SAGE: A short test for detecting Alzheimer's?

Contributed by: Dennis Fortier, President, Medical Care Corporation
The Self-Administered Gerocognitive Examination (SAGE) has received a fair amount of attention this year. Most of the attention has been driven by PR efforts and a website hosted by The Ohio State University where the test was developed.

Unfortunately, much of the press coverage has been somewhat sloppy and perhaps misleading. While the scientists who developed SAGE described it as "a test to measure thinking abilities", the press has repeatedly characterized it as an Alzheimer's test. As readers of this blog know, dozens of medical conditions can impair memory and other thinking abilities, Alzheimer's disease is merely one of them.

Press mis-characterizations aside, the data supporting the validity of the SAGE is quite minimal.  The instrument, a 15-minute self-administered questionnaire, was developed based on responses from a small sample of 63 subjects, and then validated on a sample of 1,047 subjects.

The website and support materials for the SAGE suggest that the instrument can detect dementia with fairly high accuracy of 95%. This has been the root claim for much of the ensuing press coverage. Remember though, the definition of dementia includes "impairment severe enough to interfere with social and occupational function". In this regard, a test that can detect such severe impairment may have little value, especially if it cannot detect more subtle symptoms such as those we see in the earliest stages of Alzheimer's disease.

From a clinical point of view, it is much more important to detect subtle stages of decline called Mild Cognitive Impairment (MCI). MCI is caused by a great many medical conditions and is often confused with normal, age-related changes in cognition. In fact, incorrectly assuming that subtle changes are caused by normal aging, rather than by an emerging medical condition, is the primary foe in our efforts to intervene at early stages against the various conditions that impair memory. Distinguishing the two is key to proactive management of cognitive health. According to the instrument's website, the SAGE is 62% sensitive for detecting MCI, which will provide little help for physicians trying to distinguish MCI from normal aging.

Other well-validated instruments are more useful in detecting MCI.  The most accurate test in the published literature is the MCI Screen.  It is 97% accurate in distinguishing MCI from normal aging, but it is not self-administered. Rather, it is generally administered by medical staff or a researcher. The ideal would be an instrument with the self-administering ease of the SAGE and the accuracy of the MCI Screen.