Is DSM widely accepted by Japanese clinicians?


Correspondence address: ToshiyukiSomeya Department of Psychiatry, Niigata University Graduate School of Medical and Dental Sciences, 1-757 Asahimachidori, Niigata 951-8510, Japan. Email:


Abstract The Diagnostic and Statistical Manual of Mental Disorders, 3rd edition (DSM-III), a new standardized diagnostic system with multiaxial diagnosis, operational criteria and renewed definitions of mental disorders, was introduced in 1980 and prompted movements to reform conventions in Japanese psychiatry. This review overviews the initial response of Japanese clinicians to accept DSM-III, and its effects on the development of systematic research of psychiatric diagnosis. These new research activities include those on reliability of psychiatric diagnosis, application of various evaluation tools, discussion on the concept of mental disorders, relation of personality disorders with depressive disorders, and Taijin-kyofusho, or culturally distinctive phobia in Japan. A reference database search to survey the latest trend on psychiatric research indicated that the number of papers published by Japanese workers increased sharply after 1987, and DSM apparently greatly influenced their internationalization. Twenty years after the publication of DSM-III, a questionnaire on the use of DSM-IV was set out in 2000 to survey how widely DSM is utilized in clinical practice in Japan. Two hundred and twelve psychiatrists answered the questionnaire, and the results show that DSM has been accepted positively by the younger generation, while the older generation (over 40s) has still less interest in DSM, and DSM is used mainly for research purposes rather than in daily practice.


Before the recent development of classification system standardized for mental disorders, such as the Diagnostic and Statistical Manual of Mental Disorders (3rd edition) DSM-III published in 1980, little attention was paid to the standardization of psychiatric diagnosis in Japan, and each clinician was accustomed to using his or her own nomenclature. Clinicians in this country did not widely use even the ninth revision of the International Classification of Diseases (ICD-9) in their daily practice, but they were accustomed to make use of the so-called ‘conventional diagnosis’, which was based on the German psychiatry directly imported in the beginning of 20th century.

Some psychiatrists, attentive to the new trend in psychiatry and who had not been satisfied with the situation in Japan at the time, were encouraged by the publication of DSM-III and became first interested in translating it into Japanese, expecting that this new tool could unify their diverse concepts and diagnostic conventions of mental disorders.

At first, a series of 17 introductory papers titled ‘Application of APA's DSM-III Diagnostic Criteria’ described the outline of DSM-III with critical discussions and gave tentative Japanese translation of its diagnostic criteria.1 This series had a marked impact on many psychiatrists, and some of them were eager to incorporate new terminology for mental disorders, operational criteria and multiaxial diagnosis employed in DSM-III. After that, the Japanese edition of MINI-D2 was published in April of 1982 and enthusiastically accepted by many psychiatrists in Japan.

They ran the project of multi-institutional reliability studies of DSM-III and ICD-9 with the collaboration of seven Japanese university clinics. Reliability of the DSM-III could then be proved and researchers hoped that it would substitute their rather obsolete evaluation method in Japan, and thus they could establish internationally acceptable practices in psychiatric diagnosis, which might serve educational, clinical and research purposes. The project constituted two parts: field trials using joint in-person interviews, and a study using videotaped materials. The purposes of the study were to examine whether the classification and diagnostic criteria of DSM-III ensured good diagnostic reliability and to what extent the reliability and classification differed between DSM-III and ICD-9 as used in Japan.3

In the first part of the study, 103 physicians from seven psychiatric departments of university hospitals and clinics participated. Data collected from 345 new outpatients admitted to the clinics revealed that the overall Kappa for 16 DSM-III major classes of axis I diagnosis was 0.71; a value comparable with the result of DSM-III field trials previously reported in the USA, while it was as low as 0.43 in the class of axis II personality disorders. For the ICD-9 diagnosis we obtained an overall Kappa of 0.64, the value apparently lower than that for the DSM-III diagnoses. Kappa values fluctuated widely with high values for common illnesses such as organic psychotic conditions, alcohol dependence and schizophrenic psychoses, and rather low values for the remaining classes of disorders.

In order to explore the possibility of variance among the seven institutions and the source of disagreement, the second part of our reliability study was carried out with the use of video-taped materials. Fourteen patients with different DSM-III categories were selected, and video-taped interview scenes were edited for 30 min for each patient who was willing to cooperate in our study and who gave their consent for the use of video-taped interviews. One hundred and forty raters participated in the study, and 1187 evaluation forms were collected. Overall diagnostic reliability for the major classes of axis I was 0.80. It should be noted that this was the first attempt to investigate diagnostic reliability with video-taped material in Japan. The DSM-III served to reduce variability in diagnoses among raters and thus improved diagnostic reliability.

However, we obtained the overall Kappa for the ICD-9 diagnoses, 0.75, with some disagreement in diagnoses observed among the seven institutions. These results proved that DSM-III was suitable for general use even though physicians had different medical education backgrounds and varying clinical experiences.

In November 1981, the 1st Meeting of the Japanese Committee of the International Diagnostic Criteria in Psychiatry (JCIDCP) was held in Tokyo and participants discussed several issues on psychiatric diagnosis comparing our conventional system with the international one.

Since the first meeting was held, we have met every year, and later this group was reformed to become a scientific organization called Japanese Society for Psychiatric Diagnosis (JSPD), and it celebrated its 20th anniversary in October 2000. Activities of this society prompted the publication of a journal, Archives of Psychiatric Diagnostics and Clinical Evaluation, which has its focus on this field in psychiatry. This journal has been published quarterly and has reached 44 issues. Although published in Japanese, this journal accepts only manuscripts that are well prepared and of a high standard of writing, and has facilitated our insight into the knowledge and strategy rationale for psychiatric diagnosis research. For example, the December 2000 issue was devoted to a reappraisal of Cloninger's personality theory, with nine interesting papers.4


Initial movements to accept DSM-III has generated various research activities on psychiatric diagnosis so as to catch up with this international trend.

Until recently, Japanese edition of publications on DSM have been published, including DSM-III-R, DSM-IV and their interpretative books and case books.5–12 In the fall of 2001 the translation edition of DSM-IV-TR is scheduled to be published.

Reliability studies of psychiatric diagnosis

Our own evidence of significantly better reliability in the psychiatric diagnosis using operational criteria of DSM-III,3 has greatly impacted the conventionalism in our clinical works, thus it has stimulated some clinicians to be engaged in a new research area in psychiatry.

Many scales were translated to test the inter-rater reliability of the Japanese version, and applied to further clinical research. For example, the translated edition of Positive and Negative Syndrome Scale (PANSS) was proved a reliable and efficient tool for comprehensive assessment of the schizophrenic syndrome in patient with DSM-IV schizophrenia.13 Test–retest reliability of the Personality Diagnostic Questionnaire Revised (PDQ-R), which was developed to assess DSM-III-R personality disorders, was translated and it was shown that highly reliable diagnoses of personality disorders were possible, even using a self-report questionnaire, if the axis-I symptoms of depression and anxiety had moderately subsided.14 The discriminant validity of the Japanese version of the Inventory to Diagnose Depression (IDD) was examined, a self-report instrument designed to diagnose major depressive disorder as defined in DSM-III-R. The concordance between IDD and clinical diagnoses, using a structured interview, was significant high for the group of patients with major depression, with a Kappa index as high as 0.81, and the IDD total score in this group was distinguished from the group with anxiety disorders.15 They also investigated its test–retest reliability in 30 patients with major depression, and reported that the agreement of diagnostic performance was substantial for 60 subjects, and Cronbach's internal consistency was sufficient.16

The competency of psychiatric patients to give informed consent is important in order to respect patients' decisions as well as protecting patients from undue exploration. One hundred and seventy-six experienced psychiatrists gave clinical judgements in a questionnaire of competency in five case vignettes, and their inter-rater reliability of competency judgement was reported to be slight with a generalized Kappa of 0.31 and clinicians' global judgement proved to be not sufficiently reliable.17

Application of new evaluation tools

Various scales and inventories were translated into Japanese and applied to a clinical population. Several studies investigated comorbidity and associations between mental disorders using these validated rating scales in patients with different mental disorders meeting the DSM-III, DSM-III-R and DSM-IV diagnostic criteria.

A study on the association between anxiety and depressive symptoms provided supportive evidence for the common mechanisms to anxiety and depression.18 Clinical characteristics and comorbid personality disorders were assessed in patients with both anorexia nervosa and obsessive–compulsive disorder (OCD) using Structured Clinical Interview for axis II disorders (SCID-II) and Yale–Brown Obsessive Compulsive Scale (Y-BOCS), and found that many patients with anorexia nervosa manifested significant impairment from primary OCD symptoms similar to inherent OCD patients.19 They also studied the prevalence and symptomatology of comorbid OCD in 26 patients with DSM-III-R bulimia nervosa. Comorbid OCD is a common phenomenon in Japanese bulimics (33%), similar to that suggested in patients with bulimia nervosa in Western countries.20

The prevalence of OCD was measured in Japanese students using the Japanese version of the Maudsley Obsessional-Compulsive Inventory (MOCI). Six (1.7%) of 350 interviewed students were diagnosed as OCD according to DSM-III-R.21

A Japanese translation of the Childhood Autism Rating Scale, Tokyo Version (CARS-TV) was used with 167 developmentally disabled children under the age of 16 years. The total CARS-TV score demonstrated a satisfactory level of taxonomic validity on DSM-III diagnostic groups. The total score discriminated infantile autism and other pervasive developmental disorders more efficiently from mental retardation than an IQ.22

Based on the nine dementia rating scales used in Japan, Europe and the USA, a new test was developed, and given to 203 subjects including the normal elderly as well as those suspected of suffering from dementia. Reliability and validity of this test were established against the clinical diagnosis with DSM-III as external criteria.23

Seventy Japanese DSM-III-R schizophrenic patients were assessed for 30 clinical symptoms using the Positive and Negative Syndrome Scale (PANSS). Principal component analysis was applied and disclosed five orthogonal independent symptom groups: negative, hostile/excited, thought-disordered, delusional/hallucinatory and depressive components. This supports the hypothesis that more than two dimensions are required to account for structures of the schizophrenic symptoms.24

Review of concepts of mental disorders

Since the time DSM-III was published, dispute on the concepts of common mental disorders has revived as one of the major topics in psychiatry.

Thirty-nine patients with childhood-onset schizophrenia, diagnosed according to DSM-III-R, who were under 15 years of age, were studied in two groups, those with obsessive–compulsive symptoms during prodromal phase, and those without. The former was a clinically distinct group from the latter, which suggests the possibility of subtype categorization.25

A total of 1432 outpatients attending psychosomatic clinics were assessed by the DSM-III-R or DSM-IV semistructured interview. The results seemed to reinforce the belief that the diagnoses on the axis only were not adequate enough to record psychosomatic phenomena. The most frequent diagnosis on axis I was somatoform disorders not otherwise specified, followed by bulimia nervosa, major depressive disorder, panic disorder and psychological factors affecting physical condition.26

The classification of mood disorders is one of the most highly debated topics in modern psychiatry, and the introduction of DSM-III has set a new standard in this controversy. The group for Longitudinal Affective Disorders Study (GLADS) in Japan has been conducting a multicenter prospective follow-up study of a broad spectrum of mood disorders in 23 participating centres in Japan using a polydiagnostic semistructured interview developed by the group. The classical neurotic depressions are diagnosed as major depression and not as dysthymia in DSM-IV.27

Attitudes of Japanese psychiatrists toward their patients who suffer from schizophrenia were investigated, using a self-reported questionnaire distributed to 150 executive board members of the Japanese Society of Psychiatry and Neurology. The results revealed that the grave tone of the Japanese term for schizophrenia influences a psychiatrist's decision to inform patients of the diagnosis and the term should be changed to one with less stigma.28

A structured interview using criteria from the DSM-IV major depressive episode and Research Diagnostic Criteria major depressive disorder was used to examine the 1 year incidence and prevalence of depression, among first-year university students. Twenty-three of 116 students had an onset of depression and this high rate may be explained by their readjustment after entering university.29

To demonstrate the prevalence of depression among physically ill patients and to investigate its effects on the length of hospital stay, 65 patients meeting the DSM-IV major depression were evaluated. Among the patients with benign general medical conditions, those with major depression stayed significantly longer at hospital than those without (82 vs 36 days). This is similar among cancer patients.30

Clinical research on newly listed diseases in DSM

Several concepts of diseases were revised and introduced to us through the innovation of the classification of mental disorders by DSM, which prompted clinical studies of these new disorders. Among them, panic disorder drew many researchers' attention. The clinical characteristics of panic disorder were investigated using cluster analysis to compare the symptom structure between those with panic attacks and those with limited panic attacks. Cluster analysis revealed that clusters of three and four panic symptoms in the panic attacks and limited panic attacks groups, respectively, and there were also differences in symptom structure between the two groups.31 To investigate the prevalence rates of panic disorder in the general population of Japan, a set of questionnaires were administered to 207 people who were then interviewed. Two subjects (1.0%) had DSM-III-R panic disorder and five had panic attacks but did not meet the criteria for panic disorder. There were comorbidity with agoraphobia in two cases and with major depression in five.32

Records of 44 children and adolescents were reviewed and 19 cases with monosymptomatic and 25 cases with polysymptomatic presentation of somatization disorder were identified. The latter may constitute a different entity from the monosymptomatic conversion disorder and may be an early manifestation of somatization disorder.33

Clinical characteristics of Tourette syndrome were studied in 64 patients according to DSM-III-R criteria. Generalized tics afflicting the entire body were found in 64% of subjects and coprolalia in 50%. Comorbid features were obsessive–compulsive symptoms in 62% and hyperactivity in 17%. Frequency of familial cases seemed to be lower than previously reported for Western patients.34 According to a questionnaire survey collected data of 154 patients with Tourette syndrome from respondents among psychiatrists and paediatricians, 45 (29%) had obsessive– compulsive symptoms and 10 (6.5%) had family histories of this syndrome. Many patients were associated with developmental disorders.35

Relation of personality disorders with depressive disorders

A few groups of clinicians who explored the comobidity of DSM-III-R and DSM-IV personality disorders in depressive patients, are advocators of immodithym, which is somewhat different from obsessive personality, called Kasahara's melancholy type personality, and which may predispose major depressive disorder.

A study was attempted to identify a specific personality disorder cluster associated with poor outcome after antidepressant therapy in 96 outpatients with major depression. Cluster A personality disorder was significantly correlated with the 4-month outcome of depression.36

The relationship between the melancholic type of personality and DSM-III-R personality disorders in 96 outpatients with major depression was investigated using a series of multivariate analysis. The results showed that the personality features of melancholic type personality were quite different from those of any DSM-III-R personality disorders, including obsessive–compulsive personality disorder.37

They identified three personality clusters, and patients with cluster B were less likely to have melancholic depression. Individuals with cluster B personality disorders may have depression with clinical features that have deviated from the depressive patients.38

The Structured Clinical Interview for DSM-III-R personality disorders (SCID-II) was administrated to 118 consecutive outpatients with major depression, and the frequency of personality disorders according to DSM-III-R criteria was found to be within the range of frequencies reported in North American and Europeans studies, although schizoid and narcissistic personality disorders were more frequent in a Japanese clinical sample with major depression.39

Higher prevalence of personality pathologies in early onset major depression may reflect a higher likelihood to connect into bipolar disorders, and the possibility that the predisposing personality pathology may be different in early-onset and late-onset major depression has been suggested.40

Typus melancholicus, as measured by the rigidity subscale of the Munich Personality Test (MPT), and maladaptive personality features, as measured by the Temperament and Character Inventory (TCI), was measured in 181 remitted patients with DSM-IV major depression. The results indicated that typus melancholicus and maladaptive personality features can coexist in some depressive patients.41

Whether personality is mediating the effects of adverse parenting and whether personality dimensions are related to developing lifetime depression was studied in 322 volunteer workers who were asked to complete the MPT, the Parental Bonding Instrument (PBI) and the Inventory to Diagnose Depression Lifetime version (IDDL). The results raised an objection to a hypothesis that adverse parenting experienced in childhood dispose one to a dysfunctional personality, which then predispose one to the development of depression in adulthood.42

Data from 59 patients with personality pathology were analyzed by cluster analysis and yielded eight adjective description typologies of patients, while multidimensional scaling identified three dimensions (anxious rumination versus behavioral acting out, overall severity of personality pathology, and assertiveness versus withdrawal), thus, rating for dimensional severity is proposed to be useful to define personality disorder.43

Clinical data of case reports with multiple personality disorder were collected, and compared with those from other countries. The cases in Japan differed significantly from those in North America in the mean number of personalities and prevalence of sexual and/or physical abuse.44

Is Taijin kyofusho a culturally distinctive phobia?

There is only one Japanese word that is listed in the glossary of culture-bound syndromes in DSM-IV. Taijin kyofusho is thought to be a culturally distinctive phobia in Japan, in some ways resembling social phobia in DSM-IV, and refers to an individual's intense fear that his or her body, its parts or its functions, displease, embarrass, or are offensive to other people in appearance, odor, facial expressions or movements. Several studies reported similarity and dissimilarity of the syndrome to social phobia or other phobic disorders in DSM-IV.

Avoidant personality disorder in a cultural context based on the Japanese concept of Taijin kyofusho as well as that of DSM-III-R and DSM-IV social phobia is discussed, with the data from 66 patients presented. Findings suggest that patients with avoidant personality disorder have had a long history of difficulties and share common personality problems with a milder form of Taijin kyofusho, which is conceptually different from social phobia.

They conclude that whether the syndrome is in fact one of the culture-bound syndromes is questionable, because Taijin kyofusho is only an ordinary social phobia, but the patients with this syndrome complained of their phobic anxiety in Japanese expression.45

Reference database search by Journal Citation Reports (JCR) and MEDLINE

The reference database search revealed that there were no psychiatric reports written by Japanese authors that appeared in international journals with an impace factor ≥ 1.0 before 1986. This result was obtained by computer search on journal citation systems such as JCR and MEDLINE by narrowing down references based on the author's position and nationality, thus old reports that did not indicate ‘Japan’ in the institution field might be omitted. Moreover, this survey covered only psychiatry-specific journals, while reports printed in general science journals such as Nature and Science were not included. The number of reports written by Japanese authors was 20 in 1987, 157 in 1997, and 197 in 1999, increasing by approximately 10-fold. The percentage of papers from Japan increased up to 3% of the total number of papers published in the world in 1999. As to reports on the psychiatric diagnosis issue, there was no report ever published internationally prior to the writing of a chapter in International Perspectives on DSM-III.3 It was only in 1989 that a report written by a Japanese author was printed for the first time in an international journal. They have tended to increase from 1,2,4 to 6 since 1996 and the percentage of the reports has also steadily increasing.

The reference database search showed a dramatic increase in the number of reports written by Japanese authors over the past 15 years, and considering that most of these reports used DSM diagnoses, DSM has certainly had a positive effect for the internationalization of Japanese research. Many clinicians are now convinced of usefulness of DSM diagnosis and have applied it to their research. This consequently has resulted in an increased number of reports.


Nearly 20 years have passed since the time DSM-III was established by American Psychiatric Association (APA) and was translated into Japanese and published in 1982.1 The DSM has been revised twice from DSM-III to DSM-III-R, then to DSM-IV, having been an object of discussion on diagnosis in psychiatry. How much has the DSM been accepted in the clinical practice of psychiatry? What are the results produced by DSM? To obtain the answers to these questions, a questionnaire was set out.46

Subjects and methods

Questionnaire on DSM-IV

The questionnaire was sent out in April and May 2000 to 654 psychiatrists who worked for the university hospitals or related hospitals of Fukushima Medical University, Niigata University, Toyama Medical and Pharmaceutical University, Kanazawa University, Kanazawa Medical University, or Fukui Medical University. Table 1 shows the questionnaire from which respondents made a choice from six questions and answered one question. The answers to the questionnaire were received by fax. Two hundred and twelve (32.4%) questionnaire sheets were returned.

Table 1.  Questionnaire on DSM-IV Thumbnail image of


The respondents were 212 psychiatrists (average age 46.6, SD 11.9). The age distribution was 15 psychiatrists in their 20s, 47 in their 30s, 77 in their 40s, 33 in their 50s, 30 in their 60s, nine in their 70s and one of unknown age. Average years of clinical experience was 19.7 years (SD = 12.2, 48 psychiatrists with 0–9 years, 67 with 10–19, 49 with 20–29, 28 with 30–39, 19 with over 40, and one did not answer). The position of respondents were: directors (n = 43), and medical staff of a hospital (n = 118), clinic office physicians (n = 17), internship (n = 2), university staff (n = 30), and did not provide an answer (n = 2). One hundred and seventy-nine of the 212 were Certified Mental Health Doctors.

Question 1: When did you know about DSM for the first time?

The DSM-II, DSM-III, DSM-III-R, and DSM-IV were published by APA in 1968, 1980, 1987, and 1994, respectively. Only 11 psychiatrists knew DSM-II and later versions; 143 knew DSM-III and later, 37 knew DSM-III-R and later, 19 knew DSM-IV. Corresponding to the publication date of each edition, psychiatrists in their 20s had their first experience with the DSM system from the DSM-IV, those in their 30s from DSM-III-R, and those over 40-years from DSM-III (Table 2).

Table 2.  Relationship between question 1 and age of respondents
  Age (years)
 ≤ 2930–3940–4950–5960–69≥ 70Total
  1. χ2 = 126.3; d.f. = 15; P < 0.001.

DSM-II  1 (2.1%) 5 (6.5%) 1 (3.1%) 3 (10.0%)1 (12.5%) 11 ( 5.3%)
DSM-III 1 (6.7%)17 (36.2%)63 (81.8%)28 (87.5%)26 (86.7%)7 (87.5%)142 (67.9%)
DSM-III-R 4 (26.7%)23 (48.9%) 7 (9.2%) 2 (6.3%) 1 (3.3%)  37 (17.7%)
DSM-IV10 (66.7%) 6 (12.8%) 2 (2.6%) 1 (3.1%)   19 (9.1%)
Total15 (100%)47 (100%)77 (100%)32 (100%)30 (100%)8 (100%)209 (100%)

Question 2: Have you ever used DSM to establish a diagnosis?

Most of respondents (193/212, 91%) have used DSM to establish a diagnosis. The majority of clinicians over 60 years of age had never used DSM (Table 3). It appeared that DSM was not sufficiently accepted by psychiatrists aged over 60 years of age, who have extensive experience. According to the comparison by position, DSM was used mainly in university hospitals (Table 4).

Table 3.  Relationship between question 2 and age of respondents
  Age (years)
 ≤ 2930–3940–4950–5960–69≥ 70Total
  1. χ2 = 19.2; d.f. = 5; P < 0.005.

No 1 (6.7%) 1 (2.2%) 6 (6.1%) 2 (6.1%) 5 (16.7%)4 (44.4) 19 (9.0%)
Yes14 (93.3%)46 (97.9%)71 (92.2%)31 (93.9%)25 (83.3%)5 (55.6%)192 (91%)
Total15 (100%)47 (100%)77 (100%)33 (100%)30 (100%)9 (100%)211 (100%)
Table 4.  Relationship between question 2 and position of respondents
 ManagerPost in hospitalClinicInternshipUniversity staffTotal
No 5 (11.6%)11 (9.3%) 3 (17.6%)  19 (9.0%)
Yes38 (88.4%)107 (90.7%)14 (82.4%)2 (100%)30 (100%)191 (91.0%)
Total43 (100%)118 (100%)17 (100%)2 (100%)30 (100%)210 (100%)

Question 3: What kind of criteria do you apply for outpatients?

The total number of responses was 269 due to multiple answers. One hundred and thirteen of 269 (42%) used the conventional Japanese criteria for clinical practice. Sixty-three psychiatrists (23%) used ICD and 93 (35%) used DSM. Older psychiatrists tended to use conventional criteria (49% of psychiatrists in their 50s and 67% in their 60s). In contrast, most younger clinicians used DSM or ICD. The generational boundary was at the age group in their 40s (Table 5). The results by position showed that DSM-IV or ICD-10 was generally used for outpatients in university hospitals. More than half of clinicians from university hospitals answered that they had established diagnoses for outpatients by operational criteria (Table 6).

Table 5.  Relationship between question 3 and age of respondents
 Age (years)
 ≤ 2930–3940–4950–5960–69≥ 70Total
1 (5.0%)20 (33.3%)40 (42.1%)23 (48.9%)24 (66.7%) 5 (45.5%)113 (42.0%)
ICD-8 or 9   2 (2.1%)    2 (0.7%)
ICD-109 (45.0%)14 (23.3%)19 (20.0%)10 (21.3%) 5 (13.9%) 4 (36.4%) 61 (22.7%)
DSM-III     3 (8.3%)  3 (1.1%)
DSM-III-R  1 (1.7%) 5 (5.3%)2 (4.2%) 2 (5.6%)  10 (3.7%)
DSM-IV10 (50.0%)25 (41.7)29 (30.5%)12 (25.5%) 2 (5.6%) 2 (18.2%) 80 (29.7%)
Total20 (100%)60 (100%)95 (100%)47 (100%)36 (100%)11 (100%)269 (100%)
Table 6.  Relationship between question 3 and position of respondents
 ManagerPost in hospitalClinicInternshipUniversity staffTotal
Conventional32 (62.7%) 65 (44.8%) 9 (42.9%)  6 (15.8%)112 (41.8%)
ICD-8 or 9 1 (2.0%)  1 (4.8%)   2 (0.7%)
ICD-10 8 (15.7%) 31 (21.4%) 6 (28.6%)1 (33.3%)15 (13.2%) 61 (22.8%)
DSM-III 2 (3.9%) 1 (0.7%)    3 (1.1%)
DMS-III-R  7 (4.8%) 1 (4.8%)  2 (5.3%) 10 (3.7%)
DSM-IV 8 (15.7%) 41 (28.3%) 4 (19.0%)2 (66.7%)25 (65.8%) 80 (29.9%)
Total51 (100%)145 (100%)21 (100%)3 (100%)38 (100%)268 (100%)

Question 4: How often do you use DSM?

Although 93 clinicians (35%) answered that they used DSM in an outpatient clinic in question 3, only 59 (24%) applied DSM to all outpatients. The DSM was used only for research and presentation purposes by 97 psychiatrists (39%); only for inpatients by 11 psychiatrists (5%); psychiatric evaluations in forensic settings, psychiatric evidence for involuntary civil commitment and case reports by 76 (11%) psychiatrists. The DSM was not used to establish a diagnosis for every diagnostic occasion, but was used mainly as criteria for research purposes. According to stratification by age and position, younger psychiatrists in university hospitals applied DSM to all outpatients, while experienced psychiatrists used DSM frequently only for psychiatric evidence of hospitalization (Tables 7, 8).

Table 7.  Relationship between question 4 and age of respondents
 Age (years)
 ≤ 2930–3940–4950–5960–69≥ 70Total
For all patients 6 (42.9%)18 (26.1%)25 (28.1%) 8 (20.0%) 1 (3.6%)1 (14.3%) 59 (23.9%)
Only for research 5 (35.7%)28 (40.6%)35 (39.3%)17 (42.5%)11 (39.3%)1 (14.3%) 97 (39.3%)
Only for inpatients 1 (7.1%) 6 (8.7%) 4 (4.5%)   11 (4.5%)
Psychiatric evidence 1 (7.1%)17 (24.6%)25 (28.1%)13 (32.5%)16 (57.1%)4 (57.1%) 76 (30.8%)
Never used 1 (7.1%)   2 (5.0%) 1 (14.3%) 4 (1.6%)
Total14 (100%)69 (100%)89 (100%)40 (100%)28 (100%)7 (100%)247 (100%)
Table 8.  Relationship between question 4 and position of respondents
 ManagerPost in hospitalClinicInternshipUniversity staffTotal
For all patients 4 (8.5%) 30 (21.3%) 3 (18.8%) 22 (55.0%) 59 (24.0%)
Only for research19 (40.4%) 62 (44.0%) 7 (43.8%)1 (50.0%) 8 (2.0%) 97 (39.4%)
Only for inpatients  5 (3.5%) 1 (50.0%) 5 (12.5%) 11 (4.5%)
Psychiatric evidence23 (48.9%) 42 (29.8%) 5 (31.3%)  5 (12.5%) 75 (30.5%)
Never used 1 (2.1%) 2 (1.4%) 1 (6.3%)   4 (1.6%)
Total47 (100%)141 (100%)16 (100%)2 (100%)40 (100%)246 (100%)

Question 5: What diagnostic categories in DSM-IV have you ever experienced?

Out of 30 diagnostic categories listed in question 5, the average number of experienced categories was 12.5 (SD 7.1). This value tended to decrease with age and there was a significant difference between clinicians aged in their 40s and those in 60s (Fig. 1). The average number for university staff was 18.9 (SD 4.7) and that was significantly higher than the other four groups, indicating that DSM was essential knowledge for psychiatrists teaching new psychiatry (Fig. 2).

Figure 1.

Relationship between question 5 and age of respondents.

Figure 2.

Relationship between question 5 and position of respondents.

Question 6: Have you ever used any terms associated with DSM?

The average number of terms for all respondents was 2.0 (SD 1.7). The number of terms used by clinicians in their 60s was significantly smaller than those in their 30s and 40s (Fig. 3). The results from questions 5 and 6 showed that a clear distinction exists in knowledge about DSM between clinicians under 40 years and those over 40 years.

Figure 3.

Relationship between question 6 and age of respondents.

Question 7: What are the advantages and disadvantages of DSM as compared to the conventional diagnostic criteria used in Japan?

The advantages and disadvantages of DSM were asked to evaluate DSM diagnosis. They answered that advantages were: (i) its higher reliability and objectivity, (ii) its usability as a common language, (iii) multiaxial diagnosis and comorbidity, (iv) good tool for research purposes, and (v) an internationally common criteria (Table 9). However, the disadvantages were (i) its over-detailed and burdensome features and takes a lot of time to consult criteria sets, (ii) even minor differences may result in different diagnosis, (iii) uselessness for treatment and understanding of patients, (iv) etiology was not included in its terminology, (v) may be difficult to understand for doctors in other departments, comedical staff, judicial participants, and families of the patient, and (v) different diagnoses resulting from too frequent revisions (Table 10). Opinions such as ‘too frequent revision’ and ‘over-operative criteria’ were derived from the negative acceptance of the characteristics of DSM rather than its disadvantages.

Table 9.  Summary of question 7 ‘advantages of DSM diagnoses’
DSM has high reliability and objectivity50 (of 212)
DSM can be used as a common language
(for psychiatrists in other sectors,
other specialties, judicial parties,
comedical staff.)
Multiaxial diagnoses, cormobidity37
Appropriate for research purposes26
Internationally common19
Even clinicians with little experience can use17
Solid concept of disorders, easy to comprehend
conditions, and well organized
Diagnoses can be made even in borderline cases16
Operative criteria16
Statistic uniformity15
New categories were adopted. (personality
disorder 4, children 4, V code 3, alteration of
classifications of neurosis and personality
disorders 2, bipolar 1)
Detailed images of disorder can be conveyed9
Accurately revised by adopting the latest progress7
Easy to explain to the patient and his/her family5
Easy to determine a therapeutic strategy5
Subjectivity is excluded5
Other two advantages presented by two
respondents each, and 23 advantages by one
Table 10.  Summary of question 7 ‘disadvantages of DSM diagnosis’
Over-detailed, time-consuming to consult criteria sets, troublesome, too many alternatives53 (of 212)
Inconsistency with the final diagnosis because of sticking to details. Minor differences results in different diagnosis
because of over-detailed criteria
Useless for treatment and understanding of patients23
Etiology is not considered. Difficult to understand psychopathology and true nature20
Difficult to understand for doctors in other departments, comedical staff, concerned judicial parties,
and patient’s family. Not widely used
Diagnoses are changed by revision. Revisions are too frequent19
Subtle differences of nuance can’t be conveyed. The condition of the disorder can’t be grasped because the hardness
and softness of condition and the possibility of comprehension can’t be held
Inconsistency with names of diseases in receipt and report16
Superficial comprehension. A condition is regarded as a disorder by merely meeting the requirement13
Only cross-sectional evaluation. Vertical variation is too wide12
Too many cases are classified into the “Unspecified” category12
Factors of conventional diagnoses are not reflected9
Criteria for personality disorder is insufficient8
Translation is difficult to catch on. Unfamiliarity with the name of a disease8
Many cases are yet to meet the criteria. Some categories involve insufficient descriptions6
No reason to use an American style of criteria6
Another disadvantage presented by four respondents, one by three, seven by two, and 28 by one

Discussion on the questionnaire results

Clinicians working with various institutions and in varied positions replied to the questionnaire for the survey about the present situation of DSM in Japanese psychiatry. The response rate was not high enough (32.4%) and respondents are likely to be more concerned with diagnostics than others. This is obvious from the answers to question 7; most clinicians were enlightened to a certain extent regarding DSM.

The DSM-III and later editions are well recognized by clinicians, in particular the younger generations of psychiatrists (Table 5). This may result from the postgraduate education for which DSM or ICD is used in many universities. The generations over 40s have little interest in DSM and are not familiar with the new terms applied in DSM (Figs 1,3). These differences are attributable to the generation gap caused by the introduction of DSM after clinicians over 60-year-old had finished their internship and had left their university and started their clinical works. They also reflect a negative generational attitude of those familiar with conventional diagnosis, toward the international criteria. Described above, recovery of the questionnaire was around 30%. Because clinicians only having relatively high interest in psychiatric diagnosis might reply, non-respondents in this older generation may be even less interested in DSM.

Although many psychiatrists have used DSM, only half applied DSM daily to outpatient clinics, while over 40% used conventional criteria (Tables 5,6). Few respondents applied DSM to all outpatients. The DSM diagnoses are used depending whether the situation requires communication with the ‘outside’ such as research and psychiatric examinations (Tables 7,8). This is discernible from the answers to question 7. While regarded as highly reliable and objective, DSM criteria are considered complicated, difficult to use and useless for treatment.

There are two problems with DSM use in Japan. First, there is a generation gap in the attitude for new knowledge such as DSM. Second, although DSM has been recognized as one of the professional training subjects, it is not fully used in clinical practice. What was the reason for this generation gap? The first factor is the inadequate postgraduate education system in the psychiatric field. Although the Mental Health Doctor Certification System of the Ministry of Welfare authorizes specialists in the psychiatric field, any psychiatric association does not have authorizing systems. The Mental Health and Welfare Act requires certified psychiatrists to take compulsory training every 5 years, however, the training program does not include re-education but places emphasis on the protection of human rights. Another factor is that psychiatrists who do not use DSM do not feel inconvenienced in daily practice. For example, the registration forms defined by Mental Health and Welfare Act and the certification forms for the disablement pension do not contain internationally registered diagnosis according to the operative criteria. Diagnoses based on DSM may be a disadvantage when bills for health care are submitted to the health insurance bureau where most officers are not familiar with the new terminology. Therefore, it seems that psychiatrists who have never used DSM in their postgraduate education have less necessity to learn DSM.

Isn’t a wider use of DSM hindered by misunderstanding of DSM by some individuals? ‘Traditional diagnoses’ or ‘conventional diagnoses’ that have originated from German psychiatry, are commonly used in Japan. It can be said that psychiatrists familiar with conventional diagnoses have not accepted DSM, although it is only modernized German psychiatry by American pragmatism. As to misconceptions of significance and purpose of DSM embraced by those individuals, this is vividly demonstrated by their stubborn answers to question 7.

The advantage of DSM diagnosis is its high reliability and objectivity. These advantages are derived from the clear criteria sets in DSM-III and later editions. This brings an improved communication among mental healthcare providers by using a reliable common language; an easier sharing of clinically useful knowledge by accumulating experiences of clinical courses and responsiveness to treatments; an excellent way of education and studying; enhanced objective reviews of causes, courses, diagnoses, and treatment of mental disorders; a symptomatologic review through these objective reviews; and an organization of mental disorders based on empirical evidence.11,12

Major criticisms of DSM include: over-detailed and mechanically fashioned, lack of flexibility, superficial observation, and leading to less understandings of psychopathology. For example, reduced flexibility by which lack of only a single requirement results in a different diagnosis, was often mentioned. In this aspect, the following explanation is in the DSM:

The diagnostic categories, criteria, and textual description are meant to be employed by individuals with appropriate clinical training and experience in diagnosis. The specific diagnostic criteria included in DSM-IV are meant to serve as a guideline to be informed by clinical judgement and are not meant to be used in a cookbook fashion. For example, the exercise of clinical judgement may justify giving a certain diagnosis to an individual even though the clinical presentation falls just short of meeting the full criteria for the diagnosis as long as the symptoms that are present are persistent and severe.

Diagnoses not corresponding to the clinical pictures should not be established, as much as is possible. It is practical that the accumulation of information and an oversight in symptoms are reviewed before the mechanical application of criteria.

It must be also noted that application of operational criteria is not always a complete diagnosis. DSM clearly states:

Making a DSM-IV diagnosis is only the first step in a comprehensive evaluation. To formulate an adequate treatment plan, the clinician will invariably require considerable additional information about the person being evaluated beyond that required to make a DSM-IV diagnosis. The operational criteria is only a guideline for the diagnostic classification, and that that is only a part of clinical diagnoses.

Other criticisms are ‘superficial observation’ and ‘leading to less understandings of psychopathology’. These criticisms are directed at the descriptive method of DSM that keeps the situation for etiological neutrality. Certainly, the significance and validity of diagnoses must be proved only after a group with one specific diagnosis is proven to correspond to a specific pathophysiology, clinical course, and reactivity to the treatment. However, this criticism should not be directed to DSM but to psychiatry itself. Conventional psychiatry has not yet had any standardized diagnostic system that is common among schools and countries. This means there has not been any diagnostic system that had enough validity for everyone to agree upon. Each category of DSM should be regarded as a trial-and-error system to establish more useful categories by accumulating empirical evidence and examining validity.