Article Text

Download PDFPDF
OP26 Self-reported frailty components predict incident disability, falls and all-cause mortality in later life: results from a prospective study of older British men
  1. E Papachristou1,
  2. SG Wannamethee1,
  3. S Iliffe1,
  4. AO Papacosta1,
  5. LT Lennon1,
  6. PH Whincup2,
  7. SE Ramsay1
  1. 1Primary Care and Population Health, University College London, London, UK
  2. 2Population Health Research Institute, St George’s University of London, London, UK


Background Frailty is an established state of increased vulnerability for disability, falls, and mortality. Assessments of frailty components including grip strength, weight loss, physical activity and gait speed are complex or require objective measurements which are challenging in primary care. In this study we examined the ability of self-reported frailty components to predict known adverse outcomes of frailty including disability, falls and mortality and compared it to that of an established frailty phenotype.

Methods The British Regional Heart Study is a cohort study comprising a socially and geographically representative sample of older British men, initially examined in 1978–80. In 2010–12, 1622 participants attended a physical examination (55% response rate) and completed a questionnaire (68% response rate). Frailty was based on weight loss, grip strength, exhaustion, slowness, and low physical activity. Single self-reported measures of the frailty components were selected from the questionnaire. Information on disability (problems walking 400 yards/taking stairs) and falls were collected through a postal questionnaire in 2014 after a 3-year follow-up. Data on mortality were obtained through the NHS Central Register. The discriminative ability of models with up to three single subjective measures were compared to the frailty phenotype for incident disability and falls using receiver operating characteristic-area under the curve (ROC-AUC). The predictive ability of these models for all-cause mortality was assessed using age-adjusted Cox proportional hazard models.

Results A model including single items of self-reported slow walking speed, physical inactivity and exhaustion had a significantly increased ROC-AUC (0.68, 95% CI 0.63–0.72) for incident disability compared with the frailty phenotype (0.63, 95% CI 0.59–0.68; p-value of ΔAUC = 0.003) and was a significant predictor of all-cause mortality (hazard ratio = 1.65, 95% CI 1.34–2.03). A second model including self-reported slow walking speed, physical inactivity and weight decrease had a higher ROC-AUC (0.64, 95% CI 0.59–0.68) for incident falls compared to the frailty phenotype (0.57, 95% CI 0.53–0.61; p-value of ΔAUC < 0.001) and was also a significant predictor of all-cause mortality (hazard ratio = 2.13, 95% CI 1.71–2.66).

Conclusion Simple self-report questions on walking speed, physical activity, feelings of exhaustion and weight-loss improve the risk stratification provided by an established frailty phenotype for incident disabilities and falls in older people, and are predictive of all-cause mortality. Developing such a simple assessment tool has potential implications for care pathways of older people to reduce the adverse consequences associated with frailty.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.