Given the findings of our recently published report (1) and their earlier study (2), Hermanns et al. (3) suggest that because screening for multiple conditions is not always feasible, it may make sense to use diabetes distress measures to assess both depression and diabetes-specific distress. Both studies concluded that measures of diabetes distress captured both the negative affect associated with depression and the negative affect associated with diabetes-related distress. Hermanns et al. support their view by presenting cogent arguments that are based on the sensitivity of distress measures in capturing both kinds of negative affect. We used the Center for Epidemiology Depression Scale (CESD) (4) to assess depressive affect and the Diabetes Distress Scale (DDS) (5) to measure diabetes distress. We also used a structured interview, the Composite International Diagnostic Interview (CIDI), to assess clinical depression or major depressive disorder (MDD) (6).
Although we agree with Hermanns et al. in principle, we suggest that the crucial study to support their recommendation has not yet been done and that the empirical question on which their argument is based has not yet been answered. Our study showed that the DDS captured most of the negative affect assessed by the CESD and that most of this affect was linked to behavioral and biological outcomes, not to clinical depression. Furthermore, we showed that the presence or absence of MDD was not linked to these outcomes, either individually or in combination with the other two measures. We argued that MDD, although elevated in our sample when compared with community control subjects, was related to but distinct from the kind of negative affect that was associated with diabetes distress. We also argued that most patients who scored high on the CESD and who experienced high levels of a negative affect were really experiencing the negative affect associated with diabetes distress and not the negative affect experienced with MDD.
Given these findings, we really need to screen for two conditions that occur at high rates in this population: diabetes-related distress and MDD. Unfortunately, we have no data to suggest that the DDS or a similar diabetes distress scale is a tool that effectively screens for both. The biggest current danger in using the DDS alone is obtaining a false-negative finding for MDD. That is, showing a negative screen for distress when the patient meets criteria for MDD, which could easily occur if the DDS does not screen well for MDD. Although a single generic screening instrument would be ideal, we do not yet know the relationship between the DDS and scores on an MDD screener, such as the Patient Health Questionnaire 9 (7). It may be that these two instruments assess very different things or, alternatively, that there is substantial overlap between them such that a negative DDS screen also indicates a negative screen for MDD. At this point in time, we simply do not know the answer to this important question. Findings not reported in our original article that are now available indicate that only 27.8% of patients with MDD, diagnosed by a detailed clinical interview (CIDI) and not a screener, also scored above the cutoff on the DDS. Thus, the DDS was not sensitive to a diagnosis of MDD using these methods of assessment.
A review of the DDS items suggests that its specific diabetes focus may not address many aspects of MDD, which often includes feelings of dysphoria and anhedonia. Indeed, one purpose of the DDS was to use the items to begin a conversation with the patient about their distress during a clinical encounter. Thus, we suspect that a brief integrated four- to five- item screener that includes questions for both distress and MDD may be required, to be followed by a more detailed screen for patients who score positively on either or both. No matter what the outcome of a study that assesses the relationship of diabetes distress and MDD, however, it is important to note that both conditions are considerably elevated among those with diabetes, that both are significantly linked to clinical outcomes and costs, and that both should be addressed as part of comprehensive diabetes care.
