This article is one of two in JCN by Aselage (2010a,b) on the subject of feeding difficulty in older adults with dementia. Both studies are based on reviews, the first (2010a) being a concept analysis of feeding difficulty in dementia, and this present article considering the psychometrics of instruments designed to measure feeding difficulty in older adults with dementia. This study both delights me and disappoints me in equal measure: my delight stems from the citations to my own work (with Deary and others) on nutritional assessment (Watson & Green 2006) and the Edinburgh Feeding Evaluation in Dementia (EdFED) scale (Watson 1993, 1996, Watson & Deary 1997) and the acknowledgement of its pre-eminence amongst scales for the measurement of feeding difficulty in dementia owing to its extensive psychometric testing (Watson 1997, Watson et al. 2001a,b). My disappointment stems from the paucity of work by others in this field generally and specifically in the area of measurement.

The problems of nutrition in older people, and especially the allegations of abuse and neglect of older people with dementia in hospital with respect to food and starvation, are regularly in the UK press. Solutions are proposed which are usually policy oriented and professionally focused, but very general. Nurses and others in the ‘frontline’ of care of older people with dementia are given few specific tools and strategies to help older people with dementia to eat, and it is self-evident that assessment of feeding difficulty is an essential component of any strategy – but the issue of assessment is rarely addressed.

Aselage’s present paper is an excellent ‘marker’ in the field as it uses a rigorous search of electronic databases to identify potential candidates for inclusion in the review and then applies a rigid framework for comparison and contrast of the 12 instruments identified as worthy of inclusion in the review. The instruments are broadly similar, and all have some psychometric testing carried out. While the EdFED is the one with the most extensive testing of reliability and validity – thus setting it apart from the others – as the main author of this scale I should emphasise that psychometric testing is never complete; the search for construct validity is infinite, and there is one crucial aspect that has yet to be addressed with the EdFED scale and that is its ecological validity: it has been tested in long-term geriatric and psychogeriatric settings, but its utility in acute care remains to be established. There is likely to be a very different spectrum of dementia in acute care, and stays in acute care are shorter with less time to become familiar with patients. The EdFED was developed for staff (primary nurses as they were then described) who were familiar with the people being assessed. Nevertheless, there is some evidence that the scale can be used for assessment by direct observation (Lin & Chan 2003), and also the scale has been translated and validated in Chinese (Lin et al. 2008).

One additional note that I would like to add with regard to the psychometrics of the EdFED and this is that the psychometrics refer to the six item Mokken scale reported in 1997 (Watson), and this is what I now refer to as the EdFED scale. However, confusion has arisen – generated by myself and co-workers – in that the original version, developed using Guttman scaling, was referred to as the EdFED-Q scale and when this was refined it was subsequently labelled as EdFED#2, prior to the development of the Mokken-scaled EdFED scale. This was careless and I can only apologise; for the best tested scale and the one measuring the behavioural aspects of feeding difficulty in older people with dementia, please refer to Watson (1996) and Lin et al. (2008).


  1. Top of page
  2. References