Lifetime comorbidity of DSM-IV disorders in the NCS-R Adolescent Supplement (NCS-A)

Abstract

Background

Research on the structure of comorbidity among common mental disorders has largely focused on current prevalence rather than on the development of comorbidity. This report presents preliminary results of the latter type of analysis based on the National Comorbidity Survey Replication Adolescent Supplement (NCS-A).

Methods

A national survey was carried out of adolescent mental disorders. DSM-IV diagnoses were based on the Composite International Diagnostic Interview administered to adolescents and questionnaires self-administered to parents. Factor analysis examined comorbidity among 15 lifetime DSM-IV disorders. Discrete-time survival analysis was used to predict first onset of each disorder from information about prior history of the other 14 disorders.

Results

Factor analysis found four factors representing fear, distress, behavior, and substance disorders. Associations of temporally primary disorders with the subsequent onset of other disorders (dated using retrospective age-of-onset reports) were almost entirely positive. Within-class associations (e.g., distress disorders predicting subsequent onset of other distress disorders) were more consistently significant (63.2%) than between-class associations (33.0%). Strength of associations decreased as comorbidity among disorders increased. The percent of lifetime disorders explained (in a predictive rather than causal sense) by temporally prior disorders was in the range 3.7-6.9% for earliest-onset disorders (specific phobia and attention-deficit/hyperactivity disorder) and much higher (23.1-64.3%) for later-onset disorders. Fear disorders were the strongest predictors of most other subsequent disorders.

Conclusions

Adolescent mental disorders are highly comorbid. The strong associations of temporally primary fear disorders with many other later-onset disorders suggest that fear disorders might be promising targets for early interventions.

Epidemiological surveys of child-adolescent mental disorders consistently find high comorbidity (Angold et al., 1999; Fergusson et al., 1994; Newman et al., 1996). A number of longitudinal studies have examined temporal progression among these comorbid disorders (Pine et al., 1998; Reinke & Ostrander, 2008; Stein et al., 2001). However, these studies have generally (Costello et al., 1996; Fergusson et al., 1993; McGee et al., 1992), although not always (Lieb et al., 2000; Shankman et al., 2009), focused on point prevalence in each wave or the cumulation of point prevalence estimates across multiple waves (Costello et al., 2003; Kim-Cohen et al., 2003). For example, the Great Smoky Mountain Study assessed 3-month mental disorders once yearly in three cohorts of children (Costello et al., 2003). This approach makes it impossible to distinguish associations of primary disorders with onset versus persistence of secondary disorders. Such a distinction, which would require examining temporal sequencing across lifetime rather than recent disorders, could have value in interpreting longitudinal associations and strategizing about intervention possibilities (Angold et al., 1999). There are exceptions, though. For example, the Early Developmental Stages of Psychopathology Study (Wittchen et al., 1998) is a large-scale longitudinal study that examined temporal sequencing of lifetime disorders in a cohort of adolescents followed into adulthood. Among the important findings this design allowed are that temporally prior GAD predicts the subsequent onset of other anxiety disorders more strongly than the subsequent onset of depression (Beesdo et al., 2010) and that social phobia is associated with subsequent onset of temporally secondary depression (Beesdo et al., 2007). However, this kind of analysis has never been used to examine temporal sequencing of disorder onsets across the full range of commonly occurring mental disorders.

The current report presents results of a preliminary analysis of the associations between temporally prior lifetime disorders and the subsequent first onset of temporally secondary disorders based on the data collected in the National Comorbidity Survey Replication Adolescent Supplement (NCS-A), a national survey of adolescent DSM-IV disorders (Merikangas et al., 2009). The results are only preliminary because the data are cross-sectional and temporal associations are inferred from retrospective age-of-onset (AOO) reports. Nonetheless, intriguing results emerge that could be useful in generating hypotheses to be examined in subsequent longitudinal studies. Previous studies would lead us to expect stronger inter-temporal associations within than between anxiety-mood and behavioral disorders (Copeland et al., 2009; Reinke & Ostrander, 2008) and strong associations of temporally primary behavior disorders with later substance disorders (Copeland et al., 2009; Elkins et al., 2007; Tapert et al., 2002), but we had no initial hypotheses about other specifications.

METHODS

Sample

Adolescents (ages 13-17) were interviewed between February 2001 and January 2004 in dual-frame household and school samples described elsewhere (Kessler et al., 2009a; Kessler et al., 2009b). The household sample included 904 adolescents (879 in school, 25 dropouts) from households in the National Comorbidity Survey Replication (NCS-R) (Kessler & Merikangas, 2004). The school sample included 9,244 adolescents from a representative sample of schools in the adult sample areas of that survey. The conditional adolescent response rate was 86.8-82.6% (household and school samples, respectively). The household sample includes school dropouts and adolescents residing in areas where schools refused to participate. A high (72.0%) percent of initially-selected schools refused to participate and were replaced with matched replacement schools. Comparison of household sample respondents from nonparticipating schools with school sample respondents from replacement schools found no evidence of bias in estimates of either disorder prevalence or correlates (Kessler et al., 2009a).

One parent or surrogate (henceforth referred to as parents) of participating adolescents was asked to complete a self-administered questionnaire about the adolescent’s developmental history and mental health. The conditional response rate was 82.5-83.7% (household and school samples, respectively). This report focuses on the 6,483 adolescent-parent pairs where complete data are available from both adolescents and parents. The fact that parent data were available for only a subset of adolescent respondents was taken into consideration by weighting the data in complete parent-adolescent pairs to adjust for differences with incomplete pairs. These weighting procedures are discussed in detail elsewhere (Kessler et al., 2009a; Kessler et al., 2009b).

Written parent informed consent and adolescent assent were obtained before surveying either the parent or adolescent. Each respondent was given $50 for participation. These recruitment and consent procedures were approved by the Human Subjects Committees of both Harvard Medical School and the University of Michigan. Once the survey was completed, cases were weighted for within-household probability of selection (household sample) and deviation from Census population socio-demographic/geographic distributions, making each sample nationally representative on the socio-demographic/geographic variables. The samples were then merged with sums of weights proportional to relative sample sizes adjusted for design effects in estimating disorder prevalence. These procedures are detailed elsewhere (Kessler et al., 2009a; Kessler et al., 2009b).

Diagnostic assessment

Adolescents were administered the fully-structured Composite International Diagnostic Interview (CIDI) modified to simplify language and use examples relevant to adolescents (Merikangas et al., 2009). The 15 DSM-IV disorders assessed include mood disorders (major depressive disorder/dysthymia, bipolar I-II disorder and sub-threshold bipolar disorder), anxiety disorders (panic disorder with or without agoraphobia, agoraphobia without panic disorder, social phobia, specific phobia, generalized anxiety disorder, post-traumatic stress disorder, separation anxiety disorder), behavior disorders (attention-deficit/hyperactivity disorder, oppositional-defiant disorder, conduct disorder, eating disorders [anorexia nervosa, bulimia nervosa, binge-eating behavior]), and substance disorders (alcohol and drug abuse, alcohol and drug dependence with abuse). There were no other exclusionary diagnoses. These disorders include all those assessed in most previous adolescent epidemiological studies.

Adolescent interviews assessed all disorders. Parent questionnaires assessed only disorders for which parent reports have previously been found important in diagnosis: behavior disorders (Johnston & Murray, 2003) and depression/dysthymia (Braaten et al., 2001). Parent and adolescent reports were combined at the symptom level using an “or” rule (except in the case of attention-deficit/hyperactivity disorder, where only parent reports were used based on evidence of invalidity of adolescent reports). All diagnoses were made using DSM-IV distress/impairment criteria and organic exclusion rules, but diagnostic hierarchy rules were not used because we wanted to study comorbidity among hierarchy-free disorders.

A clinical reappraisal study interviewed adolescent-parent pairs by telephone with the Schedule for Affective Disorders and Schizophrenia for School-Age Children (K-SADS) Lifetime Version (Kaufman et al., 1997). As detailed elsewhere, concordance was good between survey and clinical diagnoses (Kessler et al., 2009c), with area under the receiver operating characteristic curve (AUC) of .81-.94 for fear disorders, .79-.86 for distress disorders, .78-.98 for behavior disorders, and .92-.98 for substance disorders. Parent and adolescent reports both contributed to AUC when both were assessed for depression/dysthymia (.75, .71, and .87 for adolescent, parent, and combined reports, respectively), oppositional-defiant disorder (.71, .66, and .85), and conduct disorder (.59, .96, and .98), but only parent reports contributed to AUC for attention-deficit/hyperactivity disorder (.57, .71, and .78). Adolescent disorder AOO reports were obtained retrospectively using probes shown experimentally to maximize recall accuracy among adults (Knauper et al., 1999).

Analysis Methods

AOO curves were generated using the actuarial method (Halli et al., 1992). Tetrachoric factor analysis (principal axis method) with promax rotation was used to examine bivariate comorbidity. Temporal unfolding was studied by examining predictive associations between temporally primary disorders (based on retrospective AOO reports) and first onset of later disorders with multivariate discrete-time survival models using a person-year dataset (Willett & Singer, 1993). The details of this modeling procedure, which we have used extensively in previous NCS reports, are presented elsewhere (Kessler et al., 2005). Each model predicted first onset of one disorder from information about prior lifetime occurrence of the other 14 disorders controlling basic socio-demographic variables (sex, race/ethnicity [Non-Hispanic White, Non-Hispanic Black, Hispanic, Other], region of the country [Northeast, Midwest, South, West], urbanicity [Major Metropolitan Area, other urbanized area, rural area], parent education [less than high school, high school, some college, college graduate, coded for the parent with the higher level of education], number of biological parents with whom the adolescent lives [0-2], birth order [only child, oldest, youngest, other], number of siblings, and age).

Several more complex models allowed non-additive associations to exist among the predictor disorders. One of these added to the additive multivariate model a series of dummy variables for number of prior lifetime disorders beginning with two. These dummy variables represent gross interaction terms that require the coefficient to be the same for all combinations of disorders of a given number (e.g., all pairs of disorders, all sets of three disorders). A somewhat more complex model weighted the disorders in each of these sets to have differential slopes proportional to the main effects of the disorders. Another specification allowed for interactions between a continuous measure of number of prior lifetime disorders and each type of prior lifetime disorders. More detailed information about these models are presented in a previous report in this journal (Alonso et al., 2011). The best-fitting model among these alternatives was selected using the Akaike Information Criterion (AIC) (Burnham & Anderson, 2002). Survival coefficients and standard errors were exponentiated to produce odds-ratios (ORs) and 95% confidence intervals.

Population attributable risk proportions (PARPs) were calculated to describe strength of association between predictor disorders and outcome disorders. PARPs are the proportions of outcome disorders that would not have occurred in the absence of predictor disorders under the model if the survival coefficients represent causal effects. The proportions were calculated by generating a conditional predicted probability of first onset of each outcome disorder at each year of life of each respondent from the coefficients in the best-fitting survival model twice: once using all model coefficients and then again omitting coefficients for the predictor disorders. The actuarial method (Halli et al., 1992) was used to cumulate conditional predicted probabilities to respondent age at interview. PARP was defined as 1-R, where R represents the ratio of mean cumulative predicted probability in the second specification divided by mean cumulative predicted probability in the first specification.

As the survey data are both clustered and weighted, the design-based Taylor series linearization method implemented in the SUDAAN software system (Research Triangle Institute, 2002) was used to estimate standard errors of prevalence estimates and 95% confidence intervals of ORs. Significance of predictor sets was evaluated using Wald χ² tests based on design-based variance-covariance matrices. Statistical significance was consistently evaluated using .05-level two-sided tests.

RESULTS

Lifetime prevalence and comorbidity

A previous NCS-A report documented high lifetime prevalence of disorders (Merikangas et al., 2010). Overall prevalence is even higher in the current data because we included sub-threshold bipolar disorder, which was omitted from the earlier report. (Table 1) Comorbidity is also more common because we analyze hierarchy-free diagnoses, with 27.9% of respondents meeting criteria for two or more disorders and a mean of 3.5 disorders among those with comorbidity. All but two of the 105 tetrachoric correlations among disorder pairs (15×14/2) are positive (82.5% statistically significant), with a median of .29 and inter-quartile range (25^th-75^th percentiles) of .20-.37. (The tetrachoric correlation matrix is available on request.)

Table 1. Lifetime prevalence of estimated DSM-IV disorders along with standard errors of prevalence estimates (se) and rotated (promax) factor pattern of comorbid disorders in the sub-sample with complete parent data (n=6,483)^a.

	Prevalenceb		Rotated Factor Pattern (Standardized Regression Coefficients)
	%	se	I. Fear disorders	II. Distress disorders	III. Behavior disorders	IV. Substance Disorders
I. Fear disorders
Specific phobia	19.9	1.0	0.70	0.06	0.12	−0.10
Agoraphobiac	2.6	0.4	0.79	0.10	−0.13	0.00
Social phobia	8.5	0.6	0.67	0.09	0.03	0.09
Panic disorderd	2.4	0.2	0.68	−0.07	−0.10	0.15
Any fear disorder	26.1	1.0
II. Distress disorders
Separation anxiety disorder	7.6	0.5	0.37	0.51	0.11	−0.27
Post-traumatic stress disorder	4.7	0.4	0.02	0.79	−0.03	0.13
Major depressive episode/dysthymia	18.6	1.1	0.04	0.53	0.38	0.17
Generalized anxiety disorder	2.2	0.4	0.07	0.87	−0.08	0.01
Any distress disorder	25.4	0.9
III. Behavior disorders
Attention-deficit/hyperactivity disorder	8.1	0.6	−0.15	−0.13	0.94	−0.07
Oppositional defiant disorder	12.6	0.9	0.04	0.01	0.70	0.25
Conduct disorder	6.8	0.9	−0.02	0.09	0.49	0.44
Eating disorderse	5.1	0.4	0.20	0.11	0.54	−0.14
Any behavior disorder	22.7	1.3
IV. Substance disorders
Alcohol abusef	6.1	0.5	−0.18	0.23	−0.11	0.89
Drug abusef	8.9	0.8	0.10	−0.05	0.02	0.91
Any substance disorder	11.4	0.9
V. Other disorders
Bipolar disorderg	6.2	0.4	0.38	−0.23	0.10	0.40
VI. Total number of disorders
Any disorder	51.3	1.3
Exactly one disorder	23.4	0.9
Exactly two disorders	12.0	0.6
Three or more disorders	15.9	1.0

^aEigenvalues of unrotated factors: 5.2, 2.2, 1.1, 1.0, 0.9. Correlations among factors in promax rotated four-factor solution: I-II .44; I-III .25, I-IV .16, II-III .31, II-IV .29, III-IV .43.

^bThe prevalence estimates reported here are higher than those in an earlier NCS-A report (Merikangas et al., 2010) because disorders were defined here without diagnostic hierarchy rules and included sub-threshold bipolar disorder. Estimated comorbidity is higher than in the earlier report because of not using hierarchy rules.

^cAgoraphobia is assessed with or without panic disorder

^dPanic disorder is assessed with or without agoraphobia

^eIncludes anorexia nervosa, bulimia nervosa, and binge-eating disorder

^fWith or without dependence

^gIncludes Bipolar I, Bipolar II, and sub-threshold bipolar disorder

Factor analysis finds four factors with unrotated eigenvalues greater than 1.0. Oblique (promax) rotation shows that the first factor is similar to what previous studies referred to as a fear factor (Krueger, 1999; Watson, 2005), with high loadings for panic and phobias (.67-.79).(Table 1) The second factor represents what previous studies referred to as a distress factor, including depression, generalized anxiety, post-traumatic stress, and separation anxiety disorders (.51-.87). The third factor represents behavior disorders (.49-.94) and the fourth represents alcohol-drug disorders (.89-.91), although conduct disorder (.44) and bipolar disorder (.40) also have elevated loadings. Disorders in the fear factor are the most common (26.1%), followed by distress (25.4%), behavior (22.7%), and substance (11.4%) disorders. The factors are all significantly correlated with each other, from a high Pearson correlation of .44 between fear and distress disorders to a low correlation of .16 between fear and substance disorders.

Age-of-onset distributions

Retrospectively reported median (and inter-quartile range) AOO of any disorder is 8 (6-13) years old. A clear temporal order exists across disorders within classes. (Figure 1) Specific phobia has the earliest AOO (Median) within fear disorders (6), separation anxiety disorder within distress disorders (8), and attention-deficit/hyperactivity disorder within behavior disorders (6). The next earliest AOO distributions are for other fear (11-12) and behavior (12-13) disorders. Bipolar disorder (14), the remaining distress disorders (14), eating disorders (15), and substance disorders (16) have the latest median AOO.

Figure 1.

Age-of-onset (AOO) distributions based on retrospective age-of-onset reports among respondents with lifetime DSM-IV/Composite International Diagnostic Interview disorders

Bivariate associations between earlier and later disorders

We estimated 210 (14×15) bivariate survival models, each with one lifetime disorder as a time-varying predictor of subsequent first onset of one other disorder, controlling socio-demographics. 89.3% of coefficients were positive and 59.5% significant. (Detailed results are available on request.) The median (inter-quartile range) ORs of significant positive survival coefficients were 3.1 (2.3-3.9). Only two coefficients were negative and significant, both involving associations of substance disorders with later agoraphobia.

The best-fitting model

The best-fitting multivariate model pooled across all outcomes was one that predicted first onset of each disorder from dummy variables for prior disorders plus summary measures of number of prior disorders. (Detailed results of model-fitting are available on request.) The number-of-disorders measures include dummy variables for exactly two to seven+ disorders.

Coefficients associated with pure disorders

The disorder type coefficients in the best-fitting model represent predictive associations of disorders that occur to people with no prior disorders. As there are a large number of coefficients for these pure disorders, it is useful to focus on summary statistics. (Detailed results are available on request.) 40.0% of pure-disorder coefficients are positive and statistically significant. (Table 2) Median (inter-quartile range) ORs are 3.1 (2.4-3.9). Only one of 210 coefficients is negative and significant. Roughly two-thirds (63.2%) of within-class coefficients are significant versus 33.0% of between-class coefficients. The median (inter-quartile range) significant within-class ORs [3.9 (3.0-4.8)] are higher than comparable between-class ORs [2.9 (2.4-3.5)].

Table 2. Predictive associations (percent significant and range of significant odds-ratios) of temporally primary pure lifetime estimated DSM-IV disorder types with the subsequent first onset of other estimated DSM-IV disorders based on the best-fitting multivariate survival model (n = 6,483)^a.

	Predictor disorders
Outcome disorders	Fear disorders		Distress disorders		Behavior Disorders		Substance disorders		Bipolar disorder		All disorders
	%g	ORsh	%g	ORsh	%g	ORsh	%g	ORsh	%g	ORsh	%g	ORsh
I. Fear disorders
Specific phobia	33.3	3.9	50.0	1.8-3.2	0	-	0	-	0	-	21.4	1.8-3.9
Agoraphobiab	66.7	4.8-9.2	50.0	2.8-3.3	0	-	0	-	0	-	28.6	2.8-9.2
Social phobia	66.7	3.6-4.2	50.0	1.8-2.1	0	-	0	-	0	-	30.8	1.8-4.2
Panic disorderc	33.3	3.3	25.0	3.1	0	-	0	-	0	-	14.3	3.1-3.3
II. Distress disorders
Separation anxiety disorder	50.0	2.4-2.7	66.7	2.4-4.7	0	-	0	-	0	-	28.6	2.4-4.7
Post-traumatic stress disorder	50.0	3.0-3.9	66.7	3.7-4.7	0	-	50.0	4.7	100.0	3.2	42.9	3.0-4.7
Major depressive episode/dysthymia	100.0	1.7-3.3	33.3	3.4	100.0	2.3-3.4	50.0	4.4	0	-	71.4	1.7-4.4
Generalized anxiety disorder	50.0	2.8-3.1	100.0	2.5-5.1	0	-	50.0	4.3	0	-	42.9	2.5-5.1
III. Behavior disorders
ADHD	0	-	0	-	66.7	3.7-5.6	0	-	0	-	20.0	3.7-5.6
Oppositional defiant disorder	75.0	1.9-2.9	50.0	2.0-3.6	66.7	2.3-4.9	0	-	0	-	50.0	1.9-4.9
Conduct disorder	50.0	1.8-3.4	25.0	4.4	66.7	2.3-3.0	50.0	10.3	0	-	42.9	1.8-10.3
Eating disordersd	50.0	2.4-3.7	50.0	2.5-3.8	66.7	2.2-4.1	0	-	0	-	42.9	2.2-4.1
IV. Substance disorders
Alcohol abusee	25.0	2.6	25.0	3.1	50.0	2.8-3.5	100.0	8.3	0	-	35.7	2.6-8.3
Drug abusee	50.0	2.4	25.0	2.1	75.0	2.8-6.1	100.0	8.4	100.0	3.2	57.1	2.1-8.4
V. Other disorders
Bipolar disorderf	75.0	2.8-4.3	50.0	1.8-5.7	100.0	2.3-4.4	0		0		64.3	1.8-5.7

^aBased on a discrete-time (person-year) survival model in which first lifetime onset of each of the 15 estimated DSM-IV disorders is predicted by 14 dummy variables for prior lifetime history of the other disorders, dummy variables for number of prior lifetime disorders (2-7+), and socio-demographic controls (sex, race/ethnicity, region of the country, urbanicity, parent education, number of biological parents with whom the adolescent lives, birth order, number of siblings, and age).

^bAgoraphobia is assessed with or without panic disorder

^cPanic Disorder is assessed with or without agoraphobia

^dIncludes anorexia nervosa, bulimia nervosa, and binge-eating disorder

^eWith or without dependence

^fIncludes Bipolar-I, Bipolar-II, and sub-threshold bipolar disorder

^gPercent of coefficients that are significant at the .05 level using two-sides design-based significance tests

^hRange of significant ORs

The same general pattern holds within each class of disorders. 50.0-100% of within-class ORs are significant vs. 15.9-64.3% of between-class coefficients. The percentage of significant between-class coefficients is much lower predicting fear (15.9%) than other (29.6-64.3%) disorders. One or two disorders are the most powerful within-class predictors in each class: specific phobia and social phobia among the fear disorders; major depression for distress disorders; and attention-deficit/hyperactivity and conduct disorders for behavior disorders. Fear disorders have the highest proportion of significant between-class coefficients predicting other disorders (52.3%) followed by distress (36.4%), behavior (29.6%), and substance (15.4%) disorders. Two of the three most consistent predictors are fear disorders: social phobia (81.8%) and specific phobia (72.7%), the other being major depression (72.7%).

Coefficients associated with comorbidity

The coefficients associated with number of predictor disorders in the best-fitting model represent predictive associations of comorbidity expressed as deviations from the pure-disorder coefficients. For example, absent effects of comorbidity, respondents with three predictor disorders having pure-disorder ORs of, say, 1.3, 1.5, and 1.7 would have an expected OR of 3.3 (i.e., 1.3 × 1.5 × 1.7). If the actual OR for these respondents is 4.0, then the number-of-disorders OR would be 1.2 (i.e., 4.0/3.3), indicating that the predictive effect of comorbidity is 20% higher than expected from the pure-disorder coefficients. The best-fitting model assumes that these number-of-disorders coefficients are a function of each respondent’s weighted number of predictor disorders, with weights defined by pure-disorder coefficients.

With this interpretation in mind, 87.2% of the number-of-disorder coefficients in the best-fitting model are less than 1.0 across equations, indicating a general pattern of sub-additive interaction in the logistic specification; that is, a pattern in which the joint effects of the interacting predictors are significantly less than those estimated in a model that assumes that no interactions exist. (Table 3) One-third of these sub-additive coefficients are statistically significant, while none of the ORs greater than 1.0 is significant. This pattern of the predictive associations of comorbidity generally being less than the product of their parts becomes stronger as the number of disorders in the comorbid profile increases from a median OR of 0.8 for two disorders to 0.1 for 7+ disorders.

Table 3. Predictive associations (odds-ratios) along with 95% confidence intervals (95% CI) of number of lifetime estimated DSM-IV disorders with the subsequent first onset of other estimated DSM-IV disorders based on the best-fitting multivariate survival model (n = 6,483)^a.

	Number of predictor disorders
	2		3		4		5		6		7+
Outcome disorders	OR	(95% CI)	OR	(95% CI)	OR	(95% CI)	OR	(95% CI)	OR	(95% CI)	OR	(95% CI)
I. Fear disorders
Specific phobia	1.0	0.5-2.0	0.3	0.1-1.1	0.9	0.2-4.9	0.7	0.1-7.0	1.2	0.1-11.4	3.6	0.3-45.5
Agoraphobiab	0.6	0.2-1.7	0.4	0.0-2.4	0.4	0.0-3.8	0.1	0.0-2.0	0.0*	0.0-0.8	0.0	0.0-6.4
Social phobia	0.8	0.4-1.9	0.6	0.2-2.0	0.4	0.1-2.7	0.1	0.0-1.5	0.1*	0.0-0.7	1.0	0.0-65.5
Panic disorderc	0.9	0.4-2.2	1.1	0.2-6.7	0.4	0.0-7.6	0.3	0.0-6.4	0.4	0.0-14.1	0.0	0.0-5.1
II. Distress disorders
Separation anxiety disorder	0.7	0.3-1.6	1.0	0.3-3.2	0.9	0.2-4.7	0.7	0.1-5.4	0.0*	0.0-0.2	1.8	0.1-51.6
Post-traumatic stress disorder	0.7	0.3-1.6	0.6	0.1-2.3	0.1*	0.0-0.3	0.1*	0.0-0.7	0.1	0.0-1.7	0.0*	0.0-0.4
Major depressive disorder/dysthymia	0.8	0.4-1.3	0.4*	0.2-0.8	0.2*	0.1-0.6	0.1*	0.0-0.2	0.1*	0.0-0.4	-	-
Generalized anxiety disorder	1.0	0-.33.6	0.9	0.1-5.5	0.7	0.1-9.0	0.2	0.0-3.5	0.2	0.0-8.8	0.0	0.0-28.6
III. Behavior disorders
Attention-deficit/hyperactivity disorder	1.2	0.3-5.1	3.0	0.3-30.2	0.6	0.0-15.3
Oppositional defiant disorder	0.8	0.4-1.6	0.6	0.3-1.2	0.2*	0.0-0.9	0.1*	0.0-0.8	0.0*	0.0-0.4	0.0*	0.0-0.5
Conduct disorder	1.0	0.5-2.4	0.4	0.1-1.2	0.4	0.6-2.0	0.2	0.0-1.3	0.1	0.0-1.0	0.2	0.0-9.6
Eating disordersd	0.8	0.3-2.1	0.3*	0.1-0.8	0.2	0.0-1.2	0.2	0.0-1.4	0.1	0.0-1.3	0.0*	0.0-0.04
IV. Substance disorders
Alcohol abusee	0.6	0.2-1.8	0.6	0.1-2.8	0.3	0.0-2.2	0.1*	0.0-0.9	0.0*	0.0-0.4	0.0	0.0-1.6
Drug abusee	0.9	0.4-1.9	0.4	0.2-1.3	0.2*	0.0-1.0	0.1*	0.0-1.0	0.3	0.0-4.6	0.0*	0.0-0.4
V. Other disorders
Bipolar disorderf	0.4*	0.2-0.8	0.2*	0.1-0.6	0.2	0.0-0.7	0.0*	0.0-0.3	0.0*	0.0-0.02	0.0*	0.0-0.02

^*Significant at the .05-level, two-sided test

^aBased on a discrete-time (person-year) survival model in which first lifetime onset of each of the 15 estimated DSM-IV disorders is predicted by 14 dummy variables for prior lifetime history of the other disorders, dummy variables for number of prior lifetime disorders (2-7+), and socio-demographic controls (sex, race/ethnicity, region of the country, urbanicity, parent education, number of biological parents with whom the adolescent lives, birth order, number of siblings, and age).

^bAgoraphobia is assessed with or without panic disorder

^cPanic Disorder is assessed with or without agoraphobia

^dIncludes anorexia nervosa, bulimia nervosa, and binge-eating disorder

^eWith or without dependence

^fincludes Bipolar-I, Bipolar-II, and sub-threshold bipolar disorder

Population attributable risk proportions

PARPs vary widely across outcomes (inter-quartile range: 35.6-54.5%). (Table 4) The lowest PARPs in terms of the outcomes are associated with specific phobia (3.7%) and attention-deficit/hyperactivity disorder (6.9%), the two earliest-onset disorders. This reflects the low prevalence and generally insignificant association of prior disorders predicting these two disorders. A similar interpretation applies to separation anxiety disorder (SAD), the outcome disorder with the next lowest risk proportion (23.1%). SAD has a comparatively early AOO distribution and is predicted by a lower than average proportion of prior disorders. The other 12 disorders, when considered as outcomes, all have risk proportions of 35.0% or higher. Those with the highest risk proportions have comparatively late AOO distributions and are either significantly predicted by the majority of other disorders (alcohol abuse, bipolar disorder), very strongly predicted by a smaller number of other disorders (panic disorder, agoraphobia), or are less strongly predicted by highly prevalent disorders (generalized anxiety disorder).

Table 4. Population attributable risk proportions (PARPs) of temporally primary lifetime estimated DSM-IV disorder types predicting subsequent first onset of other estimated DSM-IV disorders based on the best-fitting multivariate survival model (n = 6,483)^a.

	Types of predictor disorders
Outcome Disorder	Fear disorders	Distress Disorders	Behavior disorders	Substance disorders	Bipolar disorder	All disorders
1. Fear disorders
Specific phobia	1.7	1.3	1.2	0.5	0.5	3.7
Agoraphobiab	63.3	23.2	− 18.8	−0.0	1.0	62.2
Social phobia	29.0	5.9	0.4	−0.2	0.2	35.6
Panic Disorderc	28.2	15.5	1.5	0.0	2.1	41.5
II. Distress disorders
Separation Anxiety disorder	19.4	7.0	0.9	0.6	0.7	23.1
Post-traumatic Stress disorder	38.2	27.0	11.4	3.8	8.1	54.5
Major depression/dysthymia	20.5	4.6	18.7	0.2	−0.0	37.9
Generalized anxiety disorder	43.0	40.6	7.4	2.8	−4.8	64.3
III. Behavior disorders
Attention deficit/hyperactivity disorder	3.0	1.1	3.6	0.0	0.8	6.9
Oppositional defiant disorder	12.0	13.3	19.0	−0.2	2.3	36.4
Conduct disorder	16.0	12.3	17.4	3.7	0.7	40.4
Eating disordersd	28.4	14.8	16.2	0.5	2.1	43.6
IV. Substance disorders
Alcohol abusee	11.4	11.2	27.4	11.1	3.8	48.8
Drug abusee	15.7	9.6	29.5	7.1	7.0	42.7
V. Other
Bipolar disorderf	32.2	25.1	19.0	0.0	--	57.2

^aBased on a discrete-time (person-year) survival model in which first lifetime onset of each of the 15 DSM-IV/CIDI disorders is predicted by 14 dummy variables for prior lifetime history of the other disorders, dummy variables for number of prior lifetime disorders (2-7+), and socio-demographic controls (sex, race/ethnicity, region of the country, urbanicity, parent education, number of biological parents with whom the adolescent lives, birth order, number of siblings, and age).

^bAgoraphobia is assessed with or without panic disorder

^cPanic Disorder is assessed with or without agoraphobia

^dIncludes anorexia nervosa, bulimia nervosa, and binge-eating disorder

^eWith or without dependence

^fIncludes Bipolar-I, Bipolar-II and sub-threshold bipolar disorder.

Focusing on predictor disorders, the fear disorders have the highest PARPs in predicting nearly three-fourths of all outcome disorders, including other fear disorders, all four distress disorders, eating disorder, and bipolar disorder. There are only five exceptions: attention-deficit/hyperactivity disorder, where, as noted above, the risk proportion is very low overall; oppositional defiant and conduct disorders, where the highest risk proportions are associated with other behavior disorders; and substance disorders, where the highest risk proportions are associated with behavior disorders. As predictors, substance disorders consistently have the lowest risk proportions. Distress disorders generally have higher proportions than behavior disorders in predicting fear and distress disorders, but lower in predicting behavior and substance disorders.

DISCUSSION

It is important to recognize that the analyses reported here focused on predictors of first onset of comorbid conditions. The investigation of persistence is a separate matter that requires additional analysis not undertaken here. Furthermore, with regard to onset the analysis examined only aggregate associations and did not consider the possibility of variation in the structure or predictors of comorbidity in childhood versus adolescence, among boys versus girls, or by other potentially important sub-grouping distinctions.

Three sample limitations are noteworthy: that the school-level response rate was quite low, the individual-level response rate relatively low, and the sample excluded adolescents not enrolled in school. Methodological analysis reported elsewhere reduces concern about the first limitation, as no evidence of bias was found due to school replacement (Kessler et al., 2009a). The finding in previous methodological studies that non-respondents have higher rates of mental illness than respondents implies that the second limitation probably led prevalence estimates to be conservative, although estimates of predictive associations might be biased either upward or downward and differentially across predictor disorders (Kessler et al., 1995). The third limitation reduces the external validity of findings.

Two limitations concerning measurement are also noteworthy: that diagnoses were based on lay interviews and questionnaires rather than clinical assessments and that lifetime diagnoses and AOO reports were based on retrospective recall rather than prospective data. Concern about the first limitations is somewhat reduced by the good concordance with clinical diagnoses (Kessler et al., 2009c). The second limitation presumably reduced lifetime prevalence estimates (Moffitt et al., 2010; Patten, 2009), distorted AOO reports (Simon & von Korff, 1995), and could have biased estimates of predictive associations differentially across disorders depending on between-disorder differences in failure to recall lifetime occurrence and/or AOO. It would be very valuable to correct these limitations by replicating the complex analyses carried out here with long-term longitudinal data.

As a number of prospective studies of child-adolescent mental disorders exist (Costello et al., 2003; Newman et al., 1996; Pine et al., 1998; Reinke & Ostrander, 2008), a question can be raised why retrospective analysis of the sort reported here has value. The answer is that most existing prospective studies, with a small number of notable exceptions (Wittchen et al., 1998; Olino et al., 2010), did not obtain data on lifetime prevalence of mental disorders, making it impossible to carry out analyses of the associations between prior lifetime disorders and subsequent first lifetime onset of secondary disorders in the majority of such studies. Replication of the kinds of analyses reported here should be encouraged in the prospective studies that would support them. The results of the current analyses should be useful in providing preliminary hypotheses to be investigated in these prospective studies. This has traditionally been a major value of retrospective studies (Mantel & Haenszel, 1988).

In the context of these limitations, the differential AOO patterns found here are consistent with considerable previous evidence that attention-deficit/hyperactivity disorder, specific phobia, and separation anxiety disorder have the earliest ages-of-onset and that generalized anxiety, depressive, and substance disorders have the latest ages-of-onset of the adolescent disorders considered here (Kessler et al., 2007). Very similar patterns were found in parallel analyses we previously carried out in a cross-national sample of adults (Kessler et al., 2011). The high comorbidity found among disorders is also consistent with previous evidence (Angold et al., 1999).

The four-factor structure found in the NCS-A is also generally similar to findings from previous studies (Kessler et al., 2009b; Krueger, 1999; Vollebergh et al., 2001; Watson, 2005). Although a number of different specifications of the factor analysis model (e.g., at the person-level vs. person-year level, with and without sub-threshold bipolar disorder, with and without corrections for differential validity of diagnoses) yielded very similar results, replication in independent datasets is needed before considering this pattern stable.

One exception to the consistency of NCS-A results with previous factor analyses is that separation anxiety disorder (SAD) loaded with distress disorders in the NCS-A but with fear disorders some previous studies (Lahey et al., 2004). It is unclear why this occurred, but it is noteworthy that the loading of SAD on the distress factor was the weakest among all distress disorders and that SAD had the largest cross-loading with the fear factor of any distress disorder.

Another discrepancy with previous factor analysis studies is that substance disorders loaded separately from behavior disorders in the NCS-A. Substance and behavior disorders have generally loaded together in previous studies. However, this might reflect the fact that most previous factor analysis studies did not include as many behavior disorders as the NCS-A. The fact that substance and behavior disorders loaded separately in the NCS-A suggests that they have unique underlying psychopathological processes in adolescence (Krueger, 1999), a possobility that is indirectly consistent with the finding in behavior genetic studies that substance disorders have unique genetic loadings not shared with behavior disorders (Kendler et al., 2003) that vary in importance across the life course (Kendler et al., 2008).

A final factor analysis result that warrants comment is that the summary measure of eating disorders in the NCS-A loaded with behavior disorders rather than with distress disorders. Previous research has generally found that anorexia nervosa is more strongly related than bulimia nervosa or binge-eating disorder to internalizing disorders (e.g., Godart et al., 2006), while bulimia nervosa and binge-eating disorder are strongly related to behavioral disorders (e.g., Marmorstein et al., 2007). It is relevant in this regard that the vast majority of NCS-A respondents with eating disorders had bulimia nervosa or binge-eating disorder rather than anorexia nervosa.

The finding of stronger within-class than between-class predictive associations is consistent with previous longitudinal studies of prevalent disorders among youth (Costello et al., 2003; Newman et al., 1996; Reinke & Ostrander, 2008), although we are unaware of previous comparable studies of first onset of lifetime disorders. Comparable results were also found in our previous cross-national study of the development of comorbidity among adults (Kessler et al., 2011). The finding that fear disorders are the most important between-class predictors is consistent with, but goes beyond, previous research showing that early-onset anxiety disorders predict later distress disorders (Costello et al., 2003; Lewinsohn et al., 1995; Pine et al., 1998; Stein et al., 2001). Substance use disorders, at the other extreme, generally do not predict later disorders but are predicted strongly by earlier disorders. Although associations of behavior disorders with later substance disorders are well-documented (Costello et al., 1999; Fergusson et al., 2007), the predictive associations of primary fear and distress disorders have been less consistently examined (Costello et al., 2003).

The asymmetric associations across time-lagged disorder pairs could be due either to unmeasured common causes (e.g., genetic influences) or to causal influences of the predictor disorders on outcome disorders. There is no definitive way to distinguish these two possibilities with either retrospective or prospective non-experimental data. To the extent that the primary disorders are causes, though, successful early intervention might help prevent onset of subsequent comorbid disorders. To the extent that predictive associations are due to underlying common causes, temporally primary disorders might be useful risk markers (Kraemer et al., 1997) in targeting indicated interventions. Fear disorders stand out as especially important in the latter regard because they are such consistently strong predictors of later disorders. It is noteworthy in this regard that childhood-onset specific phobias can often be treated very effectively with inexpensive exposure-based therapies (Hamm, 2009; Gros & Anthony, 2006).

We are aware of no previous broad-based analysis of lifetime comorbidity predicting the subsequent first onset of a range of secondary disorders other than our own prior cross-national cross-section study of adults (Kessler et al., 2011). Our finding of largely sub-additive predictive associations of comorbid disorders, which we also found in our adult study, is consequently unique. It is important to note that this sub-additivity is based on a logistic model, though, which means that an additive specification with a linear or other link function might fit the data. However, our efforts to find an alternative link function that provided a better fit of the data with an additive specification was unsuccessful. This sub-additive pattern has potentially important implications for intervention because it suggests that intervening with a single comorbid disorder might not be effective in preventing subsequent disorder onset, as incremental predictive effects of individual disorders decrease with number of other disorders. To the extent that primary disorders are causes, treatment of the entire cluster of comorbid disorders is likely to be more effective than treatment of individual primary disorders in secondary prevention among patients with comorbidity (Moses & Barlow, 2006). Perhaps even more important, though, is that we showed that the innovative methodology used here yields substantively plausible results. Although many of these results do little more than replicate previous findings, the methodology has the potential to yield much more nuanced results when applied, as we hope it will be in the future, to prospective datasets.