Can the Distress Thermometer be improved by additional mood domains? Part I. Initial validation of the Emotion Thermometers tool.

PURPOSE: To examine the value of a new screening instrument in a visual-analogue format.
METHODS: We report the design and validation of a new five-dimensional tool called the Emotion Thermometers (ET). This is a combination of five visual-analogue scales in the form of four predictor domains (distress, anxiety, depression, anger) and one outcome domain (need for help). Between March and August 2007, 130 patients attending the chemotherapy suite for their first chemotherapy treatment were asked to complete several questionnaires with validation for distress, anxiety and depression.
RESULTS: Of 81 with low distress on the Distress Thermometer (DT), 51% recorded emotional difficulties on the new ET tool, suggesting added value beyond the DT alone. Of those with a broadly defined emotional complication, 93.3% could be identified using the Anxiety Thermometer (AnxT) alone, compared with 54.4% who would be recognized using the DT alone. Using a cut-off of 3v4 on all thermometers against the total Hospital Anxiety and Depression Scale (HADS) score (cut-off 14v15), the optimal thermometer was the Anger Thermometer (sensitivity 61%, specificity 92%). Against HADS anxiety scale, the optimal thermometer was AnxT (sensitivity 92%, specificity 61%) and against the HADS depression scale, the optimal thermometer was the Depression Thermometer (DepT; sensitivity 60%, specificity 78%). Finally, against DSM-IV major depression, the optimal thermometer was the DepT (sensitivity 80%, specificity 79%). Further improvements may be possible by using a combination of thermometers or by repeating the screen.
CONCLUSION: The diagnostic accuracy of the DT can be improved by the inclusion of simple addition linear domains without substantially increasing the time needed to apply the test.