We examined issues of criterion validity and detection of depression employing the Zung Self-Rating Depression Scale (ZSDS) as a "lab test" to trigger follow-up interviews of ambulatory oncology patients by oncology staff and the possibility of subsequent algorithm-based antidepressant treatment. Sixty oncology patients were screened with the ZSDS and then interviewed using the Mini-International Neuropsychiatric Interview (MINI). We examined the sensitivity and specificity of various cutoffs on the ZSDS and a briefer version, the Brief Zung Self-Rating Depression Scale (BZSDS) as they predicted results of the MINI, which was used as the criterion. Mean age of patients was 58.3 years (SD = 11.9). Thirty-two were female (53.3%) and 28 were male (46.7%). The correlation of the ZSDS (r = -0.66, P <.0001) and BZSDS (r = -0.57, P <.0001) with the MINI overall suggested acceptable levels of criterion validity. Additionally, we examined various cutoff scores on the ZSDS and BZSDS to explore the false negative and false positive rates that are associated with each. For example, using the mild cutoff on the Zung (score > 48) to determine depression or adjustment disorder, 14 false negatives and 2 false positives were found. When the more stringent moderate cutoff (score > 56) was used, 25 false negatives and 0 false positives were found. Oncology staff can utilize such data to make decisions about where to set cut-offs that trigger follow-up based on the amount of error that is allowable in their attempts to identify depressive symptoms in their patients. We discuss that such decisions might be based on many factors including the resources available in a particular site for follow-up or the comfort of particular oncologists and nurses managing and prescribing psychotropic medications, or in providing supportive counseling.