Subjective cognitive decline (SCD) has been proposed as a marker of neurodegeneration in cognitively normal elderly. This idea is supported by the growing evidence that SCD is associated with Alzheimer's disease (AD) biomarkers and increases the risk of future cognitive impairment. Nevertheless, this evidence is not complete, since other studies have not found these associations. This discrepancy could have a methodological basis. It is well known that across the broad spectrum of degenerative disease from healthy controls to dementia, the research setting affects key characteristics of the sample such as age, educational level, or family history of dementia. However, virtually no studies have specifically tested the influence of sampling and recruitment methods in SCD research. Population-based samples are less biased and therefore they probably are more suitable for the study of memory complaints as a symptom at the population level. On the other hand, the memory clinic setting could introduce a set of biases that make these patients more likely to develop cognitive impairment. Thus, memory clinic would be the most cost-effective context in which to study the phenomenology of SCD due to AD and eventually recruit patients for secondary prevention trials. However, this general hypothesis needs to be tested. Studies that compare samples of patients with SCD from different settings are necessary. Sometimes it is difficult for patients with subtle forms of cognitive impairment to access specialized diagnostic centers. Based in our experience we state that Open House type initiatives may be useful for attracting these individuals to memory clinics.
Keywords: Alzheimer’s disease; epidemiologic research design; epidemiologic studies; patient selection; risk; sampling; selection bias; subjective cognitive decline.