Patients with major depressive disorder (MDD) exhibit depressive symptoms, such as depressed mood, anhedonia, anxiety, changes in appetite, sleep disturbance, feelings of guilt, and suicidal ideation [1,2]. In addition to these symptoms common to MDD, bipolar disorder (BD) is characterized by hyperactivity, mood swings, and risk- taking behavior and/or the history of them [1,3]. Recurrence is common in both diseases, and negatively influences social functioning [4-6].

To evaluate depression severity, several clinician-rated assessment scales (CRs) have been used, such as the Hamilton Depression Rating Scale (HAMD) [7,8] and the Montgomery−Åsberg Depression Rating Scale [9]. There are also self-report questionnaires (SRs), such as the Quick Inventory of Depressive Symptomatology-Self Report [10] and the Beck Depression Inventory (BDI) [11], which have been validated. Combined use of CRs and SRs to measure severity of symptoms may facilitate measurement-based care, an essential component of evidence-based medicine [12], although the set of assessment may often be difficult in daily clinical settings with time-constraint.

Correlations between CRs and SRs for depression severity vary from 0.2 to 0.8 [13], indicating discrepancy in several cases. Recently, the importance of “patient-centered care” has increasingly been recognized [14,15]. In this framework, patients are encouraged to be actively involved in their own treatment decisions rather than simply following the direction of physicians. Accordingly, more emphasis is placed on assessments with SRs when evaluating treatment outcome. At the same time, it would be meaningful to identify factors that relate to the discrepancy between CRs and SRs. Previous studies suggest that demographic and personality factors are associated with such discordance regarding depression severity in MD patients. Specifically, the difference between data from CRs and SRs is eminent in patients with mild-to-moderate depression [16-18]. Furthermore, depressed patients who over-reported the severity tended to show personality traits of low extraversion [19,20], agreeableness [20,21], openness to experience [22], and self-esteem [16], as well as high neuroticism [19,20]. In addition, the discrepancy between CRs and SRs has been reported to be associated with worse response to treatments and cognitive distortions in BD [23].

It is important to examine which clinical factors are associated with the discrepancy between CRs and SRs in mood disorders. Such discordance is hypothesized to be affected by psychological characteristics related to depressive symptoms. In the present study, we aimed to examine the possible associations of the SRs/CRs difference with adverse childhood experiences, autistic-like traits, and coping styles in mood disorders.

Participants comprised 100 patients with MDD (n = 60) or BD (n = 40) who had received medical examinations from October 2017 to February 2020 in the outpatient clinic of the Mood Disorder Center for Advanced Therapy at the National Center of Neurology and Psychiatry (NCNP) Hospital, Japan. We included depressed patients with BD as well as those with MDD, which enabled us to find more general tendency in depressed patients irrespective of diagnosis, and enlarged the sample size to minimize the possibility of type II error. The study procedures were explained, and informed consent was obtained from patients (agreement rate: 76%). Study exclusions consisted of the following criteria: (1) being 16 years old or younger, or (2) the primary diagnosis not being of MDD or BD. The participants were screened for Axis I psychiatric disorders by a trained research psychologist using the Japanese version of the Mini-International Neuropsychiatric Interview [24]. The diagnosis was made according to the Diagnostic and Statistical Manual of Mental Disorders 4th Edition criteria [25], based on the information from the Mini-International Neuropsychiatric Interview, additional unstructured interviews, and medical records. This study was approved by the ethics committee of the NCNP, Tokyo, Japan (A2021-081).

Clinical and Psychological Assessments

Depression severity was evaluated by trained research psychologists by using the GRID Hamilton Depression Rating Scale 17-items (HAMD-17) [7,8], while manic symptoms were measured by the Young Mania Rating Scale [26]. Subjectively perceived depressive symptoms were assessed using the Japanese version of the BDI. The BDI was developed by Beck et al. [11] as a screening tool to calculate the severity of depressive symptoms using 21 items, each of which is rated from 0 to 3, with a total score range of 0−63. The Japanese version of the BDI is a reliable and valid tool for measuring the severity of depressive symptoms [27].

A six-item short version of the Childhood Trauma Questionnaire (CTQ-6) [28] was used to self-evaluate adverse childhood experiences. Its six items are grouped into three factors: physical abuse, emotional abuse, and emotional neglect. Items are rated using a five-point Likert scale ranging from 1, “never true,” to 5, “very often true”.

The Japanese version of the Social Responsiveness Scale for Adults (SRS-A) [29-31] was used as a reliable and valid tool for measuring autistic-like traits. It evaluates self-reported autistic-like traits of respondents via 63 items that are grouped into five categories; social awareness, social cognition, social communication, social motivation, and restricted interests. Items are rated using a four-point Likert scale ranging from 0, “not true,” to 3, “almost always true”.

The Japanese version of the Ways of Coping Checklist (WCCL) [32,33], a self-report instrument, was used to evaluate coping styles. Its 47 items are grouped into six coping styles; problem-solving, positive reappraisal, social support, self-blame, wishful thinking, and escape-avoidance. Each item is rated using a four-point Likert scale ranging from 0, “not used,” to 3, “regularly used”.

Patient Categories by Discrepancies between Patient- and Clinician-rated Scales

This study aimed to examine the possible associations of the SRs/CRs difference with psychological characteristics related to depressive symptoms. The patients were classified into three groups according to a previous study [34], based on distance from a point to a line in Euclidean geometry [35]. Pearson correlations were calculated between the HAMD-17 and BDI scores, and their regression line was determined. The distance from the regression line to the BDI scores was measured and the magnitudes of the distances were divided into quartiles. The top 25% consisted of persons with BDI scores higher than the predicted value (BDI relative-overrating group, BO group), the middle 25−75% consisted of those with BDI scores similar to the predicted value (BDI relatively concordant group, BC group), and bottom 75−100% consisted of those with lower BDI scores than the predicted value (BDI relative-underrating group, BU group).

Statistical Analysis

Statistical analyses were performed using SPSS Version 25.0 (SPSS Japan). Continuous variables were compared among the BO, BC, and BU groups using Kruskal−Wallis test for mean education level, chlorpromazine-equivalent dose, HAMD-17, BDI, YMRS, SRS-A, WCCL, and CTQ-6 scores. Between-group differences were examined by pairwise multiple comparisons using rank sums, as proposed by Dunn [36]. Pairwise multiple comparisons were conducted with Bonferroni correction. Categorical variables were compared by chi-square test.

Correlation and Regression Analyses between HAMD and BDI Scores

HAMD scores were correlated significantly with BDI scores (Pearson’s r= 0.624, p < 0.001) (Fig. 1). Regression analysis by least squares method was performed with HAMD scores as the independent variable and BDI scores as the dependent variable, which yielded the equation of; BDI-II = 1.1 × HAMD-17 ＋ 7.89 (R² ＝ 0.389).

Analyses of Clinical Variables

Demographic and clinical characteristics of the participants are shown in Table 1. The median HAMD score was 13.0, whereas the median for BDI scores was 21.0. The percentage of comorbid psychiatric disorders with agoraphobia, social anxiety disorder, panic disorder, alcohol use disorders, and obsessive-compulsive disorder were 13.0%, 10.0%, 9.0%, 6.0%, and 4.0%, respectively.

Demographic and clinical characteristics of the participants allocated to the BO, BC, or BU groups are shown in Table 1. Total BDI scores were significantly different among the three groups, while their HAMD scores and other clinical variables did not significantly differ. In addition, there were no significant between-group differences in demographic and clinical characteristics.

Analyses of Psychological Measures

Data on psychological measures are shown in Table 2. Total CTQ-6 scores and emotional abuse subscale scores were significantly higher for the BO group compared to the BU group. By contrast, physical abuse and emotional neglect scores were not significantly different among the three groups.

SRS-A total scores were significantly higher for the BO group compared to the BC or BU group. The same applied to SRS-A subscales scores of social cognition, social communication, social motivation, and autistic mannerisms. On the other hand, social awareness scores were not significantly different among the three groups.

Among the six WCCL subscales, self-blame scores were significantly higher in the BO group than in the BC and BU groups. Wishful thinking and escape-avoidance scores were significantly higher in the BC group than in the BU group. By contrast, subscale scores on coping strategies of problem-solving, positive reappraisal, and social support did not significantly differ among the three groups.

To our knowledge, this is the first study to demonstrate the contribution of adverse childhood experiences, autistic-like traits, and coping styles to the difference between severity of depressive symptoms reported subjectively (with BDI) and that measured objectively (with HAMD) in patients with mood disorders. We also confirmed the correlation between BDI and HAMD scores in these patients, as reported in previous studies [37,38].

Our results indicating discordant data between the clinician-rated scale and self-report questionnaires suggest that caution is needed when assuming subjective symptoms based solely on objective observations. That is, clinician-rated scales may not effectively represent the actual distress experienced by some patients with mood disorders. These considerations suggest that the use of both SR and CR scales is desirable especially when emotional abuse experiences, autistic symptoms, and/or maladjusted coping styles are indicated through family history, history of present illness, and examination scenes.

Experience of emotional abuse was significantly more frequent in the BO group than in the BU group, suggesting negative influence of adverse childhood experiences on inner distress. Experiences of emotional abuse have been indicated to be associated with depressive symptoms in adulthood [39,40]. The current observations extend the findings from these previous studies by suggesting the link between emotional abuse and the dissociation between subjective and objective data on depressive symptoms. Also, our findings may be consistent with a greater increase in depressive symptoms in people who experienced emotional abuse compared to those who did not, when confronted with current depression-inducing stressors, such as interpersonal conflicts or academic/vocational failure [41]. Possibly, experiences of emotional abuse in childhood may enhance sensitivity to stressors, as manifested by the exaggerated subjective distress in patients with mood disorders.

Autistic-like traits (i.e., impaired social communication and social motivation, autistic mannerisms, etc.), as measured by the SRS-A, were more pronounced in patients who reported relatively severe subjective depression (BO group) compared to those who do not (BC and BU groups). In our previous study [42], SRS-A scores were associated with objective depressive symptoms, as evaluated by the HAMD, in patients with MDD. These findings suggest subjective, rather than objective, distress should be a major concern for clinicians who take care of people with autistic traits. In contrast, the BU group (patients with relatively less subjective symptoms) elicited lower SRS-A scores than the BO group, consistent with an association between under estimation of one’s depressive states and extroversion (i.e., sociable, having interests/concerns toward the outside world), as previously shown [19].

Patients with relatively severe subjective depressive symptoms (BO group) elicited higher scores for self-blame as evaluated by the WCCL, than those in the other groups (Table 2). This finding may be partially consistent with Hori et al. [43] who observed that emotion-focused coping is frequently used by patients with MDD. In agreement with the present results, exaggerated self-estimation of depressive states has been associated with cardinal factors for stress coping, such as cognitive distortions [23], as well as high neuroticism and low extraversion [19,20]. Specifi-cally, neuroticism has been reported to be moderately correlated with emotion-focused coping, such as self-blame in patients with depression [44]. These considerations may help identify patients with depressive symptoms who might benefit from psycho-education for coping strategies.

Limitations

This study has a few limitations that should be considered. First, patients in our study presented relatively milder symptoms (Table 1), with a relatively young median age, which may limit the generalizability of the results. Second, most patients were treated with antidepressant drugs following their first contact with psychiatric services; potential pharmacological effects were not considered. Further research with patients who have not been prescribed with antidepressants is warranted. Third, the relatively small sample size of this study was subject to type II errors. Fourth, this study did not include subjective measures of manic symptoms, unlike in the case for depression severity. Future research should determine the present or absence of discrepancy between subjective and objective manic symptoms. Finally, this study did not include objective measures of adverse childhood experience, autistic traits, coping style, unlike the case for depression severity. Future research should aim to investigate the presence or absence of discrepancy between subjective and objective data for these psychosocial aspects.

Conclusions

The results of the present study suggest that patients with mood disorders who express self-blame coping style, autistic-like traits, and/or adverse emotional experience perceive greater distress than that evaluated objectively by clinicians. Our data also indicate the need for inclusion of subjective assessments to effectively evaluate depressive symptoms in patients deemed to have the above psycho-behavioral problems.

Evaluation with the M.I.N.I. and the data were provided by NCNP biobank.

No potential conflict of interest relevant to this article was reported.

This work was supported by the Strategic Research Program for Brain Sciences from the Japan Agency for Medical Research and Development, AMED (Grant number: 20dm0107100h0005) (Hiroshi Kunugi), Intramural Research Grants for Neurological and Psychiatric Disorders (Grant number: 30-1) (Hiroshi Kunugi) and (Grant number: 3-1) (Tomiki Sumiyoshi) from the National Center of Neurology and Psychiatry, and Japan Health Research Promotion Bureau Grants (Grant number: 2021-B-01) (Tomiki Sumiyoshi).

Conceptualization: Hiroshi Kunugi, Risa Yamada. Trans-lated questionnaire English into Japanese (CTQ-6): Hiroaki Hori. Funding: Hiroshi Kunugi, Tomiki Sumiyoshi. Data acquisition: Tomiki Sumiyoshi, Sumiko Yoshida, Takeshi Fujii, Kotaro Hattori, Risa Yamada, Naoko Ishihara, Tomoko Kurashimo, Hiroshi Kunugi. Formal analysis: Risa Yamada, Ryo Matsumura. Writing−original draft: Risa Yamada. Super-vision: Hiroshi Kunugi, Tomiki Sumiyoshi, Hiroaki Hori, Takeshi Fujii. All authors have read and approved the revised manuscript.