Electronic Frailty Index (eFI) John Young Geriatrician, Bradford, UK National Clinical Director for Integration & Frail Elderly, NHS England

Mrs Greenaway was found on the floor (“FLOF”) with new confusion by the home care staff and taken to hospital where is was found to be poorly mobile. Fall Delirium Immobility “She was a fall waiting to happen.” Home care staff Frailty is ………………

The “new” narrative: Frailty as a long-term condition ? A LTC is: “A condition that cannot, at present, be cured but is controlled by medication and/or other treatment/therapies” (DH 2012) Frailty is: Common (25-50% of people over 80 years) Progressive (5 to 15 years) Episodic deteriorations (delirium; falls; immobility) Preventable components Potential to impact on quality of life Expensive (Harrison, Young, Clegg, Conroy Age & Ageing 2015)

Increasing frailty FitMild frailtyModerate frailtySevere frailty Frailty as a trajectory

Diagnostic Test Accuracy (DTA) for simple frailty instruments (Systematic Review) SensitivitySpecificity Gait Speed <0.8m/s99%64% Gait Speed <0.7m/s93%78% TUGT >10s93%62% PRISMA 783%83% (wide CIs) Self-reported Health83%72% (wide CIs) Groningen Frailty Indicator58%72% Polypharmacy (>5 meds)67%72% GP clinical assessment58%72% (Frailty instruments assessed against a reference standard) (Clegg, Teale, Young. Age Ageing 2014)

Phenotype Frailty Model (Cardiovascular Health Study [n=5210] Fried et al 2001) Weight loss:> 4.5kg or > 5% per year Fatigue:US Centre for Epidemiological Studies Depression Scale Sedentary Life:< 383 Kcal/week men < 270Kcal/week women Slow gait speed:Standardised cut-off times to walk 4.57m stratified by sex & height Weakness:Dynamometer measurement stratified by sex & BMI

Cumulative Deficit Model of Frailty: Frailty Index (Rockwood et al) “The more things that are wrong with you, the more likely you are to be frail” Frailty Index counts “deficits” A deficit is a thing that is wrong with you (symptom, sign, disease or disability) Frailty Index = the proportion of deficits accumulated over time Simple calculation: Zero deficits from list of 50: FI = 0/50 = 0 Ten deficits from list of 50: FI = 10/50 = 0.20 Frailty Index(s) based on deficit accumulation closely related to risk of death (Mexico, China, Canada, Europe etc. …)

Development of the primary care eFI Existing primary care EHR (“SystmOne”) Read Codes (>80,000 8,000 2,200) Read codes map onto 36 ‘DEFICITS’ Tested in “ResearchOne” (n=227,648 ≥65y) Internal Validation Process (n=227,063 ≥ 65y) External Validation Process (n=516,107 ≥ 65 y)

Outcome Mild frailty (HR, 95% CI) Moderate frailty (HR, 95% CI) Severe frailty (HR, 95% CI) 1 yr care home admission2.00 (1.68 to 2.39)2.70 (2.41 to 3.04)5.94 (4.61 to 7.64) 3 yr care home admission1.52 (1.37 to 1.69)2.70 (2.41 to 3.04)3.42 (2.84 to 4.12) 5 yr care home admission1.56 (1.43 to 1.70)2.34 (2.10 to 2.61)3.00 (2.42 to 3.70) 1 yr hospitalisation1.85 (1.81 to 1.88)2.96 (2.90 to 3.02)4.62 (4.50 to 4.74) 3 yr hospitalisation1.71 (1.69 to 1.73)2.54 (2.51 to 2.58)3.64 (3.57 to 3.70) 5 yr hospitalisation1.63 (1.61 to 1.64)2.43 (2.40 to 2.46)3.59 (3.54 to 3.65) 1 yr mortality1.91 (1.78 to 2.04)3.39 (3.15 to 3.65)5.23 (4.73 to 5.79) 3 yr mortality1.74 (1.68 to 1.81)3.02 (2.90 to 3.14)4.56 (4.29 to 4.84) 5 yr mortality1.66 (1.62 to 1.71)2.73 (2.64 to 2.81)3.88 (3.68 to 4.09) eFI outcomes

Proportion alive Time Primary care electronic Frailty Index (eFI): survival plots (n=227,648; >65y) Fit Mild frailty Moderate frailty Severe frailty 5 yrs Supported self-management Care & Support Planning Comprehensive Geriatric Assessment

FIT 32% MILD 41% MOD 20% SEV 7%

Frailty & Social Deprivation eFI score Social Deprivation (IMD rank group)

PartnersEngagement Count GP Practices22 CCGs (n=211)35 CSU1 SCN1 CLAHRC1 Public Health (regional) 3 Industry Partners2 (ACG Systems) VCS1 (Age UK Y&H) National Spread of eFI (Year 1) Clinical Commissioning Group (Pop approx. 250,000) “at risk” populations De-prescribing EoLC/ACPs Supported-self management etc, etc………..

Summary: The eFI relates to the cumulative deficit model of frailty and uses existing primary care data (Read codes) to identifies people with mild, moderate and severe frailty Currently implemented in SystmOne (30% GPs) as a practice level report eFI has potential to inform development of proactive care for older people across the frailty spectrum Good early take up by CCGs in England May enable better targeting of interventions to reduce health inequalities in older age

Collaboration for Leadership in Applied Health Research and Care, Yorkshire and Humber Delivering innovative research through effective partnerships Acknowledgements: This presentation presents independent research by the Collaboration for Leadership in Applied Health Research and Care Yorkshire and Humber (NIHR CLAHRC YH). The views and opinions expressed are those of the authors, and not necessarily those of the NHS, the NIHR or the Department of Health. CLAHRC YH would also like to acknowledge the participation and resources of our partner organisations. yh.nihr.ac.uk Thank you. University of Leeds Andrew Clegg John Young Tizzy Teale TPP Chris Bates John Parry Ankit Sharma University of Bradford M A Mohammed University of Birmingham Ronan Ryan Linda Nichols Tom Marshall