Letter to the Editor

摘要

Sir—I read the paper by Perera et al. and the accompanyingneditorial with interest as there is a need for geriatricians tonrecognise the treatments and outcomes for frail adults.nHowever, I have great difficulty in determining what cutnscore for the Edmonton Frail Scale (EFS) was used. It is notndescribed in the article's Methods section.nThere are two relevant references. Wolfson et al. havenmeasured the correlation of the EFS against a Geriatrician'snClinical Impression of Frailty questionnaire. Whilst theyndemonstrate a modest relationship (Pearson's r = 0.64), theyndo not identify a cut point in the EFS above, or below, whichna classification of 'frailty' can reasonably be ruled in or out.nDasgupta et al. have found that EFS score ranges (≤3 orn>7) have a reasonable relationship to the ability to predictnwhich of an elderly cohort would not or would havenpost-operative surgical complications. But this is scarcely anreference against a 'gold-standard' for identifying frailty, andnagain the statistical power of the association is poor withnreceiver operator curve 'area under curve' data of 0.69 fornall complications and 0.65 for cardiac complications.nSo, while there is an association for the EFS with othernmeasures of frailty, I cannot apply these data to any othernpopulation of elderly because the key defining criterion isnnot given.