The Use of GDS-15 in Detecting MDD: A Comparison Between Residents in a Thai Long-Term Care Home and Geriatric Outpatients
Abstract
Methods: This was a cross-sectional study of 156 geriatric outpatients and 81 LTC home residents. All 237 participants were given a Mini-Mental State Examination, a MDD diagnosis according to the Mini-International Neuropsychiatric Interview, and completed a TGDS-15 questionnaire. Sensitivity, specificity, overall accuracy, and positive and negative predictive values were calculated. A comparison between the two groups was carried out. Differential Item Functioning (DIF) using logistic regression and factor analytic study were also applied.
Results: Overall, 38.4% of the participants were found to have MDD. The TGDS-15 was found to perform better when used with the GOs than with the LTC home residents, revealing a sensitivity of 0.92 and a specificity of 0.87 in the GOs (cut-off score of >= 5), but a sensitivity of 100% and a specificity of 49% with the LTC home group (cut-off score of >= 8), when comparing only cognitively intact subjects. The negative predictive value (NPV) was very good for both groups, but the positive predictive value (PPV) for the GO group was much better than for those in the LTC group (83.3% vs. 31.2%). Seven uniform DIF items were found - 2 by gender and 4 by age. Cronbachs alpha was higher for the GO group than for the LTC home residents. Factor analysis supported a two-factor solution, using the depressed mood and positive mood factors, which accounted for 46.55% of the total variance.
Conclusions: The TGDS-15 scale was effective at screening for MDD in elderly cognitively intact Thais, those in both GO and LTC settings, as the sensitivity and NPV were shown to be very good in both groups. However, in the LTC setting, the low specificity and PPV found leads to the need for a further assessment to be carried among the potentially depressed individuals, based on the GDS results. Taking the factor analytic study into account, a more suitable version of the GDS should be developed.
doi: http://dx.doi.org/10.4021/jocmr1239w