Dear Editor
We enjoyed reading the systematic review and meta-analysis of Castilla-Rilo et al. (2007) concerning IADL in the screening for dementia in population studies. This is an interesting topic because an IADL scale might be a short and simple screening instrument for dementia. However, the evaluation of screening and diagnostic tests often contains biases. In this letter we would like to highlight some of these, using the Castilla-Rilo article as an example.
Instrumental Activities of Daily Living is a widely used, but poorly defined construct. According to the definition of Lawton and Brody (1969) it is included in functional status. Generally, functional status can be divided into basic activities of daily living (BADL) and IADL. BADL contains basic self-maintenance skills and IADL more complex activities of daily living (Lawton and Brody, 1969). In the article of Castilla-Rilo et al. (2007) the terms 'functional assessment' and 'IADL' are used interchangeable, while they are in fact separate terms. Moreover, no clear distinction was made between scales measuring exclusively IADL and scales measuring functional status. As a consequence of including BADL functions in the analysis, the results of this study were distorted.
Another point of interest regards the use of IADL as the index test, and the diagnosis of dementia as the reference test. To diagnose dementia, information from the index test is necessary. Diagnostic criteria to establish the diagnosis, for example DSM-IV or NINDCS/ADRDA, require a decline from a previous level of functioning (APA, 1994). As such, activities of daily living are embedded in the diagnosis and as a result of this incorporation bias, diagnostic accuracy will be overestimated (Delgado-Rodriguez and Llorca, 2004).