COMPUTER PROCEDURES FOR PSYCHIATRIC CLASSIFICATION

A Review of Psychiatric Rating Scales

Considerable attention has been paid during the past twenty years, mainly by research workers in the U.S.A., to the employment of rating scales for the assessment and recording of those characteristics of psychiatric patients which are generally classed as signs and symptoms.

As well as being used in the practical problems of diagnostic evaluation of patients, the rating scales are often employed in research endeavours in the area of psychiatric nosology. Their employment and the methodological approach which they represent might well be regarded as complementary to other sorts of activities and interpretative philosophies which might be encountered in any particular psychiatric facility, including those in which comparatively less emphasis may be laid upon considerations of diagnostic differentiation.

Rating scales are relatively objective methods of appraisal at least of certain observable aspects of personality, such as overt situational behaviour and appearance, as well as some aspects of the patients' verbal communication. In employing these scales, such personal characteristics are systematically reviewed; usually quantitatively graded; and are recorded in recoverable form. Immediate interpretation of the observed phenomena is not called for. The observations are made in a more or less standardized fashion, and more than one rater may participate in order to minimize subjective differences of appraisal. In this way, then, real life behaviour in a particular situation is assessed, in contrast to the approach involved in some of the more traditional ‘test’ methods, in which the rationale entails that certain aspects of personality are implied symbolically in the test scores or protocol. However, in so far, as certain of the scales may involve some self-report on the part of patients regarding subjective feelings and symptoms, there is clearly some affinity with certain of the older personality inventories, questionnaires, symtom check-lists and the like.

The factor analytic treatment of data derived from the use of psychiatric rating scales to assess some of the phenomena of psychopathology has led to fresh suggestive principles of classifying these. From various studies a number of dimensions have emerged in distinction from the usual disease entities though it is possible to make assimilatory links between these two sets, as, for example, Lorr (26) has done. While the nature and number of the ‘factors’ produced are dependent upon the particular patient population studied, several surveys have resulted in some measure of concordance as may be seen in the three lists of factors shown in Table 1.

The various scales themselves may be typed according to the extent of involvement of the patient; the areas covered by the particular scale; and the sort of patients for whom it is intended. In this way, the reader will be able to select scales which are appropriate for any particular purpose, such as routine observation and assessment, or research connected with changes of signs and symptoms.

Improved statistical classification methods in computerized psychiatric diagnosis

BACKGROUND

Mainstream psychiatric diagnosis involves mainly sequential, expert-system-derived, logical decision rules. Among the few statistical classification methods that have been sporadically evaluated are Bayes, k-nearest neighbor, and discriminant analysis classifiers.

METHODS

A statistical classification method based on artificial neural networks (ANN) with task-specific constrained architectures was applied to a sample of 796 clinical interviews, where the symptom evaluation and the diagnostic judgments were made using the Psychiatric State Examination (PSE) system. The proposed constrained ANN (CANN) method was compared with other statistical classification methods.

RESULTS

CANN was found to be superior to all other considered methods, having an overall "correct" classification rate of 80% when applied to test data. Similarly, the concordance coefficients of agreement with the PSE diagnostic categories were all very high. Among the other used methods, discriminant analysis had slightly inferior performance but better generalization capability.

CONCLUSIONS

The proposed CANN method has a definite utility in psychiatric diagnosis and requires further evaluation, perhaps alongside other standard classification systems and/or with larger samples.

Computerized psychiatric diagnoses based on euclidean distances: a Chinese example

Current diagnostic methods in psychiatry use sequential logical decision rules that generate a single diagnosis. Insufficient attention has been paid to parallel methods that can simultaneously determine the relative probability of many diagnoses. This study installed 45 items from various symptom scales on a portable computer and applied a euclidean distance formula to generate immediate diagnoses based on responses to the items. The reliability and validity of the method were assessed using Chinese psychiatric inpatients. Interrater reliability was excellent (kappa = 0.91) and 3-week test-retest reliability was fair (k = 0.50). The concordance of this method with clinicians' diagnoses and with diagnoses based on standardized Chinese diagnostic criteria was excellent (k = 0.73 and 0.76). Concordance with DSM-III-R diagnoses and ICD-10 diagnoses was fair (kappa = 0.55 and 0.65). The clinical utility of such parallel methods of psychiatric diagnosis deserves further evaluation.

Expert systems in psychiatry. A review

Existing computer-based decision aids in the areas of psychiatric diagnosis and consultation are reviewed, and the prospects for expert system development within the mental health field are discussed. Emphasis is placed upon the decision-making models used in these systems rather than on their particular application area. The decision-making paradigms discussed are (1) data bank analysis, (2) statistical pattern recognition, (3) Bayesian analysis, (4) logical flow chart method, and (5) knowledge-based (expert system) approaches. For each paradigm, its essential features, its strengths and weaknesses, and some example applications are presented.

Clinical computer applications in mental health

Direct patient-computer interviews were among the earliest applications of computing in medicine. Yet patient interviewing and other clinical applications have lagged behind fiscal, administrative, and research uses. Several reasons for delays in the development and implementation of clinical computing programs are discussed. Patient interviewing, clinician consultation, and other applications of clinical computing in mental health are reviewed, as well as changes that will facilitate their appropriate use.

Computer-aided medical diagnosis: literature review

The difficulty of the medical diagnostic task and the advantages of the computer as an aid in this task are discussed. The general strategy and structure of any computer-aided system is presented, and the relationship of diagnostic accuracy to key variables involved in the development, test and use of a computer-aided diagnostic system is examined. These variables include: the computer algorithm, the source of the information used to develop the data base, the number and type of diseases under investigation, the number and type of indicants used, the source of the test sample, and the source of the validated diagnosis. A table of 58 empirically tested computer-aided medical diagnostic systems is presented; each system is summarised in relation to the variables mentioned above and diagnostic accuracy.

Profile stability of a clinical rating scale

A statistic for describing the reliability of clinical profiles is described. Videotaped interviews are utilized to hold day-to-day variability constant, permitting the evaluation of rater-specific errors. The stability of clinicians' conceptions of the pathology associated with “typical” patients representing various diagnostic categories was evaluated. Even when rater-specific variance is averaged out, patients' observed pathology does not match well with the profile of pathology ascribed to the diagnostic group to which the patient has been clinically assigned.

Doctors as decision-makers: a computer-assisted study of diagnosis as a cognitive skill

When viewed as a sequence of decisions clinical diagnosis becomes amenable to detailed investigation in terms of standard statistical concepts. A study of six clinicians diagnosing identical sets of cases of non-toxic goitre is used to illustrate an objective technique for studying the diagnostic process with the aid of a digital computer. Considerable variation in clinicians' routes to correct diagnosis is shown when these routes are compared in detail by five statistical measures related to the effective use of the information available to the clinicians. For rapid analysis of diagnostic skill two visual methods are presented. These can be developed for teaching undergraduates the interpretative skills involved in diagnosis and for studying such skills in experienced clinicians.