Comparative validity of the WHO-5 Well-Being Index and two-question instrument for screening depressive symptoms in patients with type 2 diabetes

A simple tool consisting of two questions for screening depressive symptoms has been shown to be useful in primary care settings, but its validity in patients with diabetes has yet to be evaluated. We compared the test performance of this two-question instrument with that of WHO (The World Health Organization)-5. We consecutively enrolled 153 patients with type 2 diabetes who visited a diabetes clinic in Japan. Using the Center for Epidemiologic Studies Depression Scale as a reference standard of depressive symptoms, we calculated the sensitivity and specificity of the two-question instrument and WHO-5, and compared the area under the ROC curves of these tests. The two-question instrument had a sensitivity of 53.6% (95% CI, 39.7–67.0%) and specificity of 67.7% (95% CI, 58.1–74.9%). With the conventional cutoff point equal to or less than 13 points, the WHO-5 had a sensitivity of 57.1% (95% CI, 43.2–70.3%) and specificity of 82.5% (95% CI, 81.9–94.9%). The area under the ROC curve for the WHO-5 and two-item questionnaire, an indicator of discriminatory power, was 0.81 and 0.73, respectively, showing a statistically significant difference (P = 0.0453). The two-question instrument had statistically lower discriminatory power than the WHO-5 in screening depressive symptoms in patients with diabetes. We do not recommend the use of the two-question instrument for screening depressive symptoms in patients with diabetes.