TO THE EDITOR: Drs. Swanwick and Wrigley raise a number of important points to which we are pleased to respond. While standard systems such as the Research Diagnostic Criteria may or may not be ideal for elderly depressed patients, an alternative "gold standard" for this population has yet to achieve common acceptance. Alternative criteria have been proposed for elderly patients who are physically ill (R1557291D), but they largely identify the same depressed patients as standard criteria (R1557292D, R1557293D). Hence, we doubt that their application in this particular study would have made much difference. We would, however, encourage additional research to develop and validate appropriate techniques to identify depression in this patient population.

Depressed elderly patients may indeed be less likely than younger patients to complain of low mood. We, in fact, recommended a two-item screening approach (assessing loss of interest as well as depressed mood), since it provides complete coverage of the core criterion symptoms of depression. The same logic is incorporated into most structured interviews, which allow skipping the remaining symptoms if depressed mood and loss of interest are both absent. In our study, however, addition of the loss-of-interest item did not actually improve diagnostic accuracy over the single-item approach. The accuracy of single-item screening in identifying depressive disorders in older adults has also been reported by other investigators (R1557294D).

Finally, our results are unarguably criterion specific. It is worth noting that even with standard criteria, depression among the terminally ill tends to be underdiagnosed (R1557295D). This may reflect, in part, clinician discomfort in probing too deeply into the psychological experiences of dying patients (R1557296D). To suggest that single-item screening diminishes psychiatry's important role in addressing these issues is simply reading too much into our results. We hope our study serves as a gentle reminder to clinicians to ask their dying patients—even in a simple way—about their mood states. Doing so may help increase recognition of depression in this vulnerable population.