Psychometric Properties and Factor Structures of the Korean Version of Children's Depression Rating Scale-Revised

Abstract

Objective:

The Children's Depression Rating Scale-Revised (CDRS-R) is a valid instrument for the assessment of depressive symptoms in youth, but this measure is yet to be validated in Korea. Thus, the present study aimed to evaluate the validity of the Korean version of the CDRS-R and to determine its factor structures.

Methods:

This study included 66 youths between 12 and 17 years of age who participated as part of an ongoing study investigating biomarkers of the antidepressant response and suicidal events in depressed youth. At baseline, the participants were assessed using the Kiddie-Schedule for Affective Disorders and Schizophrenia-Present and Lifetime version (K-SADS-PL), CDRS-R, Children's Depression Inventory (CDI), Beck Depression Inventory (BDI), Clinical Global Impressions-Severity (CGI-S), Children's Global Assessment Scale (CGAS), and Child Behavior Checklist (CBCL). Based on their diagnosis, each participant was assessed either five (major depressive disorder [MDD], n = 36) or two (non-MDD, n = 30) more times with the CDRS-R, CDI, BDI, CGI-S, and CGAS; thus, 306 assessments were included in the final analyses of the present study. Pearson's r and Cronbach's α values were used to determine validity and internal consistency, and exploratory factor analysis was performed to determine the factor structures of the Korean version of the CDRS-R.

Results:

The CDRS-R score was significantly correlated with scores on the K-SADS-PL depression subscale (r = 0.93, p < 0.01), CDI (r = 0.86, p < 0.01), BDI (r = 0.85, p < 0.01), CGI-S (r = 0.84, p < 0.01), and CGAS (r = −0.86, p < 0.01), indicating that this measure has strong criterion validity. In addition, the high correlation of the CDRS-R with the depression subscale of the CBCL (r = 0.67, p < 0.01) and lower correlations with the subscales of the CBCL that assess delinquent behavior, aggressive behavior, and externalizing problem support the discriminant validity of this instrument (r _s ≤ 0.50, p < 0.01). Internal consistency was high, as evidenced by a Cronbach's α of 0.91, and the exploratory factor analysis revealed that the Korean version of the CDRS-R comprised three factors as follows: subjective depressed mood, daily functional impairment, and observed depressive affect. These factors differed from those used in previous studies, which were performed with the English version of the CDRS-R.

Conclusions:

This study of the Korean version of the CDRS-R provides initial promising data regarding its criterion validity, discriminant validity, internal consistency, and factor structures. These properties were significantly strong, which suggests that the Korean version of the CDRS-R is a valid and reliable instrument for the assessment of depressive symptoms in youth.

Introduction

Major depressive disorder (MDD) is a common and debilitating disorder that is a major cause of morbidity and mortality during adolescence. Even though there are not many studies of MDD that have been conducted in Korea, the prevalence among Korean adolescents has been reported to be high. A Korean study using the Center for Epidemiologic Studies Depression Scale (CES-D) (Radolf 1977) found that 17.4% of boys and 20.6% of girls reported experiencing definite depressive mood (Cho et al. 2001). Another study of middle school students reported the prevalence of major depression to be 19.8% in girls and 11.1% in boys (Kim 2008). In Korean adolescents, depression often results in related problems, including suicidal ideation and attempts. Notably, suicidal ideation and attempts are 5.31 and 3.19 times more likely to occur, respectively, among Korean adolescents with high depression scores than among those with lower depression scores (Juon et al. 1994). Furthermore, suicide is the leading cause of death in Korean adolescents (Statistics Korea 2011).

The Korean Ministry of Education is attempting to screen for mental health and behavioral problems in seventh and tenth graders across the country every year to help those who are experiencing a variety of mental health problems, including depression (Min and Kang 2015). These efforts indicate that more adolescents with depressive symptoms will be referred to mental health service centers and hospitals for assessments and treatment. Due to the increasing number of adolescents with depression and the expected rise in the number of visits for precise assessments based on governmental screenings, mental health services should be equipped with valid and reliable assessment tools for more accurate diagnosis and proper treatment.

Several depression scales, including the Children's Depression Inventory (CDI) (Kovacs 1982), Beck Depression Inventory (BDI) (Beck and Ward 1981), and CES-D, are currently used in Korea, and the Korean versions of these instruments have been validated (Hahn et al. 1986; Cho and Lee 1990; Cho and Kim 1998). However, there is no validated interviewer rating scale for assessing depression in Korean adolescents. Symptoms of depressed adolescents have contextual variations, and perspectives of the adolescents and their parents on mental health symptoms may be different (Myers and Winters 2002; De Los Reyes et al. 2015; Youngstrom et al. 2015). These variations might cause considerable informant discrepancies of reports on the symptoms between the adolescents and their parents. Therefore, a rating scale that integrates reports from adolescents and parents is required to provide a more comprehensive picture of the youth's mental health.

The Children's Depression Rating Scale-Revised (CDRS-R), which was developed based on the Adult Hamilton Depression Scale, is one of the interviewer rating scales for which a Korean version is available. This measure comprises 17 items that are rated based on a clinician's observations and individual interviews with the children and their parents; this scale was previously used in several clinical studies in Korea and is also used for the assessment of depressive symptoms in Korean youth (Goo et al. 2013; U.S. National Institutes of Health 2015, 2016). However, despite the strong validity and reliability of the original English language version, the psychometric properties of the Korean version have yet to be examined.

The English version of the CDRS-R has been reported to consist of multifactorial structures. It was originally proposed, by Poznanski et al. (1984), to incorporate four factor structures as follows: mood, somatic, subjective, and behavior. Using an exploratory factor analysis, Guo et al. (2006) reported that the CDRS-R consists of five factor structures as follows: observed depressive mood, anhedonia, morbid thoughts, somatic symptoms, and reported depressive mood. Another study, which used a longitudinal design and a youth sample, suggested that the CDRS-R consists of two factor structures as follows: clinician-observed signs and reported symptoms of depression (Isa et al. 2014). Therefore, it can be inferred that the Korean version of the CDRS-R also consists of multifactor structures.

Cross-cultural studies have reported similarities and differences in the profiles of depressive symptoms in different populations. For instance, a study conducted in 10 countries, including Korea, found different prevalence rates (low rates in Korea and Taiwan), but similar patterns of symptoms in depressed patients (Weissman et al. 1996). A comparative study of depression assessing American and Korean college students revealed similar experiences in depressed patients in terms of psychological affect (Yoo and Skovholt 2001).

In contrast, some researchers have reported cultural differences in the symptom profiles of depressed patients. While Kleinman and Good (2004) reported that depressed patients in East Asia experience more physical symptoms than psychological symptoms, Chang et al. (2007) reported that, in comparison to Western people, Koreans with depression are more likely to complain of appetite change, but less likely to complain about feelings of worthlessness or guilt. In addition, previous studies have reported that, relative to Western cultures, the South Korean culture is collectivistic and emphasizes group awareness over individual interests (Schwartz et al. 2002). Furthermore, due to the cultural emphasis on academic achievement in South Korea, school performance is strongly correlated with acceptance by one's peer group (Schwartz et al. 2002).

Exploration of the factor structures of the Korean version of the CDRS-R and comparisons between them and those found in other studies performed using the English version of the CDRS-R might indicate differences in the presentation of depression between Korean- and English-speaking youth.

Thus, the present study sought to determine the validity and factor structure of the Korean version of the CDRS-R. It was expected that the CDRS-R would be a valid and reliable instrument for the assessment of depressive symptoms in Korean youth. In addition, considering the distinct profiles of depressive symptoms among various countries, as reported in the literature, it was hypothesized that the Korean version of the CDRS-R would have a different factor structure than that reported for the English version of the CDRS-R.

Methods

Participants

This study included 66 youths between 12 and 17 years of age who participated as part of an ongoing study investigating biomarkers of the antidepressant response and suicidal events in depressed youth (NRF-2015R1A2A2A01004501). The study sample consisted of 26 boys (39.4%) and 40 girls (60.6%), who were divided into patient (n = 36; 54.5%) and control (n = 30; 45.5%) groups based on their diagnoses using the Kiddie-Schedule for Affective Disorders and Schizophrenia-Present and Lifetime version (K-SADS-PL) (Kaufman et al. 1997; Kim et al. 2004).

Participants were excluded from the study if their intelligence quotient (IQ) was <70, if they had any chronic medical diseases, a history of psychotic disorders such as schizophrenia or bipolar disorder, a history of eating disorder, or if they had been diagnosed with any developmental disorders such as autism, had abused alcohol or other substances within the past 6 months, had any neurological or physical diseases, had first degree relatives with a history of bipolar I disorder, or were currently taking any psychiatric medications (excepting treatments for ADHD). Control participants were excluded from this study if they had a history of any psychiatric disorder or first degree relatives with a history of any psychiatric disorder.

The present study was approved by the Institutional Review Board for Human Subjects at the Seoul National University Hospital. Detailed information about the study was provided to the parents and youth, and written informed consent was obtained before study entry.

Assessments

At baseline, all participants were assessed using the K-SADS-PL, CDRS-R, CDI, BDI, Clinical Global Impressions-Severity (CGI-S) (Guy 1976), Children's Global Assessment Scale (CGAS) (Shaffer et al. 1983), and Child Behavior Checklist (CBCL) (Achenbach and Edelbrock 1983). Based on their diagnoses, participants were assessed either three (control group: at baseline and weeks 2 and 8) or six (patient group: at baseline and at weeks 1, 2, 4, 6, and 8) times with the CDRS-R, CDI, BDI, CGI, and CGAS. At the study exit, all participants were assessed using the CDRS-R, CDI, BDI, CGI, CGAS, and CBCL.

The CDRS-R was administered by two interviewers (S.N. and A.-H.J.) who were blind to participants' CDI, BDI, CGI, and CGAS scores during each assessment (intraclass correlation coefficient = 0.96, 95% confidence interval = 0.89–0.98), and the K-SADS-PL, CGI, and CGAS were administered by a psychiatrist (J.-W.K.) who was blind to the CDI and BDI scores during each assessment.

Our data were derived from a treatment study and included multiple assessments of the same participants. This method has been used in previous studies about the psychometric properties and subscales of the CDRS-R (Mayes et al. 2010; Isa et al. 2014). Therefore, we analyzed the validity and reliability of the scores with baseline data only and with all available data. When using all available data, a total of 306 assessments with the CDRS-R were included in the analyses.

To investigate the correlation between the CDRS-R and the K-SADS-PL depression subscale, scores on the K-SADS-PL depression subscale were calculated by summing scores on several items, including depressive mood, excessive or inappropriate guilt, anhedonia/lack of interest/low motivation, fatigue/lack of energy/tiredness, difficulty concentrating/inattention, psychomotor agitation, psychomotor retardation, insomnia, hypersomnia, lack of appetite, increased appetite, and suicidal ideation/attempts (Birmaher et al. 2010).

Statistical analyses

Demographic variables were analyzed with independent t-tests or Chi-square tests, and paired t-tests were used to compare the baseline and exit scores on each scale. To examine the criterion, convergent, and discriminant validity of the Korean version of the CDRS-R, Pearson's correlation coefficients measuring the relation of CDRS-R scores with scores on the K-SADS-PL, CDI, BDI, CGI-S, and CGAS and subscales of the CBCL were calculated.

The diagnostic efficiencies of the self-rating scale (CDI), parent rating scale (anxious/depressed subscale of CBCL), and CDRS-R were compared using receiver operating characteristics (ROC) analyses. The ROC analyses were performed with baseline assessments only. The areas under the ROC curves (AUCs) of the three scales were compared statistically using a nonparametric approach (DeLong et al. 1988). Item–total correlations (r _it) and Cronbach's α values were generated to determine the reliability and consistency of the CDRS-R.

Our baseline sample size satisfied the minimum required sample size for a factor analysis (Mundfrom et al. 2005), and an exploratory factor analysis was performed to investigate the factor structures of the Korean version of the CDRS-R. Maximum likelihood, which provides better parameter estimates than does principal component analysis, was used to extract the factors (Guo et al. 2006), and a scree test (Cattell 1966) was used to decide the number of factors; factors with at least 5% common variance were considered. Finally, the number of factors was decided considering proper clinical interpretability; a Promax rotation, which allows for correlated factors, was performed to determine the factor structures (Cureton and Mulaik 1975). SPSS version 22.0 for Windows was used to analyze all data.

Results

Demographics and clinical characteristics

There were no significant differences between the patient and control groups in terms of age or IQ (Table 1). However, there were significant differences between the two groups in baseline scores on the K-SADS-PL depression subscale, CDRS-R, CDI, BDI, CGI-S, and CGAS. In addition, the scores on the CDRS-R, CDI, BDI, CGI-S, CGAS, and most of the subscales of the CBCL differed significantly between baseline and exit in the patient group. In contrast, the control group exhibited significant differences on the CDI and BDI scales only between the baseline and exit results.

Table 1.

Demographics and Clinical Characteristics

	Patient (n = 36)	Control (n = 30)	t or χ2^*	p ^*
Gender, n (%)
Male	11 (30.6)	15 (50.0)	2.59	0.10
Female	25 (69.4)	15 (50.0)
Age, mean (SD)	14.61 (1.64)	14.67 (1.40)	−0.15	0.88
IQ, mean (SD)	104.67 (14.83)	108.08 (12.99)	−0.99	0.33

	Baseline	Exit	t^**	p^**	Baseline	Exit	t^***	p^***
K-SADS-PL depression section, 12 items	25.67 (3.09)				12.13 (0.51)				25.87	0.00
CDRS-R	57.83 (12.19)	43.36 (13.09)	6.76	0.00	23.27 (3.47)	23.00 (4.63)	0.32	0.75	16.24	0.00
CDI	25.58 (12.56)	18.67 (15.58)	3.30	0.00	6.20 (5.40)	3.77 (4.68)	2.80	0.01	8.37	0.00
BDI	26.53 (16.06)	17.53 (19.06)	3.47	0.00	3.77 (4.55)	2.23 (3.05)	3.01	0.01	8.12	0.00
CGI-S	4.97 (0.56)	3.81 (0.79)	9.04	0.00	1.00 (0.00)	1.00 (0.00)			42.57	0.00
CGAS	45.83 (6.55)	60.06 (11.45)	−8.43	0.00	90.23 (5.01)	89.53 (3.80)	0.77	0.44	−30.43	0.00
CBCL
Withdrawn	67.94 (12.34)	61.17 (15.48)	3.83	0.00	50.00 (11.34)	48.53 (9.82)	1.01	0.32	6.10	0.00
Anxious/depressed	64.81 (11.39)	54.78 (15.33)	6.28	0.00	45.67 (12.89)	45.50 (11.38)	0.11	0.92	6.40	0.00
Delinquent behavior	53.69 (11.19)	50.50 (11.91)	2.07	0.05	43.80 (6.67)	42.73 (5.84)	1.15	0.26	4.25	0.00
Aggressive behavior	53.31 (10.24)	48.44 (14.37)	2.52	0.02	41.43 (9.25)	40.67 (8.87)	0.87	0.39	4.90	0.00
Internalizing problem	65.86 (9.59)	57.11 (14.23)	5.16	0.00	46.60 (12.47)	44.83 (11.54)	0.97	0.34	7.09	0.00
Externalizing problem	53.22 (10.88)	48.19 (14.63)	2.80	0.01	41.17 (9.01)	40.07 (8.84)	1.11	0.28	4.84	0.00

Differences of baseline between patient and control.

Differences between baseline and exit of patient.

***

Differences between baseline and exit of control.

BDI, Beck Depression Inventory; CBCL, Child Behavior Checklist; CDI, Children's Depression Inventory; CDRS-R, Children's Depression Rating Scale-Revised; CGAS, Children's Global Assessment Scale; CGI-S, Clinical Global Impressions-Severity; IQ, intelligence quotient; K-SADS-PL, Kiddie-Schedule for Affective Disorders and Schizophrenia-Present and Lifetime version.

Criterion validity

The correlation between the CDRS-R and K-SADS-PL depression subscale at baseline was significant (r = 0.93, p < 0.01; Supplementary Table S1; Supplementary Data are available online at www.liebertpub.com/cap), as well as the correlations between the CDRS-R and the internalizing subscales of the CBCL, including the withdrawn, anxious/depressed, and internalizing problem sections, at baseline (r _s ≥ 0.65; Supplementary Table S1). There were significant correlations between scores on the CDRS-R and scores on the CDI, BDI, CGI-S, and CGAS at baseline (r _s ≥ 0.86, p _s < 0.01). Furthermore, the correlations between scores on the CDRS-R and scores on the CDI, BDI, CGI-S, and CGAS at all assessments were also significantly high (Table 2), and the correlation of the mean change in scores on the CDRS-R from baseline to exit with CGI-S, CGAS, CDI, and BDI scores was significant (Supplementary Table S2).

Table 2.

Mean, Standard Deviation, and Correlations Among Children's Depression Rating Scale, Children's Depression Inventory, Beck Depression Inventory, Clinical Global Impressions, and Children's Global Assessment Scale at All Assessments (n = 306)

	CDRS-R	CDI	BDI	CGI-S	CGAS
CDRS-R	1
CDI	0.86^*	1
BDI	0.85^*	0.96^*	1
CGI-S	0.84^*	0.69^*	0.65^*	1
CGAS	−0.86^*	−0.74^*	−0.70^*	−0.96^*	1
Mean (SD)	40.19 (16.26)	16.34 (14.19)	15.20 (16.58)	3.25 (1.62)	64.24 (18.83)

p-Values <0.01.

BDI, Beck Depression Inventory; CDI, Children's Depression Inventory; CDRS-R, Children's Depression Rating Scale-Revised; CGAS, Children's Global Assessment Scale; CGI-S, Global Impressions-Severity; SD, standard deviation.

Discriminant validity

The CDRS-R showed relatively lower correlations with the externalizing subscales of the CBCL, including the delinquent behavior, aggressive behavior, and externalizing problem sections (r _s ≤ 0.50), compared with correlations with the internalizing subscales of the CBCL (r _s ≥ 0.65; Supplementary Table S1). Similar patterns were observed for the correlations between the mean change in CDRS-R score from baseline to exit and changes in the scores for each subscale of the CBCL (Supplementary Table S2). Although the correlations between the mean change in CDRS-R scores and changes in scores on the withdrawn, anxious/depressed, and internalizing problem subscales (r _s ≥ 0.37) were significant, those with changes in the delinquent behavior, aggressive behavior, and externalizing problem subscales (r _s ≤ 0.25) were low.

Correlations on the CDRS-R based on self-report scales and clinician-rated scales

The correlations between the scales completed by the same informant were very high (BDI and CDI: r = 0.96; CGI-S and CGAS: r = 0.96; Table 2). However, the correlations between the self-report scales (BDI and CDI) and the clinician-rated scales (CGI and CGAS) were relatively low (r _s ≤ 0.74). In particular, the correlations between the self-report scales and the parent-report scale (anxious/depressed subscales of the CBCL) at baseline were much lower (BDI and anxious/depressed: r = 0.58; CDI and anxious/depressed: 0.61).

ROC analyses and AUC comparisons of self-rating, parent rating, and interviewer rating scales

The ROC analyses indicated that a CDRS-R score of 40 was the optimal cutoff value, with a sensitivity of 0.97 and a specificity of 0.88. The AUCs of the CDI, anxious/depressed subscale of the CBCL, and CDRS-R were 0.88 (standard error [SE] = 0.05), 0.85 (SE = 0.05), and 0.97 (SE = 0.02), respectively. The pairwise comparisons of the AUCs among the three scales indicated that the AUC of the CDRS-R was significantly larger than those of the CDI (p = 0.01) and the anxious/depressed subscale of the CBCL (p = 0.001).

Internal consistency and item–total correlation

Internal consistency was good, as evidenced by a Cronbach's α of 0.94 and 0.91 with CDRS-R scores at baseline and for all assessments, respectively. Depressed feeling had the strongest item–total correlation (r _it = 0.89 for baseline and 0.80 for all assessments); item–total correlations ranged from 0.45 to 0.89 for baseline scores of the CDRS-R and from 0.30 to 0.80 for all assessments (Table 3).

Table 3.

Internal Consistency and Item–Total Correlations (n = 306)

Item	Mean (SD)	Corrected item–total correlation
Schoolwork	4.15 (2.23)	0.72
Capacity to have fun	2.9 (1.45)	0.63
Social withdrawal	2.26 (1.8)	0.58
Sleep	2.5 (1.50)	0.59
Appetite or eating	1.97 (1.23)	0.52
Excessive fatigue	3.23 (1.92)	0.67
Physical complaints	1.93 (1.33)	0.52
Irritability	2.44 (1.62)	0.74
Guilt	1.47 (0.98)	0.35
Self-esteem	4.03 (1.64)	0.76
Depressed feelings	2.52 (1.65)	0.80
Morbid ideation	1.3 (0.80)	0.30
Suicidal ideation	2.11 (1.79)	0.66
Weeping	1.71 (1.33)	0.54
Depressed affect	1.96 (1.15)	0.61
Tempo of speech	1.44 (0.65)	0.47
Hypoactivity	2.35 (1.17)	0.56
Total Cronbach' α	0.91

SD, standard deviation.

Factor structure

The Korean version of the CDRS-R comprised three factors, which were labeled subjective depressed mood, daily functional impairment, and observed depressive affect (Table 4).

Table 4.

Factor Loadings and Factor Internal Consistency (n = 306)

Factor	Item		Factor loading	Cronbach's α
Subjective depressed mood				0.83
	7	Physical complaints	0.50
	8	Irritability	0.67
	9	Guilt	0.61
	11	Depressed feelings	0.81
	12	Morbid ideation	0.33
	13	Suicidal ideation	0.81
	14	Weeping	0.72
Daily functional impairment				0.86
	1	School work	0.96
	2	Capacity to have fun	0.43
	3	Social withdrawal	0.58
	4	Sleep	0.36
	5	Appetite or Eating	0.26
	6	Excessive fatigue	0.52
	10	Self-esteem	0.51
	17	Hypoactivity	0.60
Observed depressive affect				0.78
	15	Depressed affect	0.89
	16	Tempo of speech	0.87

Subjective depressed mood consisted of seven items, and of these items, suicidal ideation and depressed feelings presented the highest factor loadings (0.81 and 0.81, respectively); the factor loadings for all items ranged from 0.33 to 0.81. Daily functional impairment consisted of eight items, with schoolwork presenting the highest factor loading (0.96). Although appetite, sleep, and self-esteem were included in the daily functional impairment factor, they showed factor loadings on two factors (Supplementary Table S3). Observed depressive affect consisted of two items, with factor loadings of 0.89 and 0.87. The Cronbach's α values were 0.83, 0.86, and 0.78 for the subjective depressed mood, daily functional impairment, and observed depressive affect factors, respectively.

Discussion

The present study examined the validity and reliability of the Korean version of the CDRS-R by comparing it with several measures of depression and other psychiatric symptoms. These measures used a range of modalities and a variety of informants and included the K-SADS-PL, CDI, BDI, CGI-S, CGAS, and CBCL. CDRS-R total scores at baseline were highly correlated with the K-SADS-PL depression scores. CDRS-R total scores at all assessments and the mean change in CDRS-R scores from baseline to exit were significantly correlated with scores on the CDI, BDI, CGI-S, and CGAS and with their changes from baseline to exit. These findings are consistent with those of a previous study that found significant relationships of CDRS-R total scores with diagnoses made using the K-SADS-PL (r = 0.64; p < 0.01) and scores on the CGI-S, CGI, and CGAS (Mayes et al. 2010). Thus, the Korean version of the CDRS-R has high criterion validity.

Moreover, although the correlations between the CDRS-R and the internalizing domain subscales of the CBCL at baseline were significantly high, those with the externalizing domain subscales were much lower. This finding indicates that the Korean version of the CDRS-R has high discriminant validity.

In the present study, the CDRS-R was administered by two interviewers whose evaluations were independent of clinician assessments using the CGI and CGAS. Although assessments of youth depression provided by multiple informants provide more detailed information in terms of diagnosis, the poor level of agreement among informants for childhood psychiatric disorders, including children, parents, and clinicians, has led to criticism (Grills and Ollendick 2003). For instance, reports of depressive symptoms in youth often vary because the respondents may either underreport their symptoms in a search for social acceptance or overreport their symptoms when they feel overwhelmed (Myers and Winters 2002). Furthermore, parents may report symptoms for internalizing disorders, such as major depression and anxiety, in their children much less than they report externalizing disorders because the symptoms in the latter case are much more visible (Cantwell et al. 1997).

The present findings were consistent with those of previous studies. The correlations between the rating scales completed by different informants were lower than the correlations between the rating scales completed by the same informants. This indicates that there are discrepancies between reports of the depression symptoms of adolescents. To examine the diagnostic efficiency of the CDRS-R, the present study compared the AUCs of the self-rating scale (CDI), parent rating scale (anxious/depressed subscale of CBCL), and CDRS-R using ROC analyses and pairwise comparisons of the AUCs. The diagnostic efficiency of the CDRS-R was significantly higher, with a higher AUC, than diagnostic efficiencies of the other scales.

Using exploratory factor analyses, the present study identified three factors of the Korean version of the CDRS-R and determined that the factor structure was both similar to and different from those identified in previous studies (Supplementary Table S4). Items such as depressed feelings, irritability, weeping, guilt, morbid ideation, and suicidal ideation were included in the subjective depressed mood factor in this study. Similarly, these items contributed to the mood and subjective factors identified by Poznanski et al. (1984) and the reported depressive mood and morbid thought factors identified by Guo et al. (2006).

In contrast, whereas physical complaints were included in somatic factors by Poznanski et al. (1984) and Guo et al. (2006), they were included in the subjective depressed mood factor in the present study. Somatic complaints are often experienced by children and adolescents as a result of their restricted emotional expression (Gupta Karkhanis and Winsler 2016). Compared with Western cultures, Asian cultures do not encourage emotional expression (Kim and Sherman 2007). It is possible that physical complaints were more likely to be associated with subjective mood due to the influences of these cultural factors.

Sleep, appetite, and fatigue were included in the daily functional impairment factor in the present study. They were included in the somatic factor in Poznanski's and Guo's (Supplementary Table S4). Furthermore, three additional items (schoolwork, capacity to have fun, and social withdrawal) were included together in the daily functional impairment. These items belonged to the behavior factor in the study by Poznanski et al. (1984) and to the somatic symptoms and anhedonia factors in the study by Guo et al. (2006).

Our findings indicate that schoolwork, the capacity to have fun, and social withdrawal are associated with vegetative symptoms, such as sleep disturbance, appetite change, and fatigue. In addition, schoolwork and social withdrawal might be categorized in the same factor due to South Korean culture. In South Korea, school performance is highly correlated with acceptance by one's peer group due to cultural emphasis on academic achievement (Schwartz et al. 2002). Similarly, it may be inferred that self-esteem was categorized together with schoolwork and social withdrawal because self-esteem in Korean adolescents may be highly influenced by schoolwork and one's peer group relationships; in contrast, self-esteem reflects the cognitive aspect of depression and was included in the subjective domain in the two previous studies.

In particular, schoolwork and self-esteem presented the highest severity scores over four points among all items (Table 3), as well as being highly correlated (r = 0.69, p < 0.01) in the present study, whereas irritability presented the highest score in the previous studies (Guo et al. 2006; Mayes et al. 2010).

Depressed affect and tempo of speech were originally intended to be assessed based on the observations of the interviewers and categorized as elements of observed depressive mood (Guo et al. 2006). Similarly, they were included in the clinician-observed sign factors in another study (Isa et al. 2014). In the present study, they were included in observed depressive affect, which was consistent with these previous findings.

Limitations

This study has several limitations that should be noted. The sample in the present study was not randomly recruited from among the general population, as it included depressed patients who were undergoing medical treatment; this may limit the generalizability of the results to the general population with depression. The present study differed from some others with regard to the factor structures of the CDRS-R. However, the samples and methods used in our study differed from those of the previous studies. For instance, the results reported by Poznanski et al. (1984) can be applied only to children. The study by Guo et al. (2006) included only one assessment point, and the age range of the participants was between 7 and 17 years.

Many studies that used factor analyses of other depression scales have indicated that the factor structure of a scale can differ among samples with various characteristics. Because the present sample was limited to depressed adolescents referred to a hospital and healthy controls, further studies using factor analyses with various samples are warranted to elaborate on the factor structure of the Korean version of the CDRS-R. In addition, the AUCs of the present study may need careful consideration in interpreting. Since these analyses included measures that are used to differentiate between depressed youth and healthy youth, the effect size may have been inflated (Youngstrom 2014; Youngstrom et al. 2015).

The small sample size is another limitation of our study. In addition, this study included repeated evaluations using the same depression scales for the same participants over relatively short-term intervals (1 or 2 weeks), which could cause a bias due to testing effects. However, when we analyzed the criterion validity and internal consistency with the baseline data only and with all the available data, respectively, we did not find any significant difference between the two analyses. In addition, the factor analysis with the baseline data indicated that the CDRS-R consists of three factors. The same result was obtained when the analysis included repeated measurements. Further studies with larger samples with more variety and without repeated measures are warranted to confirm the validity and factor structures of the Korean version of the CDRS-R.

Despite several limitations, our study has several clinical implications. We identified that, as a single scale, the Korean version of the CDRS-R is as valid, reliable, and diagnostically efficient for the assessment of depressed adolescents as the English version of the CDRS-R. In addition, the factor analysis indicated that the Korean version of the CDRS-R assesses various symptom domains of depressed Korean adolescents, including subjective depressed mood, daily functional impairment, and observed depressive affect.

Conclusions

The present data suggest that the Korean version of the CDRS-R is a reliable tool for the assessment of adolescent depression and that it has high criterion and discriminant validity for use with Korean adolescents. In addition, these findings suggest that the Korean version of the CDRS-R has a three-factor structure that includes subjective depressed mood, daily functional impairment, and observed depressive affect.

Clinical Significance

The use of valid and reliable tools for the assessment of depression in adolescents is important. Although several self-rating scales for adolescent depression, including Korean versions of the CDI, BDI, and CES-D, have been validated, no interviewer rating scales for adolescent depression have been validated in Korea. The present validation study of the Korean version of the CDRS-R provides Korean psychiatrists with a tool that allows for the precise assessment of adolescent depression. Furthermore, the investigation of its factor structure revealed that this measure has similarities and differences relative to the factor structure of the English version of the CDRS-R. These relationships imply that the characteristics of depression might differ between Korean and Western adolescents; thus, further studies investigating cross-cultural differences in adolescent depression are warranted.

Footnotes

Disclosures

No competing financial interests exist.

References

Achenbach

, Edelbrock

: Manual for the Child Behavior Checklist and Revised Behavior Profile. Burlington (Vermont), University of Vermont Department of Psychiatry, 1983.

Beck

, Ward

, Mendelson

, Mock

, Erbaugh

: An inventory for measuring depression. Arch Gen Psychiatry, 4:561–571, 1981.

Birmaher

, Axelson

, Goldstein

, Monk

, Kalas

, Obreja

, Hickey

, Iyengar

, Brent

, Shamseddeen

, Diler

, Kupfer

: Psychiatric disorders in preschool offspring of parents with bipolar disorder: The Pittsburgh Bipolar Offspring Study (BIOS). Am J Psychiatry, 167:321–330, 2010.

Cantwell

, Lewinsohn

, Rohde

, Seeley

: Correspondence between adolescent report and parent report of psychiatric diagnostic data. J Am Acad Child Adolesc Psychiatry, 36:610–619, 1997.

Cattell

: The scree test for the number of factors. Multivariate Behav Res, 1:245–276, 1966.

Chang

, Sohn

, Lee

, Choi

, Cho

, Jeon

, Hahm

, Lee

, Cho

: Characteristics of diagnostic criteria for depression in Korea. J Korean Neuropsychiatr Assoc, 46:447–452, 2007.

Cho

, Kim

: Use of the Center for Epidemiologic Studies Depression (CES-D) scale in Korea. J Nerv Ment Dis, 186: 304–310, 1998.

Cho

, Lee

: Development of the Korean form of the Children's Depression Inventory. J Korean Neuropsychiatr Assoc, 29:943–956, 1990.

Cho

, Jeon

, Kim

, Lyoo

, Cho

: Prevalence and correlates of depressive symptoms among the adolescents in an urban area in Korea. J Korean Neuropsychiatr Assoc, 40:627–639, 2001.

10.

Cureton

, Mulaik

: The weighted varimax rotation and the promax rotation. Psychometrika, 40:183–195, 1975.

11.

DeLong

, DeLong

, Clarke-Pearson

: Comparing the areas under two or more correlated receiver operating characteristic curves: A nonparametric approach. Biometrics, 44:837–845, 1988.

12.

De Los Reyes

, Augenstein

, Wang

, Thomas

, Drabick

, Burgers

, Rabinowitz

: The validity of the multi-informant approach to assessing child and adolescent mental health. Psychol Bull, 141:858–900, 2015.

13.

Goo

A-J

, Won

E-K

, Lee

J-I

, Jung

H-Y

, Kim

: Long-term tolerability of wscitalopram in Korean adolescents. J Korean Acad Child Adolesc Psychiatry, 24:207–212, 2013.

14.

Grills

, Ollendick

: Multiple informant agreement and the anxiety disorders interview schedule for parents and children. J Am Acad Child Adolesc Psychiatry, 42:30–40, 2003.

15.

Guo

, Nilsson

, Heiligenstein

, Wilson

, Emslie

: An exploratory factor analysis of the Children's Depression Rating Scale-Revised. J Child Adolesc Psychopharmacol, 16:482–491, 2006.

16.

Gupta Karkhanis

, Winsler

: Somatization in Children and Adolescents: Practical implications. J Indian Assoc Child Adolesc Ment Health, 12:79–115, 2016.

17.

Guy

: ECDEU assessment manual for psychopharmacology. Rockville (Maryland), U.S. Department of Health, Education, and Welfare; Public Health Service; Alcohol, Drug Abuse, and Mental Health Administration; National Institute of Mental Health, Psychopharmacology Research Branch, Division of Extramural Research Programs, 1976.

18.

Hahn

, Yum

, Shin

, Kim

, Yoon

, Chung

: Standardization study of Beck Depression Inventory in Korea. J Korean Neuropsychiatr Assoc, 25:487–502, 1986.

19.

Isa

, Bernstein

, Trivedi

, Mayes

, Kennard

, Emslie

: Childhood depression subscales using repeated sessions on Children's Depression Rating Scale-Revised (CDRS-R) scores. J Child Adolesc Psychopharmacol, 24:318–324, 2014.

20.

Juon

, Nam

, Ensminger

: Epidemiology of suicidal behavior among Korean adolescents. J Child Psychol Psychiatry, 35:663–676, 1994.

21.

Kaufman

, Birmaher

, Brent

, Rao

, Flynn

, Moreci

, Williamson

, Ryan

: Schedule for Affective Disorders and Schizophrenia for School-age Children-Present and Lifetime version (K-SADS-PL): Initial reliability and validity data. J Am Acad Child Adolesc Psychiatry, 36:980–988, 1997.

22.

Kim

, Sherman

: “Express yourself”: Culture and the effect of self-expression on choice. J Pers Soc Psychol, 92:1–11, 2007.

23.

Kim

: The prevalence of depression of middle school students in a city, and the effect of depression on the school adjustment. Korean J Couns, 9:645–658, 2008.

24.

Kim

, Cheon

, Kim

, Chang

, Yoo

, Kim

, Cho

, Seo

, Bae

, So

, Noh

, Koh

, McBurnett

, Leventhal

: The reliability and validity of Kiddie-Schedule for Affective Disorders and Schizophrenia-Present and Lifetime Version-Korean version (K-SADS-PL-K). Yonsei Med J, 45:81–89, 2004.

25.

Kleinman

, Good

: Culture and depression. New Eng J Med, 351:951–952, 2004.

26.

Kovacs

: The Children's Depression Inventory: A self-rated depression scale for school-aged youngsters. Unpublished manuscript. Pittsburgh (Pennsylvania), University of Pittsburgh, 1982.

27.

Mayes

, Bernstein

, Haley

, Kennard

, Emslie

: Psychometric properties of the Children's Depression Rating Scale-Revised in adolescents. J Child Adolesc Psychopharmacol, 20:513–516, 2010.

28.

Min

, Kang

: Analysis of School-based Mental Health Policy Stream based on Kingdon's Policy Stream Model. J Korean Soc Sch Health, 28:139–149, 2015.

29.

Mundfrom

, Shaw

, Ke

: Minimum sample size recommendations for conducting factor analyses. Int J Test, 5:159–168, 2005.

30.

Myers

, Winters

: Ten-year review of rating scales. I: Overview of scale functioning, psychometric properties, and selection. J Am Acad Child Adolesc Psychiatry, 41:114–122, 2002.

31.

Poznanski

, Grossman

, Buchsbaum

, Banegas

, Freeman

, Gibbons

: Preliminary studies of the reliability and validity of the children's depression rating scale. J Am Acad Child Psychiatry, 23:191–197, 1984.

32.

Radolf

: The CES-D scale: A self-report depression scale for research in the general population. Appl Psychol Meas, 1:385–401, 1977.

33.

Schwartz

, Farver

, Chang

, Lee-Shin

: Victimization in South Korean children's peer groups. J Abnorm Child Psychol, 30:113–125, 2002.

34.

Shaffer

, Gould

, Brasic

, Ambrosini

, Fisher

, Bird

, Aluwahlia

: A children's global assessment scale (CGAS). Arch Gen Psychiatry, 40:1228–1231, 1983.

35.

Statistics Korea: Causality of Mortality in 2010.. Daejeon (South Korea), Statistics Korea, 2011.

36.

U.S. National Institutes of Health: Safety and Tolerability Study of Once-weekly Oral Aripiprazole in Children and Adolescents with Tourette's Disorder. 2015. Available at https://clinicaltrials.gov/ct2/show/NCT01416441?term=CDRS-R+Korea&rank=2, Accessed November 26, 2016 .

37.

U.S. National Institutes of Health: Lurasidone Pediatric Bipolar Study. 2016. Available at https://clinicaltrials.gov/ct2/show/NCT02046369?term=lurasidone+KOREA&rank=2, Accessed November 26, 2016 .

38.

Weissman

, Bland

, Canino

, Faravelli

, Greenwald

, Hwu

, Joyce

, Karam

, Lee

, Lellouch

, Lepine

, Newman

, Rubio-Stipec

, Wells

, Wickramaratne

, Wittchen

, Yeh

: Cross-national epidemiology of major depression and bipolar disorder. JAMA, 276:293–299, 1996.

39.

Yoo

S-K

, Skovholt

: Cross-cultural examination of depression expression and help-seeking behavior: A comparative study of American and Korean college students. J Coll Couns, 4:10–10, 2001.

40.

Youngstrom

. A primer on receiver operating characteristic analysis and diagnostic efficiency statistics for pediatric psychology: We are ready to ROC. J Pediatr Psychol, 39:204–221, 2014.

41.

Youngstrom

, Genzlinger

, Egerton

, Van Meter

. Multivariate meta-analysis of the discriminative validity of caregiver, youth, and teacher rating scales for pediatric bipolar disorder: Mother knows best about mania. Arch Sci Psychol, 3:112–137, 2015.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.04 MB