Clinician Word Use in Dementia Evaluation Reports as a Function of Cognitive Impairment.

Objectives: To examine the extent to which levels of cognitive status influence patterns of word use in dementia evaluation reports. Methods: We utilized neuropsychological evaluation reports from 61 geriatric primary care patients referred for suspected dementia. Linguistic Inquiry Word Count analysis was utilized to examine clinician language use in patient reports and whether language use differs dependent on the diagnosis rendered. ANOVA analyses were used to analyze group differences in LIWC word counts across clinical indices of cognitive functioning: dementia diagnosis. Results: Our analysis revealed significant differences in language use across diagnostic categories. ANOVA analyses yielded differences in broad negative emotion, F(2,58) = 4.010, p = .023 as well as other subgroups; anxiety-related word groups, F(2,58) = 4.706, p = .013; insight words, F(2,58) = 3.815, p = .028; causation words, F(2,58) = 3.497, p = .037; certainty words, F(2,58) = 6.581, p = .003; negation words, F(2,58) = 3.165, p = .05; time-related words; F(2, 58) = 7.521, p < .001; and human-related words, F(2,58) = 6.512, p = .003. Conclusion: The differences in clinician language use across different diagnostic groups may be reflections of implicit emotional reactions. Many of the patterns found in this study can be linked to previous research concerning word use and underlying thought processes. Clinical Implications: Awareness of language use is helpful in clinical relationships to attenuate stigma and facilitate treatment and research.

Psychometric Evaluation of the Symptoms of Dementia Screener (SDS) in a Geriatric Primary Care Population.

Objectives: The Symptoms of Dementia Screener (SDS) is an 11-item scale developed to screen for cognitive impairment. We aim to evaluate the psychometric properties of the SDS for use in primary care. Methods: We analyzed data from 192 patients: 25 not impaired, 42 with mild cognitive impairment (MCI), and 125 with dementia. Cronbach's reliability, convergent validity, and clinical utility were examined. The SDS was investigated at the item level using binary two-parameter model item response theory (IRT) techniques. Results: The SDS demonstrated good reliability (11 items; α = .74). We found negative correlations between SDS and the Mattis DRS-2 (r = -.523, p < .01). Receiver operating characteristic (ROC) curves demonstrated acceptable clinical utility for detecting MCI and dementia with sensitivities and specificities of 83% and 52% for MCI; 78.4% and 84% for dementia, and 91% and 52% for any impairment. IRT analyses revealed 10 out of 11 items were moderately to very highly related to underlying latent factors of impairment. Conclusions: The SDS demonstrates good psychometric properties and is useful for detecting cognitive impairment in primary care settings. Clinical Implications: The SDS is an effective screening tool that does not require special training for its use in primary care. A positive screen indicates a need for further cognitive testing.