Diagnostic disclosure: a tale in two cultures.

Diagnostic communication between doctors and patients is thought to differ radically between Japan and Western countries. To understand diagnostic disclosure to psychiatric patients, a questionnaire with six case vignettes was sent to practising psychiatrists in Japan (N = 166) and North America (N = 112). While over 90% of both groups would inform patients with affective and anxiety disorders of their diagnoses, only 70% of North Americans and less than 30% of Japanese would similarly inform patients with schizophrenia or schizophreniform disorders. The Japanese preferred alternative was to give a vague alternative diagnosis such as neurasthenia. North Americans would discuss differential diagnoses with the patient instead. Nearly all in both groups would inform the family, but North Americans would do so only with patient consent. For disorders for which there are effective treatments, diagnostic disclosure is common to both cultures; when prognosis is uncertain or the diagnosis is feared, as in schizophrenia, culturally constructed views of patienthood govern disclosure practice.

The Longitudinal Interval Follow-up Evaluation. A comprehensive method for assessing outcome in prospective longitudinal studies.

The Longitudinal Interval Follow-up Evaluation (LIFE) is an integrated system for assessing the longitudinal course of psychiatric disorders. It consists of a semistructured interview, an Instruction booklet, a coding sheet, and a set of training materials. An interviewer uses the LIFE to collect detailed psychosocial, psychopathologic, and treatment information for a six-month follow-up interval. The weekly psychopathology measures ("psychiatric status ratings") are ordinal symptom-based scales with categories defined to match the levels of symptoms used in the Research Diagnostic Criteria. The ratings provide a separate, concurrent record of the course of each disorder initially diagnosed in patients or developing during the follow-up. Any DSM-III or Research Diagnostic Criteria disorder can be rated with the LIFE, and any length or number of follow-up intervals can be accommodated. The psychosocial and treatment information is recorded so that these data can be linked temporally to the psychiatric status ratings.

The stability of diagnosis with an application to bipolar II disorder.

A temporal stability study (where independent diagnostic interviews are conducted at widely separated time points) is discussed and compared to a test/retest reliability study. We introduce a new battery of statistics, based on the sensitivity, specificity, and true base rate of a disorder, that quantifies either reliability or stability over time, and provide a table that relates these underlying parameters to the opening characteristics of the battery. We analyze data on 50 relatives who participated in the family study component of the National Institute of Mental Health's collaborative "Psychobiology Depression Program." The subjects received an interview 5 years after their initial evaluation by raters without knowledge of the initial assessment. The stability of mania, hypomania, major depression, and alcoholism is considered. Although the kappa coefficient for hypomania was small (0.09), all diagnoses of hypomania in the relatives occurred in the families of bipolar probands, suggesting that the low value of kappa is due to a low sensitivity rather than to a low specificity. This is compatible with earlier findings and demonstrates the value of multiple independent assessments when studying this disorder.

Informed versus blind: the reliability of cross-sectional ratings of psychopathology.

Fifty subjects at the New York facility of the National Institute of Mental Health Clinical Research Branch Collaborative Program on the Psychobiology of Depression were evaluated for cross-sectional psychopathology and functional impairment at the 2-year followup by two professional raters, one blind and one with extensive knowledge of the subject's psychiatric history and course of illness and treatment (the cohort rater). Ratings were made on the change version of the Schedule for Affective Disorders and Schizophrenia. Interrater reliability, as tested using the intraclass correlation coefficient, was quite good for most of the symptom ratings and fair on measures of functional impairment and social functioning. However, there was a consistent tendency for the blind rater to score higher levels on symptoms of depression and anxiety than the informed cohort rater. Reasons for disagreement in different symptom areas are discussed. Using only the informed rater provides a very adequate picture of severity and patterns of psychopathology on a cross-sectional basis.

Assessing personality: effects of the depressive state on trait measurement.

The influence of the clinically depressed state on personality assessment was evaluated by comparing self-report personality inventories of patients while clinically depressed and at follow-up 1 year later. The authors examined two groups from the National Institute of Mental Health (NIMH)-Clinical Research Branch Collaborative Program on the Psychobiology of Depression: Clinical Studies--patients whose symptoms had completely remitted and those who had not recovered. The clinically depressed state strongly influenced assessment of emotional strength, interpersonal dependency, and extraversion. Assessment of rigidity, level of activity, and dominance did not change after symptomatic recovery.

The reliability of retrospective treatment reports.

High levels of agreement are reported for retrospective assessment of treatment history using the Longitudinal Interval Followup Evaluation Baseline (LIFE Base), a recently developed structured psychiatric interview, in a sample of 47 subjects with a moderate to severe affective disorder. In a paired rater/observer design with interviews conducted at five different research centers, most coefficients of reliability (Kappa) were above 0.91. This is the first published reliability study of treatment data and should be encouraging to practitioners and researchers. Future research aimed at assessing the validity of treatment information obtained from patient records and clinical interviews is recommended.

Assessment of reliability in multicenter collaborative research with a videotape approach.

The authors, as part of the ongoing NIMH Collaborative Study on the Psychobiology of Depression, used an analysis of variance design and videotaped interviews to explore the effects of sources of variance on the reliability of the measures being used by the NIMH study. In spite of substantial differences among interviewers in background or orientation, the authors found that diagnoses and symptom ratings were made with a high level of reliability. These results suggest that the use of structured interviews and diagnostic criteria, when combined with a careful and systematic training program, can lead to good levels of diagnostic reliability.

Mastering the art of research interviewing. A model training procedure for diagnostic evaluation.

A training program for researchers using standard clinical assessment procedures has been developed to enhance comparability of findings across studies through reducing critical sources of unreliability. The need for such a program is especially important in large multifacility collaborative studies in which the pooling of diagnostic and rating-scale data is planned. The program described here focuses on the Schedule for Affective Disorders and Schizophrenia and Research Diagnostic Criteria, two instruments widely used in clinical research. It consists of the following four phases of training: use of written case vignettes, videotaped interviews, in-person training interviews, and ongoing monitoring to maintain reliability. Checks on reliability are built into each successive phase of training. This program can serve as a model for training in the use of other standard interview assessment procedures requiring clinical judgment.

Reliability of lifetime diagnosis. A multicenter collaborative perspective.

It is important to determine the reliability of lifetime diagnosis in a nonpatient population, for this type of diagnostic data and this type of sample are used in many genetic, epidemiological, and nosological studies. We examined the reliability of lifetime diagnosis when the Schedule for Affective Disorders and Schizophrenia-Lifetime Version and Research Diagnostic Criteria were used to interview ill and well relatives of probands in the National Institute of Mental Health Collaborative Study of the Psychobiology of Depression. Subjects were interviewed three times, so data are available concerning both short- and long-interval test-retest reliability. Short-interval test-retest reliability was excellent for both diagnoses and symptoms. Reliability was also quite high in the long-interval test-retest study. We conclude that it is possible to make lifetime diagnoses reliably in a nonpatient population.

Reliability studies of psychiatric diagnosis. Theory and practice.

The existing literature on the reliability of psychiatric diagnosis falls into two periods, the earlier reporting low reliability and the latter reporting much higher figures. The reasons for this trend are examined in the context of a discussion of the design of diagnostic reliability studies. The problems of research design and execution in studies of diagnostic reliability are reviewed, and statistical problems are examined. Solutions to many of these problems ae suggested, including recommendations of appropriate reliability coefficients and data analyses.

Test-retest reliability of assessing psychiatrically ill patients in a multi-center design.

In a test-retest reliability study involving 25 psychiatric patients and 5 professional raters we demonstrate that research clinicians from collaborating institutions are able to achieve good reliability for most areas of the SADS and RDC when assessing psychiatrically ill patients under interview conditions that provide even less data than ideally obtained in the practice of clinical research. We expect greater reliability in the actual use of the SADS/RDC on most items and diagnoses since the SADS is intended to be used in conjunction with information obtained from relatives, friends, and treatment staff to confirm and clarify the judgements made by the raters on the patient interviews. Moreover, we are reassured that the diagnosis of schizo-affective disorders and schizophrenia is protected from the item unreliability found with specific delusions and hallucinations. Similarly, the difficulties in determining the episodic and chronic nature of the present episode does not substantially interfere with making an RDC diagnosis of the current condition. A complex diagnostic interview system such as the SADS and RDC requires multiple complementary techniques to determine reliability. We find that establishing explicit procedures for raters to discuss and categorize the reasons for their disagreements on individual items and diagnoses provides valuable data for understanding reliability problems. This has helped us to identify specific areas of the interview and criteria that require further clarification and more intensive rater training to improve ratings made by interviewers.

Reliability of lifetime diagnoses and symptoms in patients with a current psychiatric disorder.

In this study we assess the reliability of rating past psychiatric symptoms and lifetime diagnoses in a currently ill population using the SADS and RDC. Five raters from different centers interviewed 25 subjects in a short-interval test-retest design. Subjects had a wide diversity of affective and non-affective diagnoses and high levels of manifest psychopathology. Our results demonstrate that it is possible for raters from different research centers to reliably rate lifetime diagnoses and previous symptoms. Two important exceptions to the high reliability are the lifetime diagnoses of hypomania and the recurrent unipolar subtype of major depressive disorder, and we alert clinicians and researchers to be cautious when diagnosing these conditions. We conclude that the next step in measuring reliability should be a long-interval test-retest design with separate interviews conducted at the later assessment, one by the original rater and the other by a blind rater.