Validating Hospital Episode Statistics based measures of frailty risk

In Chapter 4 we described work to create a hospital frailty risk score that uses routinely collected inpatient data from the HES. 131 In this chapter, we examine agreement between the derived score using HES data and other frailty scores derived from more detailed clinical data sets.There are a number of different ways that frailty can be conceptualised and an even greater number of ways that it has been measured. 132 No single standardised measure of risk or frailty exists, and classifications of people as frail or non-frail can vary depending on which tool is used. The two most prominent measurement approaches, the phenotype model (Fried) 133 and the cumulative deficit model (Rockwood), 133 show overlap in their identification of frailty, 134 although agreement in practice is only moderate. 3 The Fried and Rockwood instruments have different approaches to defining frailty and are seen as complementary.