Neuropsychopharmacology: The Interface Between Genes and Psychiatric Nosology

By Thomas A. Ban


8. Diagnostic Instruments for Research

There are two diagnostic instruments specially devised to provide more homogenous populations for research than the diagnostic categories of consensus-based classifications: the Diagnostic Criteria for Research Budapest-Nashville (DCR) and the Composite Diagnostic Evaluation (CODE) System.


8.1 DCR Budapest-Nashville

The DCR Budapest-Nashville (Pethö, Ban, et al. 1984, 1988) is an "eclectic classification" devised on the basis of theoretical considerations. It is an integration of nosologic contributions from different schools of psychiatry into a classification in which the original diagnostic concepts are retained.

At the core of the DCR is Leonhard's (1957) classification of "endogenous psychoses." However, the DCR also includes the Scandinavian diagnostic concept of "psychogenic reactive psychoses" (Stromgren 1974; Wimmer 1916) and a composite of the Cerman diagnostic concept of "delusional development" (Gaupp 1914; Kretschmer 1927) and the French diagnostic concept of "delusional psychoses" (Baruk 1974; Magnan 1893).

The decision to adopt Leonhard's (1957) classification of "endogenous psychoses" in the DCR was based on findings in epidemiological genetic and psychopharmacological research which were supportive of Leonhard's (1957) diagnostic concepts (Ban 1990; Ban and Udabe 1995). The incidence of "nosologic homotypy" was high, from 57% to 77% in the parents and siblings of patients with "cycloid psychoses" (Perris 1974a; Ungvari 1985); and was as high as 80% in twin pairs concordance for "polarity" concordant for "mood disorders" (Tsuang and Vandermey 1980; Zerbin-Rudin 1969). Morbidity risk for "endogenous psychoses" was higher in the relatives (parents and siblings) of patients with "bipolar affective psychoses" than in the relatives of patients with "unipolar phasic psychoses," or "eycloid psychoses"; with "unsystematic schizophrenias" than with "systematic schizophrenias" or "cycloid psychoses"; and with "cycloid psychoses" than with "systematic schizophrenias" (Perris 1974b; Trostorff 1968, 1975). Responsiveness to neuroleptic treatment was higher in the "unsystematic schizophrenias" (79%) than in the "systematic schizophrenias" (22%) (Astrup 1959; Fish 1964). The different forms (and subforms) of schizophrenia in Leonhard's (1957) classification were to become suitable endpoints for molecular genetic research, yielding to the demonstration that "periodic catatonia," one of the three forms of "unsystematic schizophrenia," is associated with a "major disease locus" that "maps" to chromosome 15915 (Stober et al. 2000).

The diagnostic process in the DCR is based on a decision tree model that consists of 524 variables organized into 179 diagnostic decision clusters yielding a total of 179 (undifferentiated: 11, atypical: 37, tentative: 21, provisional: 44, working: 45 and final: 55) diagnoses. An undifferentiated diagnosis in the DCR implies that patient qualifies for "psychosis," but does not meet any of the DCR diagnoses; and an atypical diagnosis implies that he/she fulfills only cross-sectional diagnostic criteria of a specific DCR diagnosis.

8.2 CODE System

The CODE System provides a methodology for the detection of the forms of mental illness identified in the different nosologies which are biologically the most homogenous. It is a set of diagnostic instruments which can assign simultaneously a diagnosis from several diagnostic systems to a patient by specially devised algorithms. Each instrument ("CODE") can provide for a polydiagnostic evaluation in a distinct category of mental illness by the employment of an integrated criteria list and standardized data collection. To achieve its purpose, each "CODE" consists of a set of symptoms ("codes") which yield diagnoses in all the component diagnostic systems; a semi-structured interview, suitable for the elicitation of all the symptoms in terms of "present" or "absent"; and diagnostic decision trees which organize symptoms into distinct psychiatric disorders (Ban 1991). The CODE System differs from other polydiagnostic systems by the inclusion of all distinct diagnostic formulations relevant to the conceptual development of a diagnostic category and by its capability to provide readily accessible information relevant to the diagnostic process from the lowest to the highest level of decision making.

The prototype of the CODE System is CODE-DD, the CODE for "unipolar depressive disorders" (Ban 1989). It consists of a 90- item Rating Scale for Depressive Diagnoses (RSDD) with a 40-item subscale, the Rating Scale for the Assessment of Severity in Depressive Disorders; a Semi-Structured Interview for Depressive Disorders, suitable for the elicitation of the presence or absence of the 90 variables of the RSDD; and decision trees which provide diagnoses within 25 different classifications of depression. Many of the classifications of depression included in CODE-DD are empirically derived, e.g., Kiloh and Garside´s (1963), Winokur's (1979). Some are based on the conceptual development of depression in Europe, e.g., Schneider's (1920), and others on the conceptual development of depression in North America, e.g., Robins and Guze (1972). Included also in CODE-DD are consensus-based classifications, e.g., DSM-III-R of the American Psychiatric Association (1987), and diagnostic criteria for research, e.g., Feighner et al. (1972).

One would expect low inter-rater agreement in such a complex system like CODE-DD. However, in the first reliability study there was an 87.8% inter-rater agreement (regarding the presence or absence of) the 90 RSDD variables (Morey 1991). In the second, agreement increased to 100% (Ban et al., 1993). In a validation study which included 230 patients with a clinical diagnosis of major depression, there was a 99.6% correspondence between the clinical DSM-III-R and the CODE-DD diagnosis of major depression. In another validation study which included 322 patients, the correspondence was 97.2% (Ban 1992).