Introduction
Major Depression Disorder (MDD) is a prevalent mental disorder. Studies in various countries have reported its lifetime prevalence of, for example, 20.5% in Chile [
1], 21% in France [
2], 6.7% in South Korea [
3] and 3.4% in China [
4]. In 2008, the World Health Organization (WHO) ranked MDD as the third leading contributor to the global burden of disease, predicting that it would ascend to the primary position by 2030 [
5,
6]. Among Chinese children and adolescents, depressive symptoms have become more prevalent and the pooled prevalence of depressive symptoms among them was 22.2% [
7]. Depression often leads to a cascade of consequences, including suicidal ideation, school dropout, behavioral disturbances, and substance abuse among children and adolescents [
8]. Additionally, 10%-25% of patients with mild symptoms are at risk of scoring in the severe range over one to three years without intervention in time [
9]. Adolescence is frequently regarded as a pivotal period for the early identification and prevention of adult depression [
10], highlighting the vital importance of prompting diagnosis and intervention to prevent the onset of depression in adolescents.
Using simple and efficacious screening tools is an effective way to improve the detection of mental disorders [
11]. The US Preventive Services Task Force (USPSTF) recommended to adopt screening tools like the Patient Health Questionnaire for Adolescents (PHQ-A) and the primary care version of the Beck Depression Inventory (BDI) for the identification of depression among adolescents in primary care settings [
12]. Self-rating scales are simple and efficacious screening tools to screen depression among adolescents. It can not only help clinicians to quantify the patients' subjective feelings, but also enable patients understand the severity of their distress. Meanwhile, the Measurement-based care (MBC), which focus on the periodical assessments of the treatment in the process of quality controls, has been widely recommended in psychiatric practice. The MBC allows clinicians to make personalize treatment decisions for the patients, thereby improving the adoption of appropriate treatment strategies, reducing treatment resistance, and increasing treatment quality [
13].
In clinical practice, the depression assessment scales used for children and adolescents mainly include the Children's Depression Inventory (CDI), the Center for Epidemiological Studies Depression Scale (CES-D), the primary care version of the Beck Depression Inventory (BDI), the Mood and Feelings Questionnaire (MFQ) and the Patient Health Questionnaire for Adolescents (PHQ-A/PHQ-9 M) [
12,
14].
The PHQ-A/PHQ-9 M is adapted from the 9-item Patient Health Questionnaire-9 (PHQ-9). The PHQ-9 was initially developed by Kroenke and Spitzer in 2001 for assessing depression in adult primary care and then was extended to adolescent depression [
15‐
18]. The advantage of the PHQ-9 was that it exclusively focuses on the 9 diagnostic criteria of MDD in the fourth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV), which makes it more specific for major depression and may accurately discriminate depression from anxiety or even general psychological distress [
19]. However, previous studies reported that the specificity of the PHQ-9 was lower when used in adolescents, which could lead to an increased likelihood of false positive rate [
16,
20], suggesting that the PHQ-9 may not be the most suitable screening tool for assessing depression in adolescents. This could potentially be attributed to the specific features of adolescent depression that may differ from adult depression. Although some of the symptoms of MDD may be similar for adults and adolescents, its clinical feature and prominence may be different [
21]. Irritability is an impairing clinic manifestation that refers to easy annoyance and touchiness, which has been the most frequently reported symptom in adolescent depression [
21,
22]. Although the DSM-IV identifies irritability as a characteristic of adolescent depression, it states that ‘it is not a criterion for major depressive episode’ [
21]. Yet both the DSM-5 and the 11th edition of the International Classification of Diseases (ICD-11) include the criteria to specifically define MDD for adolescent with the statement ‘Depressed mood (subjective or observed) can be irritable mood in children and adolescents’ [
23], suggesting the importance of irritability on the diagnosis of MDD among adolescents. Comparing to the PHQ-9, the PHQ-A adds ‘irritability’ in the description of item 1, which is the major revision. Moreover, two minor revisions include that the order of item4 (Fatigability) and item5 (Appetitive problems) reverse, and the PHQ-A added ‘school work’ in item7 (Concentration problems), which adapts to children`s daily activities. With these revisions, the PHQ-A may be more appropriate for screening depression in adolescents and adjusts the DSM-5 and ICD-11 diagnosis better.
The English version of the PHQ-A showed satisfactory sensitivity, specificity and overall diagnostic accuracy and its reliability and validity was proved satisfactory [
24,
25]. The PHQ-A has been translated into other languages such as Urdu, Thai, Portuguese and Arabic [
26‐
29]. All of these translations demonstrated satisfactory reliability and validity in relative countries and populations. In the Urdu, Thai, and Arabic version, only one factor was extracted by exploratory factor analysis (EFA). Furthermore, the unidimensional factor structure is verified by confirmatory factor analysis (CFA) in the Urdu and Arabic version [
26‐
29]. Therefore, the PHQ-A is a promising screening tool and merits further evaluation among adolescents in China.
In this study, we hypothesized that the PHQ-A would fit into a unidimensional structure and demonstrate good psychometric properties among Chinese children and adolescents with MDD. We would test the hypotheses and recommend a cut-off value for remission.
Discussion
We investigated the psychometric properties of the PHQ-A among Chinese children and adolescents with MDD in the study.
The CFA showed that except for the item 3 (Sleeping problems), all items showed relatively adequate loadings on the latent factor. This finding of the relatively lower loading on item 3 was consistent with the study of the Jordan version of the PHQ-A [
26], suggesting that the item 3 might be less predictive. Generally, all the goodness-of-fit indices of the model fell within the acceptable parameter range, indicating acceptable structural validity of the scale. The CFA of the PHQ-A confirmed its unidimensional structure, which is consistent with previous research findings [
26‐
29].
In terms of the reliability of the PHQ-A, the internal reliability was examined by Macdonald Omega coefficient. The value of the coefficient was 0.87, indicating good internal consistency and high item homogeneity within the PHQ-A. These findings are consistent with previous research on the Arabic, Urdu, Mozambique, and Thai versions of the PHQ-A [
26‐
29]. The Spearman correlation coefficient of the scale was 0.70, which suggested good test–retest reliability for the scale, indicating the stability of the PHQ-A. These findings align with prior studies [
27‐
29]. The satisfactory internal reliability and fair test-retest reliability of the PHQ-A confirmed its acceptable reliability among Chinese adolescents with MDD.
For criterion validity, the CES-D and MFQ were selected as the criterion scales. The total score of the PHQ-A was highly positively correlated with CES-D and MFQ, indicating that the PHQ-A had good criterion validity with them. The results of the Kruskal-Wallis H Test demonstrated that the PHQ-A could successfully discriminate different levels of severity of depressive symptoms among Chinese children and adolescents, consistent with the results of earlier research [
25,
26,
28,
29].
This study showed that the area under the ROC curve (AUC) was 0.99, suggesting that the PHQ-A might be a valid tool. However, the AUC of the Thai and Mozambique versions were 0.88 and 0.85 respectively [
27,
29]. The very high AUC value in this study might be due to the small sample size of remission (
N = 19). The optimal cut-off score for the PHQ-A was 7 (7 will be a stricter criterion for remission which is clinically helpful), conforming with the recommended scores from previous studies on other PHQ-A versions [
25,
26,
28,
29]. Both the Thai and Mozambique versions of the PHQ-A had an optimal cutoff value of 8. The sensitivity and specificity of the Thai version of PHQ-A were 76% and 81% [
29]. The sensitivity and specificity of the Portuguese version were 78% and 80% [
27]. Both were lower than the results in our study.
The results indicated that the psychometric properties, the factor structure and the optimal cut-off score of the PHQ-A performed relatively stable across different cultures and languages. There are several limitations in this study. Firstly, since our study focused on only MDD patients, its findings might not apply to other clinical population. Secondly, the sample was recruited from a single mental health center, which was less ideal than multi-center recruitment. Thirdly, this research is cross-sectional, which is unable to confirm whether the scale is sensitive to changes of symptoms after treatment. Finally, the relatively small sample size of remission might cause bias in AUC. Therefore, future studies are needed to test the psychometric characteristics of the PHQ-A in different populations and verify its sensitivity to treatment changes.
Acknowledgements
This study was completed in Guangdong Mental Health Center, Guangdong Provincial People's Hospital and the authors appreciate all the colleagues for their support and encouragement. Besides, the authors wish to express their gratitude to the suggestion and support of Professor Jia Fu-Jun, Professor Hou Cai-Lan, Professor Zheng Hui-Rong, Professor Jiang Mei-Jun, Professor Ning Bu, Doctor Tang Yi, Doctor Yang Yuan, Doctor Xu Cai-Feng and Doctor Zeng Shu-Ning.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.