Purpose: To examine the added value of an algorithmic combination of visual-analogue thermometers compared with the Distress Thermometer (DT) when attempting to detect depression, anxiety or distress in early cancer.
Methods: We report Classification and Regression Tree and logistic regression analyses of the new five-domain Emotion Thermometers tool. This is a combination of five visual-analogue scales in the form of four mood domains (distress, anxiety, depression, anger) as well as need for help. 130 patients attending for their first chemotherapy treatment were assessed. We calculated optimal accuracy for each domain alone and in combination against several criterion standards.
Results: When attempting to diagnose depression the Depression Thermometer (DepT) used alone was the optimal approach, but when attempting to detect broadly defined distress or anxiety then a combination of thermometers was most accurate. The DepT was significantly more accurate in detecting depression than the DT. For broadly defined distress a combination of depression, anger and help thermometers was more accurate than the DT alone. For anxiety, while the anxiety thermometer (AnxT) improves upon the DT alone, a combination of the DepT and AnxT are optimal. In each case the optimal strategy allowed the detection of at least an additional 9% of individuals. However, combinations are more laborious to score. In settings where the simplest possible option is preferred the most accurate single thermometer might be preferable as a first stage assessment.
Conclusion: The DT can be improved by specific combinations of simple thermometers that incorporate depression, anxiety, anger and help.