Case Study 9: Healthy Ageing and Risk Stratification using electronic Frailty Index Birmingham Pilot

In a nutshell: The Five Year Forward View (NHS England, 2014) argues that the sustainability of the NHS depends on a radical upgrade in prevention and public health. People with frailty are more likely to fall than those who are not frail; and older people presenting with one of the frailty syndromes (such as falls or immobility) may well already have established frailty. Birmingham’s Falls and Fracture Public Health Programme Lead is using the eFI to understand more fully the needs of frail older people to provide a deeper understanding of current service provision and/or to identify any gaps - which in turn could inform new ways of working and/or commissioning decisions to improve care for frailty older people at risk of falls and fracture.

In partnership with Public Health Birmingham, 5 SystmOne GP Practices are using the eFI to analyse the current service provision for patients identified using the eFI as having frailty, according to the different severity grades of frailty. The GP Practices are acting as test beds as part of the locality’s Falls and Fracture Prevention Programme to give insight within each of the 3 CCGs as to existing care provision for people with frailty.

eFI data will not be used in isolation but alongside social deprivation data, Health & Wellbeing Board priority areas and Joint Strategic Needs Assessment health profiles to provide a richer picture of population need. In Birmingham, people experience, on average, 19 years of ill health which is the second worse in the West Midlands. This has a significant impact on health and social care expenditure. The hypothesis is that by targeting interventions in line with the needs of people with frailty, this would lead to earlier intervention and/or prevention through crisis recognition and management, and self-care; which could potentially contribute to a lessened burden on the Health and Social budget.

The project aims to use the eFI to:

The anticipated outputs from the project include increased understanding of:

 eFI Analysis

Following discussions with the Birmingham clinical commissioning groups (CCGs), 5 GP practices agreed to participate in the eFI pilot, equating to a population fewer than 40,000 patients, with 30% of this cohort being aged 65+. Data was extracted by practice staff and shared with public health for further analysis.

Step 1: eFI banding against Individual Ages
Once we received the dataset the eFI scores were grouped as per the bandings detailed above and ages as of 01 January 2016 calculated.  Initial analysis was simply plotting individual ages against this for each practice and for all practices combined.

Step 2: eFI banding against Age Bands
We then added age bands, in ten year intervals and plotted these against the eFI bands for the same groups.

Step 3: Mean Age
Mean age for each frailty group (including “no frailty”) and the population were calculated across all five practices combined.  Standard Deviation from this mean average was also calculated.


Please see the tables and graph below for more detailed extracts from the analysis.

Table 1: GP Practice and patient eFI characteristics

 General Practice  Practice 1 Practice 2 Practice 3 Practice 4 Practice 5  Total
 Patients aged 18+ (Mean and SD) 55.17
 Fit Age (Mean and SD) 50.25
 Mild Frailty aged 18+ (Mean and SD) 71.07
 Moderate Frailty aged 18+ (Mean and SD) 82.43
 Severe Frailty aged 18+ (Mean and SD) 85.82
 eFI Score (Median and IQR) 0.06
 99th Percentile eFI Score 0.33  0.36 0.39 0.31 0.33 0.36


Graph 1: Patients and number in severity bands agains age (all five practices in trial)


Table 2: Percentage of patients overall and in seveity bands against 10-year age bands (all five practices in trial combined)

The results of the stratification exercise indicate that an individual’s journey from a fit to mild category takes an approximate twenty years, movement from mild to moderate takes 11 years and moderate to severe takes 5 years.


Initial results from the pilot show that the eFI could have a population benefit with the aptitude to identify people with frailty earlier and allow for targeted interventions to be offered with the aim of avoiding adverse outcomes.  A point of further investigation would be around the clustering and spread of the Read codes (or deficits) which make up individual eFI scores. At a population level, this may provide insight into deficit accumulation and possibly enable the identification of distinct frailty trajectories. For example, what makes an individual with multi-morbidity with an eFI score suggestive of moderate frailty more resilient than an individual living with one long-term condition with an eFI score suggesting mild frailty?


The greatest utility of eFI is as an initial stratification tool, which needs to be supplemented with health professional judgement and clinical knowledge of their individual patients and a comprehensive geriatric assessment approach in order to identify care and support needs and provide effective treatments to manage frailty at an individual level. The utility of eFI within health and social care could be through the ability to inform frailty pathway redesign, resource allocation for commissioners and most importantly, allow for a more proactive approach to healthy ageing for older people.


Contact: Alison Doyle, Falls and Fracture Prevention Public Health Programme Lead and Hashum Mahmood, Evidence Base Manager – both Public Health Birmingham

Email: or