Dementia (CSI-D) ฉบับภาษาไทย The diagnostic accuracy of the Thai version of Brief Community Screening Instrument for Dementia (CSI-D)

Abstract

          Objectives   To validate the CSI-D Thai version in the clinical settings against other screening tests such as IQCODE, clock drawing test and standard neuropsychological tests i.e., ADAS-COG, Trail Making Test (TM) among Thai dementia patients.

          Materials and methods   There were 3 steps of study namely; first, to translate CSI-D into Thai. Secondly, neuropsychological tests such as Boston Thai naming card (BT), word list (WL) were prepared and then clinical evaluations and neuropsychological battery tests were conducted at Srithanya and Suanprung Hospital during October 2012 - February 2015. Descriptive statistical analysis was used. The receiver operating analysis curve (ROC) was applied to compare the area under curves (AUC) and the cut-off scores of each test were calculated.

          Results   There were 147 cases divided into 54 control, 33 early dementia, 37 mild dementia and 23 moderate to severe dementia cases. The AUCs of CSI-D among these cases range from

0.59- 0.83. The cut-off score was ≥7 with a fair validity (sense = 59.6, spec = 68.5). The other tests such as ADAS-COG, BT, WL, TM were in details.

          Conclusion  The CSI-D was validated in dementia patients at the psychiatric hospital and the accuracy was comparable to other screening tests. However, the new screening tool for each level such as primary, secondary level should be further developed.