We read with interest the recent article by Fisher et al. (1) that concluded that subthreshold de-pression could be regarded as a proxy marker of diabetes-related distress rather than a proxy marker of clinical depression. Fisher et al. support their conclusion by referring to two key findings of their study. Their first finding was a markedly higher correlation between diabetes-related distress and the intensity of depressive symptoms than that between diabetes-related distress and clinical depression. Their second finding was significant associations between subthreshold depression and both poor glycemic control and diabetes-related behaviors (but these risk factors were unrelated to clinical depression). Our comments regarding this study address recommendations for screening emotional problems like depression and diabetes-related distress.
In our own study (2), we showed that the assessment of diabetes-related distress using the Problem Areas of Diabetes (PAID) scale has two key advantages. First, of 53 diabetic patients with clinical depression, 43 could be correctly identified by assessing diabetes-related distress (sensitivity 81.1%). The PAID scale also demonstrated good screening performance for detecting subclinical depression, as 99 of 124 diabetic patients with subclinical or clinical depression were correctly identified (sensitivity 79.0%). Second, the use of depression questionnaires to identify diabetic patients with a high diabetes-related distress (PAID score >40) resulted in a rather poor screening performance of both the Beck Depression Inventory (which correctly identified 70 of 116 patients; sensitivity 60.3%) and the Center for Epidemiological Studies Depression Scale (which correctly identified 60 of 116 patients; sensitivity 49.1%). Thus, diabetes-related distress used as a screener for clinical and subclinical depression demonstrated a reasonable screening performance comparable to that of depression questionnaires; however, the use of depression questionnaires to identify patients with high diabetes-related distress would have missed 40–50% of diabetic patients burdened with a high amount of diabetes-specific distress.
Another argument could be derived from intervention studies. Diabetes education, which can be regarded as a method to increase patients’ ability to cope with diabetes-related distress, has proved to markedly reduce the proportion of subclinically depressed diabetic patients (3). This outcome would also support the findings of Fisher et al. (1) that subclinical depression appears to be more closely related to diabetes-related distress than clinical depression.
Thus, because multiple screening in clinical practice is hardly feasible, it seems reasonable to use diabetes-related distress measures only to screen for depression in diabetic patients and to assess diabetes-related strain. Furthermore, asking patients about diabetes-related distress might be better suited to meeting patients’ expectations regarding seeking treatment for diabetes.