Second, results are affected by the instrument used to measure depression. Van Gorp's group conducted its own study (unpublished), that looks at elderly individuals with syndromal depression.
"We looked at people who were either high or low on the Beck Depression Inventory [BDI] and compared their performance on a battery of neurocognitive tests. We found numerous differences on measures of memory, psychomotor speed and attention," he said.
When they repeated the study using the same subjects assessed with the Geriatric Depression Scale (GDS), which asks patients their view about themselves and their future, the researchers found absolutely no relationship between the measure of depression and cognitive test performance.
Third, sensitivity and breadth of examination are important.
"The study that tends to separate dementia and depression based on some of the more superficial mental status exams is in for a lot of trouble because, in a sense, we get what we pay for," Van Gorp said. "You can't administer a five-minute screening test to examine a complex question. I think that clinicians who don't work in this area have the illusion that it's an easier question than it is."
Van Gorp cited a recent meta-analytic study that compared depressed subjects who had nonneurologic illnesses and looked at how depression impacts cognition. This study was superior to other meta-analytic studies, he said, because only patients with non-CNS illnesses were used, and because the size of the effect and the variability across studies were factored in. The study offered an empirically based opinion on how depression affects cognition. He described some of the results.
Using inattentiveness as a key index of depression is a mistake because patients with depression did not largely have attentional disturbance. "This goes against the convention that depressed patients are uniformly inattentive," he said.
Verbal fluency is a slightly better indicator when trying to separate depression from normal aging; 10% of depressed subjects were in the clinically impaired range as compared with control subjects. Verbal learning-the acquisition of new information-was impacted in 14.5% of the depressed patients as compared to control subjects. In nonverbal learning acquisition, about 15% were impaired.