0.72(0.19–2.82)
Quite a bit0.19(0.06–0.56)c0.25(0.06–1.08)0.38(0.12–1.2)
Very much0.16(0.06–0.44)c0.23(0.06–0.95)c0.36(0.12–1.11)
Most of the time, your mother is warm and loving toward you.
Strongly disagree111
Disagree1.99(0.38–10.45)0.69(0.11–4.23)1.58(0.29–8.49)
Neither agree nor disagree0.79(0.15–4.28)0.45(0.07–2.75)0.78(0.14–4.24)
Agree0.43(0.1–1.9)0.26(0.05–1.27)0.65(0.14–3.04)
Strongly agree0.28(0.06–1.26)0.18(0.04–0.91)c0.58(0.12–2.67)

aDemographic adjustment included race, income, age and gender.

bAdjusted for mood at baseline. Mood was measured by CES-D score.

cIndicated p < .05.

School and peers 

Better school performance and greater peer acceptance were protective against future risk of a depressive episode (Table 4). Higher grades in English and History classes predicted three- to fourfold lower risk for developing a depressive episode. Perceiving that “teachers were unfair” predicted a threefold increase in risk, but only in the bivariate model. Involvement in sports, not having failed a grade, and having a driver's license also had protective effects in the bivariate models; however, these effects did not persist after controlling for gender, income, age, ethnicity, or depressed mood. Agreement with statements such as feeling that one is “part of [his or her] school,” “close to [his or her] classmates,” “socially accepted,” “part of this school,” and “happy at [his or her] school” were associated with a four- to fivefold reduction in risk of a depressive episode. Reporting having “trouble getting along with other students” predicted a three- to fourfold increase in risk even after adjustment for demographic factors. Believing that “friends care about me,” having a best friend, lack of perceived prejudice, and engaging in outings with friends did not affect depression risk.

Table 4. School and peers
Odds ratio (95% confidence interval)
UnivariateDemographic adjustmentaCES-D adjustmentb
School
Have you ever repeated a grade?
No111
Yes1.68(1.09–2.6)c1.25(0.73–2.16)1.32(0.85–2.06)
At the most recent grading period, what was your English grade?
D or lower111
C0.47(0.18–1.23)0.51(0.18–1.44)0.45(0.17–1.2)
B0.42(0.17–1.05)0.37(0.14–0.99)0.46(0.18–1.23)
A0.31(0.1–0.9)c0.25(0.07–0.87)c0.4(0.13–1.2)
[Constructed from questions about which sports teams the respondent belongs to]
Does not belong to any teams111
1 team0.72(0.37–1.4)0.53(0.21–1.34)0.86(0.43–1.71)
2 teams0.7(0.31–1.57)0.98(0.41–2.39)1.01(0.44–2.34)
3 teams0.24(0.06–0.87)c0.42(0.11–1.53)0.37(0.1–1.38)
4 or more teams0.61(0.18–2.08)1.15(0.32–4.15)0.75(0.2–2.86)
At the most recent grading period, what was your grade history/social studies?
D or lower111
C0.36(0.12–1.04)0.55(0.16–1.85)0.41(0.14–1.23)
B0.65(0.28–1.48)0.84(0.3–2.38)0.85(0.35–2.08)
A0.24(0.09–0.69)c0.28(0.08–0.95)c0.4(0.13–1.2)
The teachers at this school treat students fairly.
Strongly agree111
Agree1.9(0.74–4.89)1.32(0.48–3.63)1.66(0.63–4.35)
Neither agree nor disagree1.57(0.57–4.29)1.08(0.34–3.43)1.06(0.37–2.99)
Disagree1.05(0.28–3.89)0.18(0.02–1.63)0.6(0.16–2.28)
Strongly disagree3.69(1.15–11.8)c2.3(0.61–8.63)1.6(0.46–5.51)
I feel safe in my school.
Strongly disagree111
Disagree0.56(0.12–2.5)0.74(0.14–3.84)0.89(0.19–4.29)
Neither agree nor disagree0.44(0.16–1.25)0.38(0.1–1.4)0.83(0.27–2.59)
Agree0.25(0.1–0.66)c0.35(0.1–1.18)0.66(0.21–2.09)
Strongly agree0.19(0.06–0.55)c0.23(0.06–0.82)c0.53(0.15–1.84)
Peers
I feel like I am part of this school.
Strongly disagree111
Disagree0.12(0.03–0.44)c0.04(0.01–0.23)c0.15(0.04–0.58)c
Neither agree nor disagree0.26(0.11–0.6)c0.26(0.1–0.68)c0.46(0.19–1.11)
Agree0.2(0.09–0.44)c0.2(0.08–0.55)c0.46(0.18–1.16)
Strongly agree0.18(0.08–0.43)c0.24(0.09–0.62)c0.53(0.2–1.41)
I feel socially accepted.
Strongly disagree111
Disagree0.17(0.05–0.62)c0.17(0.03–0.89)c0.28(0.06–1.24)
Neither agree nor disagree0.17(0.05–0.53)c0.15(0.03–0.67)c0.39(0.1–1.59)
Agree0.05(0.02–0.15)c0.04(0.01–0.18)c0.2(0.05–0.81)c
Strongly agree0.04(0.01–0.13)c0.04(0.01–0.16)c0.2(0.05–0.85)c
I feel close to people at this school.
Strongly disagree111
Disagree0.45(0.2–1.02)0.42(0.14–1.23)0.52(0.21–1.25)
Neither agree nor disagree0.25(0.12–0.53)c0.27(0.11–0.67)c0.32(0.14–0.74)c
Agree0.2(0.09–0.44)c0.22(0.09–0.57)c0.36(0.15–0.88)c
Strongly agree0.17(0.07–0.39)c0.16(0.05–0.48)c0.33(0.13–0.84)c
Since school started this year, how often have you had trouble: getting along with other students?
Never111
Just a few times0.84(0.5–1.4)0.96(0.51–1.78)0.63(0.38–1.05)
About once a week1.91(0.99–3.7)2.19(1.04–4.62)c0.94(0.46–1.92)
Almost every day1.92(0.87–4.26)2.6(0.99–6.8)0.92(0.4–2.11)
Every day3.17(1.31–7.67)c4.24(1.57–11.42)c1.3(0.44–3.84)
I'm happy to be at this school.
Strongly disagree111
Disagree0.27(0.08–0.87)c0.18(0.05–0.71)c0.41(0.12–1.42)
Neither agree nor disagree0.33(0.15–0.72)c0.27(0.11–0.7)c0.51(0.23–1.15)
Agree0.23(0.11–0.52)c0.22(0.09–0.54)c0.5(0.21–1.2)
Strongly agree0.13(0.05–0.33)c0.1(0.03–0.29)c0.29(0.11–0.82)c

aDemographic adjustment included race, income, age and gender.

bAdjusted for mood at baseline. Mood was measured by CES-D score.

cIndicates p < .05.

Community 

A supportive community and constructive involvement through religious activities conferred some protection against new-onset depressive episodes, whereas delinquency and adverse events increased risk (Table 5). For example, believing that “adults care about [me]” predicted lower risk for depression, but not after adjustment for baseline depressed mood. Many religious activities were strongly protective in all models (twofold decrease in risk); protective activities included weekly prayer, describing oneself as an adherent follower of an organized religion, and attendance at a religious youth group (but not regular services). Having neighbors of the same age, talking to neighbors, thinking that one's neighbor's care about him or her, the absence of neighborhood litter and drug problems, and feeling safe in one's neighborhood were not protective. Criminal activity, drug and alcohol use, sexual activity, emotional trauma, and physical trauma predicted a twofold increase in risk. These relations between drug use and sexual activity and depressive episodes persisted even after adjustment for depressed mood. Interestingly, behaviors like lying to parents, skipping school, smoking cigarettes, running away from home, and using drugs other than inhalants and marijuana were not associated with an increased risk of depression.

Table 5. Community environment and engagement
Odds ratio (95% confidence interval)
UnivariateDemographic adjustmentaCES-D adjustmentb
Community involvement
How much do you feel adults care about you?
Not at all111
Very little111
Somewhat0.68(0.28–1.64)0.78(0.25–2.42)0.73(0.27–1.96)
Quite a bit0.39(0.16–0.96)c0.35(0.11–1.14)0.68(0.25–1.87)
Very much0.31(0.13–0.74)c0.32(0.1–0.97)0.65(0.24–1.76)
Do you have a drivers license? [excluded if under 16 years old]
No111
Yes0.4(0.21–0.77)c0.69(0.29–1.66)0.52(0.26–1.03)
What is your religion?
None, do not know, or refused111
Responded with religion0.53(0.33–0.87)c0.52(0.29–0.91)c0.57(0.34–0.96)c
How often do you pray?
Never111
Less than once a month0.37(0.13–1.05)0.32(0.1–1.07)0.35(0.13–0.98)
At least once a month0.46(0.21–0.98)0.49(0.21–1.14)0.48(0.22–1.06)
At least once a week0.51(0.29–0.88)c0.5(0.26–0.94)c0.52(0.29–0.94)c
At least once a day0.57(0.34–0.94)c0.42(0.22–0.79)c0.66(0.38–1.14)
Many churches, synagogues, and other places of worship have special activities for teenagers such as youth groups, Bible classes, or choir. In the past 12 months, how often did you attend such youth activities?
Never1 1
Less than once a month0.71(0.37–1.39)0.66(0.32–1.38)0.72(0.36–1.41)
At least once a month0.34(0.14–0.79)c0.32(0.11–0.91)c0.37(0.15–0.88)c
Once a week or more0.54(0.3–0.96)c0.36(0.17–0.77)c0.62(0.34–1.12)
Delinquency
Committed a violent crime in previous year (constructed).
Did not commit violent crime111
Committed violent crime1.9(1.31–2.75)c2.03(1.36–3.04)c1.35(0.91–2)
In the last 18 months—since {MONTH, YEAR}—have you had a special romantic relationship with any one?
Yes1.74(1.15–2.64)c2.01(1.17–3.46)c1.33(0.86–2.04)
No111
Have you ever had sexual intercourse? When we say sexual intercourse, we mean when a male inserts his penis into a female's vagina.
No111
Yes1.74(1.15–2.64)c2.01(1.17–3.46)c1.33(0.86–2.04)
During your life, how many times have you used marijuana?
Never111
1–10 times2.87(1.73–4.77)c2.54(1.35–4.76)c2.13(1.26–3.61)c
10+ times2.41(1.41–4.11)c3.18(1.65–6.12)c1.52(0.86–2.69)
During your life, how many times have you used inhalants, such as glue or solvents?
Never111
At least once2.2(1.18–4.08)c2.25(1.11–4.57)c1.37(0.71–2.64)
Adverse events
Number of different kinds of emotional trauma experienced in previous year (constructed).
No events111
1 or more events2.43(1.58–3.75)c2.23(1.35–3.66)c1.6(1–2.55)
Number of different kinds of physical trauma experienced in previous year(constructed).
No events111
1 or more events1.58(0.99–2.53)1.77(1.06–2.94)c1.19(0.74–1.92)
Have any of your friends tried to kill themselves during the past 12 months?
No events111
1 or more events1.58(0.99–2.53)1.77(1.06–2.94)c1.19(0.74–1.92)

aDemographic adjustment included race, income, age and gender.

bAdjusted for mood at baseline. Mood was measured by CES-D score.

cIndicates p < .05.

Sensitivity analysis 

Comparison of the odds ratios and 95% confidence intervals between the original and the alternative outcome measure (CESD >28) are qualitatively similar across all models, with a few exceptions highlighted below. Several predictor variables—“lots of energy,” “feel fit,” “go with gut without thinking,” dad's warmth, and “closeness to dad”—were not significant when CESD >28 was used as the dependent variable after adjustment for baseline depressed mood. In the case of “like self as is” and “family has fun together,” there are no significant relationships between the variable and the CES-D >28 outcome.

Back to Article Outline

Discussion 

This study gives insight into the development of new-onset depressive episodes in adolescents. Nonwhite ethnicity, lower household income, female gender, headaches, greater levels of depressed mood (and poorer affect regulation) predicted greater risk of subsequent depressive episodes. Higher levels of fitness and sleep quality, active coping, and favorable self-concept were all protective. Although parental conflict increased risk, family connectedness, shared activities, and warmth were strongly protective even after adjustment for baseline depressed mood. Better school performance, greater connectedness to school, and higher levels of peer acceptance reduced future risk of depressive episodes. Feeling that adults care and greater participation in religious activities reduced future risk of depression, whereas delinquent activities and greater numbers of adverse events increased risk.

Female gender, nonwhite ethnicity, lower levels of parent education and income, and worse physical health predicted increased risk of a depressive episode. The observed twofold increase in risk for depression in females, older age, and more somatic symptoms is consistent with previous research [5]. Although inconsistent, previous research has generally found an association between being white and a lower incidence of depression [30]. Our results are consistent with prior research that demonstrates low education status and poverty have been associated with risk of depressive episodes in children [31], [32].

Our results confirm the strong connections between robust active coping approaches, favorable view of self, more effective affect regulation, lower levels of current depressed mood, and lower risk of depressive episodes that have been found in smaller, less representative study samples. The protective effects of positive self-concept and active coping [33] are consistent with prior research, but the persistence of their effects after adjustment for baseline depressed mood is a new finding. Academic success has been shown to potentially reduce the risk of depressive episodes; our results support that finding [34]. The lack of effect of higher participation in hobbies or school activities may suggest that declines in these forms of behavioral activation occur closer to onset [34]. Consistent with prior work, better affect regulation and lower levels of current depressed mood are associated with greater risk of future depression [35].

Consistent with prior studies, supportive peer and family relationships and connection to community were highly protective in this study [36]. This study's findings on peer support place new importance on a sense of acceptance and belonging and on lack of conflict [37]. Acceptance, rather than warmth or support, may be the critical protective factor in the peer context [37]. Family connectedness, including a sense of closeness to either parent, positive parent–child relations, and global family function, appear to be the primary protective family factors, even after adjustment for baseline depressed mood [36]. Although previous work has suggested conflicting results [38], we found that religious belief and practice were protective after adjustment for demographic characteristics and baseline depressed mood. Delinquent behaviors, sexual intercourse, and adverse events predicted lower future risk of a major depressive episode consistent with prior work [39].

The greatest strengths of this study are the universality of the sample, the longitudinal design, and availability of multiple variables relating to established models of depression vulnerability and protection that enable us take a broader view of the individual adolescent in social context. Although the National Longitudinal Study of Adolescent Health is now more than a decade old, it remains the most recent, representative cohort study of U.S. adolescents [10], [11], [12], [13], [14], [15], [16], [17], [18], [19], [20], [29]. The primary limitations of this study centered on the construction of a derived outcome variable from the CES-D and multiple colinearity of the outcome variables However, our CES-D method was validated by Radloff, and has been used in previous studies; the levels of depression severity are consistent with clinical populations; and the incidence of major depression in our sample is similar to that found in other cross-sectional incidence studies using community samples [27]. Another limitation of this study is the absence of family mental health history and genetic information to evaluate gene/environment interactions.

Back to Article Outline

Conclusions 

Our results suggest that vulnerability and protection against the onset of depressive episodes is derived from multiple fields of adolescent experience and engagement, with experiences within the family playing a preeminent role. Physicians should be aware that low-income status, nonwhite ethnicity, physical symptoms, family discord, lack of peer acceptance, delinquency, and adverse events may increase risk for depression. Advising families on ways to reduce conflict and enhance family warmth should also be considered. Recommending active coping, problem solving, and involvement in youth-oriented community and religious activities to adolescents may be reasonable. Prevention investigators should consider alternative preventive approaches that are less abstract and more pragmatic (compared with the current, primarily cognitive–behavioral interventions). Tangible and easily defined behaviors, such as increasing religious and school involvement and pleasurable family activities, may be particularly amenable to community-based models that employ lay leaders such as teachers, guidance counselors, and youth pastors. On a policy level, public health officials, school administrators, and religious leaders should consider the important roles they can play in protecting adolescents from depression by providing them with a supportive and engaging community setting.

Back to Article Outline

Acknowledgements 

Dr. Van Voorhees is supported by a NARSAD Young Investigator Award, Robert Wood Johnson Foundation Depression in Primary Care Value Grant, and a Career Development Award from the National Institutes of Mental Health (NIMH K-08 MH 072918-01A2). The support of Dorothy Reeves Williams (1916–2005), a long-time contributor and supporter of the NARSAD foundation, is gratefully acknowledged.

Back to Article Outline

References 

  1. Kessler RC, Walters EE. Epidemiology of DSM-III-R major depression and minor depression among adolescents and young adults in the National Comorbidity Survey. Depress Anxiety. 1998;7(1):3–14
  2. Friedman RA. Uncovering an epidemic—screening for mental illness in teens. N Engl J Med. 2006;355(26):2717–2719
  3. Allan V, Horwitz JCW. The Loss of Sadness: How Psychiatry Transformed Normal Sorrow into Depressive Disorder. New York: Oxford University Press; 2007;
  4. Luthar SS, Cicchetti D, Becker B. The construct of resilience: a critical evaluation and guidelines for future work. Child Dev. 2000;71(3):543–562
  5. Hankin BL. Adolescent depression: description, causes, and interventions. Epilepsy Behav. 2006;8(1):102–114
  6. Vulnerability to depressionamong adolescents: implications for cogntive treatment. Reinecke MaS, A. Cognit Behav Pract. 2005;12:166–176
  7. Gutman LM, Sameroff AJ. Continuities in depression from adolescence to young adulthood: contrasting ecological influences. Dev Psychopathol. 2004;16(4):967–984
  8. Lewinsohn PM, Hops H, Roberts RE, et al. Adolescent psychopathology: I (Prevalence and incidence of depression and other DSM-III-R disorders in high school students). J Abnorm Psychol. 1993;102(1):133–144
  9. Goodman SH, Lahey BB, Fielding B, et al. Representativeness of clinical samples of youths with mental disorders: a preliminary population-based study. J Abnorm Psychol. 1997;106(1):3–14
  10. Reinherz HZ, Giaconia RM, Hauf AMC, et al. Major depression in the transition to adulthood: Risks and impairments. J Abnorm Psychol. 1999;108(3):500–510
  11. Hankin BL, Abramson LY, Moffitt TE, Silva PA, McGee R, Angell KE. Development of depression from preadolescence to young adulthood: emerging gender differences in a 10-year longitudinal study. J Abnorm Psychol. 1998;107(1):128–140
  12. Allen JP, Insabella G, Porter MR, et al. A social-interactional model of the development of depressive symptoms in adolescence. J Consult Clin Psychol. 2006;74(1):55–65
  13. Cole DA, Martin JM, Peeke LG, et al. Are cognitive errors of underestimation predictive or reflective of depressive symptoms in children: a longitudinal study. J Abnorm Psychol. 1998;107(3):481–496
  14. Fergusson DM, Horwood LJ, Ridder EM, Beautrais AL. Subthreshold depression in adolescence and mental health outcomes in adulthood. Arch Gen Psychiatry. 2005;62(1):66–72
  15. Fombonne E, Wostear G, Cooper V, et al. The Maudsley long-term follow-up of child and adolescent depression. 1 (Psychiatric outcomes in adulthood). Br J Psychiatry. 2001;179:210–217
  16. Fombonne E, Wostear G, Cooper V, et al. The Maudsley long-term follow-up of child and adolescent depression. 2 (Suicidality, criminality and social dysfunction in adulthood). Br J Psychiatry. 2001;179:218–223
  17. Knapp M, McCrone P, Fombonne E, et al. The Maudsley long-term follow-up of child and adolescent depression: 3 (Impact of comorbid conduct disorder on service use and costs in adulthood). Br J Psychiatry. 2002;180:19–23
  18. Lee AS, Murray RM. The long-term outcome of Maudsley depressives. Br J Psychiatry. 1988;153:741–751
  19. McCrone P, Knapp M, Fombonne E. The Maudsley long-term follow-up of child and adolescent depression (Predicting costs in adulthood). Eur Child Adolesc Psychiatry. 2005;14(7):407–413
  20. Resnick MD, Bearman PS, Blum RW, et al. Protecting adolescents from harm (Findings from the National Longitudinal Study on Adolescent Health). JAMA. 1997;278(10):823–832
  21. National Longitudinal Study of Adolescent Health (Add Health), Wave I and II, 1994-1996. Los Altos, CA: Sociometrics Corporation; 1998;
  22. Oxman TE, Sengupta A. Treatment of minor depression. Am J Geriatr Psychiatry. 2002;10(3):256–264
  23. In:  Attkisson CC,  Zich JM editor. Depression in Primary Care: Screening and Detection. lorence, KY: Taylor and Francis/Routledge; 1990;
  24. Lewinsohn PM, Gotlib IH, Seeley JR. Adolescent psychopathology: IV (Specificity of psychosocial risk factors for depression and substance abuse in older adolescents). J Am Acad Child Adolesc Psychiatry. 1995;34(9):1221–1229
  25. Schoenbach V, Kaplan BH, Wagner EH, et al. Prevalence of self-reported depressive symptoms in young adolescents. Am J Publ Health. 1983;73:1281–1287
  26. Schoenbach VJ, Kaplan BH, Grimson RC, Wagner EH. Use of a symptom scale to study the prevalence of a depressive syndrome in young adolescents. Am J Epidemiol. 1982;116(5):791–800
  27. Radloff LS. The use of the Center for Epidemiologic Studies Depression Scale in adolescents and young adults. J Youth Adolesc. 1991;20(2):149–166
  28. In:  Udry JR editors. The National Longitudinal Study of Adolescent Health: References, Instruments, and Questionnaires Consulted in the Development of the Adhealth In-home Adolescent Interview. Chapel Hill, NC: University; 2001;
  29. Rushton JL, Forcier M, Schectman RM. Epidemiology of depressive symptoms in the National Longitudinal Study of Adolescent Health. J Am Acad Child Adolesc Psychiatry. 2002;41(2):199–205
  30. Roberts R, Roberts CR, Chen R. Ethnocultural differences in prevalence of adolescent depression. Am J Community Psychol. 1997;25:95–110
  31. Klein DN, Lewinsohn PM, Rohde P, et al. Psychopathology in the adolescent and young adult offspring of a community sample of mothers and fathers with major depression. Psychol Med. 2005;35(3):353–365
  32. Pachter LM, Auinger P, Palmer R, Weitzman M. Do parenting and the home environment, maternal depression, neighborhood, and chronic poverty affect child behavioral problems differently in different racial-ethnic groups?. Pediatrics. 2006;117(4):1329–1338
  33. Lewinsohn PM, Gotlib IH, Seeley JR. Depression-related psychosocial variables: are they specific to depression in adolescents?. J Abnorm Psychol. 1997;106(3):365–375
  34. Boone EL, Leadbetter BJ. Game on: diminishing risks for depressive symptoms in early adolescence through positive involvement in team sports. J Res Adolesc. 2006;16(1):79–90
  35. Silk JS, Steinberg L, Morris AS. Adolescents' emotion regulation in daily life: links to depressive symptoms and problem behavior. Child Dev. 2003;74(6):1869–1880
  36. Lewinsohn PM, Roberts RE, Seeley JR, et al. Adolescent psychopathology: II (Psychosocial risk factors for depression). J Abnorm Psychol. 1994;103(2):302–315
  37. Liu Y-L. The role of perceived social support and dysfunctional attitudes in predicting Taiwanese adolescents' depressive tendency. Adolescence. 2002;37(148):823–834
  38. Cotton S, Larkin E, Hoopes A, et al. The impact of adolescent spirituality on depressive symptoms and health risk behaviors. J Adolesc Health. 2005;36(6):529
  39. Beyers JM, Loeber R. Untangling developmental relations between depressed mood and delinquency in male adolescents. J Abnorm Child Psychol. 2003;31(3):247–266

PII: S1054-139X(07)00494-6

doi:10.1016/j.jadohealth.2007.11.135

Journal of Adolescent Health
Volume 42, Issue 6 , Pages 605-616, June 2008

Article Tools

* Image Usage & Resolution