Determining the Cut-Off Points of the Self-Rated Health Visual Analogue Scale for Patients with Depression in Greece

Introduction: Depressed individuals usually rate their health more poorly than non-depressed ones. Aim: Because little is known regarding the correspondence between the scales used to measure self-rated health in patients with depression, the objective of this study is to determine the cut-off points of the Visual Analogue scale that indicate bad, moderate and good health for patients with depression. Methods: For the purpose of this study, data from a Panhellenic cross-sectional survey were used. The survey was conducted in 2017 and used stratified random sampling. The study focused on 71 patients with depression. The respondents were asked to rate their health on a 5-point Likert scale and in a Visual Analogue scale. In order to determine the cut-off points of the Visual Analogue scale, an ordinal logistic model was applied. The dependent variable was the Likert scale and the independent variable was the Visual Analogue scale. In addition, a multinomial logistic model was applied and the 33.3rd and 66.6th percentiles of the Visual Analogue scale were calculated. Results: According to the ordinal logistic regression model, the cut-off points of the Visual Analogue scale are 24 and 76. In addition, according to the multinomial logistic regression model, the cut-off points are 21 and 77. The cut-off points that correspond to the 33.3rd and 66.6th percentiles of the Visual Analogue scale are 50 and 70, respectively. Finally, the Gwet’s AC2s of the regression methods were found to be significantly higher than the percentiles’ method. Conclusions: The results of this study confirm international bibliography in the sense that depression is positively related to poor perception of health. Because the cut-off point of poor health, which is based on the percentiles method, is relatively high, we argue that the percentiles method is inappropriate. This conclusion is also derived from the Gwet’s AC2s’ comparison.