The Distress Thermometer in Spanish cancer patients: convergent validity and diagnostic accuracy

Abstract

International organizations such as National Comprehensive Cancer Network and NICE recommend implementation of routine screening programs for detecting and managing psychosocial distress among cancer patients. The selection of an adequate screening tool is crucial to the effectiveness of these programs. The present study examines the emotional symptomatology captured by the Distress Thermometer (DT) and its accuracy and validity as a screening tool in cancer. It also explores the possible discrepancy between patient distress and the use of psycho-oncology resources. A heterogeneous sample of 962 adult cancer patients completed the DT, the Problem List (PL), the Brief Symptom Inventory-18 (BSI-18), and the Psychosocial Questionnaire. The DT was significantly correlated with BSI-18 symptoms and the emotional problems listed on the PL. Receiver Operating Characteristic analysis showed good diagnostic accuracy for the DT (area under the curve = .82, 95 %CI [.79-.85]). For a selected DT cutoff of 5, standard measures (sensitivity = 90 %, specificity = 64 %, predictive positive value = 35 %, and negative predictive value = 97 %) and Clinical Utility Indexes (utility index negative = .62 and utility index positive = .32) indicated that the DT was adequate for "screening" while it was limited for "case finding." Finally, 81.30 % of patients with clinical distress had not sought or were not receiving professional psychosocial support at the time of the study. The DT is appropriate for use as a rapid screening instrument for cancer patients in a Spanish population because it assesses a broad concept of distress including both anxiety and depression symptoms. The diagnostic accuracy of the DT could be improved with minor proposed modifications to the DT and the inclusion of nonemotional ultrashort measures.