Main

Adolescent depression has been a focus of recent attention (Cole et al. 1999; Goodman and Capitman 2000). Depressive symptoms during adolescence are associated with a number of adverse health consequences including suicide and future risk of major depressive disorder (MDD) (Frost et al. 1999). Recent longitudinal, population-based studies have examined predictors of development of depressive symptoms in adolescents. In a previous study, we found that gender and cigarette smoking were associated with the onset of depressive symptoms (Choi et al. 1997). In contrast to research examining the course of MDD (Goodyer et al. 1991), there are few epidemiological data on the persistence of depressive symptoms during adolescence. The factors that promote development of depressive symptoms may differ from those that contribute to persistence or remission of depression. An improved understanding of the persistence of depressive symptoms as well as the potentially modifiable correlates of depression is needed to develop more effective intervention programs for depressed adolescents.

In this study, we examine factors associated with persistence of depressive symptoms in a longitudinal sample of 1,176 United States adolescents.

METHODS

Data were gathered from the 1989 National Teenage Attitudes and Practices Survey (TAPS I) (Allen et al. 1992). TAPS I succeeded in interviewing 9,965 adolescents aged 12–18 years. Only 9,135 respondents reached by telephone were eligible for follow-up survey in 1993. The follow-up telephone survey (TAPS II) was completed in 1993 by 87% of these adolescents (n = 7,960) who by that time were aged 15–22 years. The National Health Interview Survey (NHIS) uses a multistage sample design to provide national estimates of the non-institutionalized population. All data were weighted according to probability of selection and to represent the United States adolescent population with respect to gender, age, and race/ethnicity. This paper focuses on the 1,176 longitudinal-sample adolescents who reported notable depressive symptoms at baseline.

The measures are described in detail in another report (Patten et al. 2000). Demographic and psychosocial predictors assessed in 1989 included age, gender, race/ethnicity, cigarette smoking status, and sleep problems. A 5-item modified version of a previously validated scale (Kandel & Davies 1982) was used to assess depressive symptoms in 1989 and 1993. Total scores can range from 15 to 30, and scores of ⩾21 indicate notable depressive symptoms.

RESULTS

Table 1 presents the baseline distribution of adolescents reporting notable depressive symptoms in 1993 among the 1,176 reporting notable depressive symptoms in 1989. Of these 1,176 adolescents (433 girls, 743 boys), 453 (38.5%) reported notable depressive symptoms at follow-up. Using multiple logistic regression, significant predictors of notable depressive symptoms in 1993 among those with notable depressive symptoms in 1989 were female gender, engaging in physical fights, and lack of participation in sports (Table 2). As shown in Figure 1, changes in sleep problems and cigarette smoking status from 1989 to 1993 were also predictive.

Table 1 Demographic Characteristics of Persistent Depression, Among Adolescents With Notable Depressive Symptoms at Baseline (N = 1,176)
Full size table
Table 2 Multivariate Predictors of Persistent Depressive Symptoms
Full size table
Figure 1
figure 1

Percentage of adolescents with notable depressive symptoms in 1993 by changes in frequent sleep problems and cigarette smoking status among all 1,176 adolescents with notable depressive symptoms in 1989. YY indicates those adolescents with frequent sleep problems at both years; YN indicates those who reported frequent depressive symptomsat baseline but not at follow-up; NY indicates those without frequent sleep problems at baseline but who reported frequent sleep problems at follow-up; NN indicates those without frequent sleep problems at both years. Similar coding was used for cigarette smoking status.

Full size image

DISCUSSION

This study adds to the knowledge on persistence of depressive symptoms from an epidemiological sample of United States adolescents. The study of the trajectory of depressive symptoms and predictors of persistent depression are important components of a developmental psychopathology perspective of adolescent depression (Cole et al. 1999). Nearly 40% of adolescents reported continued depressive symptoms over a period spanning four years. Assuming the majority of depressed adolescents are not receiving treatment, the findings suggest the need for secondary interventions to prevent escalation to more serious problems (e.g., suicidal attempts, MDD, alcohol abuse). In particular, future research is needed to evaluate whether interventions focused on reduction or prevention of sleep problems and cigarette smoking reduce the likelihood of continued depression in adolescents.