Developing and validating the Community-Oriented Frailty Index (COM-FI)

• Published in: Archives of gerontology and geriatrics 2020-11, Vol.91, p.104232-104232, Article 104232
• Main Authors: Garner, Ian W., Burgess, Adrian P., Holland, Carol A.
• Format: Article
• Language: English
• Subjects: COM-FI; Community; Frailty Index; Predicting social care; Validation
• ISSN: 0167-4943; 1872-6976
• DOI: 10.1016/j.archger.2020.104232

•The COM-FI is a valid and reliable tool, comparable to other frailty indices.•The COM-FI is comprised of physical, psychological, and medical components.•The COM-FI is a tool for use in integrated care settings.•The COM-FI prognostic capabilities include predicting need for care.•The COM-FI offers a platform to build towards a holistic frailty tool. Methods for measuring frailty over-emphasise physical health, and consensus for a more holistic approach is increasing. However, holistic tools have had mixed success in meeting the validation criteria required of a frailty index. We report on the further development and validation of a Frailty Tool designed for use in the community with a greater emphasis on psychological markers, Holland et al’s Community-Oriented Frailty Index (COM-FI). A total of 351 participants aged 58–96 were recruited from Retirement Villages and local communities across the West Midlands of the UK. Participants completed a series of measures designed to assess frailty and outcomes associated with frailty over a 2-year period. All three candidate items (‘polypharmacy’, ‘exercise frequency’, and the Coronary Heart Disease and Diabetes ‘joint effect’) were incorporated into the tool, and one variable, ‘falls’ was removed from the index. The revised COM-FI was shown to be valid and met Rockwood's validation criteria (Rockwood et al., 2006), with the exception that in this specific sample there was no significant gender difference and the index did not predict mortality. Overall, the COM-FI is a valid and reliable tool, although the capacity for the COM-FI to predict mortality over a 2-year period remains inconclusive given the small numbers of people at the higher ends of the frailty range. Prediction of need for social care was good, showing the utility of this community based tool.