 |
 |
 |
|
|
Disordered thinking is a symptom found in many mental disorders, including schizophrenia, mania, depression, obsessive–compulsive disorder, and others. Unfortunately, agreement about operational definitions of disordered thinking has eluded researchers, and the value of various assessment techniques remains controversial (1). In a study of first-lifetime–episode, neuroleptic-naïve subjects with schizophrenia, we found that most had insight that something was wrong with their thinking (2). Previous research has indicated that patients with mental disorders can give subjective reports of their illness; for this reason, we developed a self-report technique to assess disordered thinking (3). Our hope was to develop a questionnaire that might be useful for the early detection of prodromes of mental disorders, such as schizophrenia, or that might predict the course and outcome of certain mental disorders (4). We describe the development of the Thought Disorder Questionnaire (TDQ), including data on reliability and validity, and we explain the questionnaire’s feasibility and diagnostic accuracy (5,6). The TDQ was developed by following the methods outlined in Jackson’s sequential system for personality scale development (7). This procedure emphasizes the importance of the following steps: 1) to begin with psychological theory, 2) to make suppressing response style a necessity, 3) to ensure convergent and discriminant validity, and 4) to scale homogeneity and generalizability. Our hypothesis was that disordered thinking is the fundamental symptom of first-lifetime–episode, neuroleptic-naive schizophrenia. At the same time, such symptomatology does not represent a unitary concept, nor is it necessarily specific to schizophrenia (8). Accordingly, multiple subscales were considered necessary, in part to reflect the multidimensionality of the concept. Second, elevated scale scores certainly were possible among patients other than those diagnosed with schizophrenia. We made an effort to tap an appreciable portion of the domain of disordered thought, or cognitive functioning. We did not attempt, however, to separate patients with schizophrenia from others on the basis of their language structure or on their unique features of positive vs negative symptomatology (8). Moreover, the construction of our measure was not constrained to the customary formal definition of thought disorder used in diagnosis, with its corresponding specific clinical significance. Before any inventory items were generated, specific definitions of each subscale were developed. To assess thinking in the general population, the following 13 subscales were developed and specified by dimensions:
We included 2 additional scales: an infrequency scale designed to control for nonpurposeful responding and a social desirability scale designed to control for response styles, in which subjects choose answers for the purpose of placing themselves in a positive light. Items were generated to cover all aspects or conditions of each subscale. Questions were also developed to reflect both the positive and negative dimension of each item. ResultsPreliminary Analysis We provided an initial version of the TDQ, including 304 items to a 97-subject sample comprising 6 individuals with a diagnosis of schizophrenia and 91 healthy volunteers. To reduce the questionnaire to a more manageable length, items that showed no statistical variance (that is, answered the same way by more than 95% of the sample) were removed from the next version of the questionnaire. By eliminating such items, reliability and validity of the scale were undoubtedly enhanced. Because disordered thinking is hypothesized to exist on a continuum that extends into the healthy population, intuitively, items endorsed by both healthy respondents and respondents with mental disorders, with variations in degree of endorsement, will be expected to yield the most valuable information (9,10). The questionnaire consists of a 4-point rating scale (0 to 4), and the responses for each question are simply totalled for each subject. Table 1 provides sample items from each scale. Initial reliability analysis performed on the entire 304-item scale yielded remarkably high alpha levels, ranging from 0.8141 to 0.9412. In this preliminary analysis, performed on an extremely small sample of subjects, the questionnaire successfully distinguished between schizophrenia patients and healthy respondents. Patients were diagnosed as having schizophrenia, based on 2 present state examinations (PSE), conducted by 2 independent psychiatrists. On 13 of the 15 subscales, inpatients with schizophrenia accrued significantly higher mean scores than did healthy control subjects. This result did not, in itself, prove the scale’s predictive and diagnostic capabilities; however, schizophrenia patients did not attain scores that significantly differed from the scores of healthy control subjects on both the Infrequency and Social Desirability scales. This suggests that patients with schizophrenia were conscientiously answering items, rather than randomly endorsing them. If schizophrenia patients’ responses had differed from those of healthy control subjects on either of these 2 scales, the usefulness of the scale would indeed be suspect. Aside from the promising data that the questionnaire generated, interviews with schizophrenia patients about its value also provided some interesting results. While the entire 304-item version of the questionnaire required approximately 45 minutes to complete, no patients with schizophrenia appeared to have difficulty with completing the items. For the most part, all schizophrenia patients provided very positive feedback on the questionnaire, suggesting, perhaps, that not only is the self-report technique feasible but that it may also provide the opportunity for schizophrenia patients to give a closer approximation of their experiences than other assessment techniques permit.
|
||||||||||||