Frailty has always existed; it was featured in Shakespeare's As you Like It (Abbott and Dykes, 2017) and described in the 1980s as ‘failure to thrive’ (Jamison, 1997). However, an established universal definition of frailty is lacking (Berrut et al, 2013; Morley et al, 2013; Waldon, 2018). The British Geriatric Society (BGS) (2014) describes frailty as how, over time, a person's inner reserve and strength is gradually depleted. Their response to a relatively minor physical or mental stressor is excessive to how a person without frailty would respond. A minor infection, medication or environmental change may cause an adverse response, such as delirium or immobility (Turner and Clegg, 2014), hospitalisation or even death (Garre-Olmo et al, 2013; Morley et al, 2013).
Frailty is described from a physiological and functional approach, as an increased vulnerability due to reduced homeostatic reserves (Clegg et al, 2013; BGS, 2018) leading to increased dependence on others (Morley et al, 2013). Topinkova (2008) suggested that frailty may be predisposed by sarcopenia, osteopaenia, balance disorders, nutritional problems, hormonal dysfunction, a decline in the immune system and protein catabolism and elevated inflammatory levels. Intrinsically, the body is compromised from chronic wear and tear. The multisystem inefficiencies negatively affect physical function and lead to detrimental outcomes (Berrut et al, 2013).
There are clear physiological and biomedical elements contributing to frailty, but there are also non-physical extrinsic factors. Environmental, social and psychological issues can all increase vulnerability (Woo et al, 2005), resulting in a complex and unpredictable trajectory. Older people will manage better in their own homes with person-centred, community-based support (BGS, 2014; Turner and Clegg, 2014). Frailty is a dynamic continuum that can be prevented, reversed or, indeed, palliated, depending on the severity and impact (Fairhill et al, 2011; Wallington, 2016). Waldon (2018) recognised it as multidimensional, requiring a multidisciplinary, integrated approach. There are theoretical models of frailty that when applied, aid in the understanding and recognition of frailty in patients.
Models of frailty
The Canadian Study of Health and Ageing (CSHA) collected data from 10 263 people aged 65 years or over, sampled from 36 different communities across Canada. The subjects represented the community-dwelling population of older adults and those residing in residential institutions (CSHA Working Group, 1994). Of the 10 263 people examined initially, 2305 (22.5%) were examined again clinically 5 years later, in order to develop and determine the criteria for clinical deficits that accumulate and contribute to clinical frailty (Rockwood et al, 2005). Detailed analysis of these data provided the basis of the cumulative deficit model of frailty, the CSHA Frailty Index and Clinical Frailty Scale (CFS) (Rockwood et al, 2005). This model described frailty as an accumulated sum of problems or deficits (Rockwood and Mitnitski, 2007). The cumulative deficit model adopts a quantitative approach: the greater the number of deficits, the higher is the probability that they will contribute to an adverse outcome, for example, hospital admission or mortality (Rockwood and Mitnitski, 2007).
An alternative model of frailty in adults is the phenotype model (Fried et al, 2001), developed using data from the Cardiovascular Health Study (CHS), a prospective, observational study of 5888 individuals over the age of 65 years from the US, between 1989 and 1993 (Fried et al, 1991). The participants were interviewed to identify clinical conditions, cardiovascular diagnoses, overall physical function and performance with activities of daily living. Prevalence of falls was considered, as well as propensity for depressive symptoms. Data from standardised clinical examination, such as electrocardiography (ECG), echocardiography or blood pressure ratio, were also collected. Both sets of data were reviewed by clinicians, along with mortality data and a review of medical records. This review formed the basis for the development of a clinician-determined consensus for the phenotype model of frailty.
This model proposes five descriptors: self-reported exhaustion; reduced muscle strength; low physical activity; slowed walking speed; and unintentional weight loss (Fried et al, 2001). A patient demonstrating three or more of these characteristics is considered to have frailty. This model also allows for a classification of no frailty (no descriptors present) and pre-frailty (one or two descriptors present). The phenotype model offers an evidence-based description of what is found to be the core clinical presentations of individuals with frailty. It also strongly predicts negative outcomes of falls, hospital admissions, disability and mortality (Fried et al, 2001). This model can be applied subjectively during a face-to-face clinical encounter and is, therefore, very relevant to clinical practice. A limitation is that this model does not specifically have a psychological or cognitive dimension (Strandberg and Pitkala, 2011; Berrut et al, 2013).
Cesari et al (2014) accurately pointed out that the two models described above must be compared with caution, as although both theoretical models aim to identify frailty, they have been developed for different purposes. The phenotype model confirms the presence of the clinical syndrome of frailty, while the cumulative deficit model calculates the risk of frailty (Cesari et al, 2014). Although these models view frailty from different perspectives, the associated instruments are both robust indicators of adverse outcomes (Landi et al, 2016).
Frailty identification
A valid standardised screening model will enable appropriate intervention to target frailty (Fried et al, 2001). However, McGuigan et al (2017) noted up to 50 different frailty instruments assessing single or multiple factors. The CSHA Frailty Index (FI) lists a broad range of 70 items incorporating mobility, continence, functional ability, cognition, psychological elements and comorbidities (Topinkova, 2008; BGS, 2014). It weights items equally, which may not be equally relevant to the individual, and factors in disabilities, reducing its value to targeting the prevention of frailty (Strandberg and Pitkala, 2011; Garre-Olmo et al, 2013). Yet, it gives a comprehensive perspective (Fairhill et al, 2011) and considers the frailty syndromes: falls, incontinence, sudden immobility, side effects to medication and delirium (Gwyther et al, 2018).
In order to explore whether the FI is a valid, adequate screening tool for primary care, Drubbel et al (2014) conducted a systematic review of studies evaluating its application. This review omitted discussion of the reliability of the index, as the tool is designed as an automated screening procedure and, therefore, not susceptible to intra- or inter-rater variability. From a database search yield of 867 papers, 20 were selected for inclusion for review against PRISMA criteria. One study was cross-sectional, with 19 cohort studies with a follow-up ranging from 1 to 12 years. The overall quality and risk of study bias were assessed using the Quality in Prognostic Studies (QUIPS) tool. Of the sample (n=20), 5.1% of QUIPS domains showed a high risk of bias, 25.5% showed a moderate risk of bias and 69.4% showed a low risk of bias. Drubbel et al (2014) concluded the FI to be a valid screening method for frailty in adults in primary care and predicting adverse health outcomes.
Manual calculation of the FI is not practical prior to clinical assessment or following brief clinical assessment. Considerable quantities of information are needed to calculate the FI score, and this is possible only after comprehensive clinical assessment. The FI screening calculation, although extremely useful as part of a full comprehensive geriatric assessment, would not be feasible or practical for primary care or community practitioners during opportunistic clinical contacts.
The Electronic Frailty Index (eFI) used in primary care is theoretically based on the cumulative deficit model and FI. Developed by Clegg and colleagues (2016), the eFI considers 36 variables, drawn from 2171 NHS-read code data collected from primary care electronic patient records. The retrospective cohort study analysed anonymised patient data collected between October 2008 and 2013 from 207 814 patients. These baseline data were internally validated against a second cohort (n=201 720) and validated externally as well (n=516 007). Hazard ratios were predicted using bi-variate and multivariate Cox regression analyses, with both discrimination and calibration also predicted using robust statistical processes. Clegg et al's (2016) study concluded that the eFI demonstrates robust predictive validity and good discrimination for nursing home admission, hospitalisation and mortality. Calibration scores, however, were low, indicating that the eFI did not explain variability in these outcomes. The broad range of coded data used to calculate the eFI reflects the multidimensional nature of frailty, but this screening tool is not practical to apply in an individual clinical setting (Strandberg and Pitkala, 2011; Berrut et al, 2013), nor does it identify specific factors contributing to frailty (Walston and Bandeen-Roche, 2015). The eFI offers significant clinical importance in primary care at a population level. Both NHS England (2019) and the BGS (2019) recommend that in patients at greatest risk of frailty, and, therefore, poorer outcomes, the eFI must be supplemented by clinical judgement before reaching a diagnosis of frailty, and, therefore, it must be used alongside alternative frailty screening tools or assessments in clinical practice.
Physical measures such as gait speed are endorsed by the BGS (Strandberg and Pitkala, 2011; BGS, 2014). Frailty is indicated if mobility exceeds 5 s over a 4-m distance. Measuring a singular variable, the gait speed test is both valid in estimating physical function and deterioration due to frailty, and simple to execute (Berrut et al, 2013). This test is not universally applicable, as exclusion is necessary for individuals with an established physical disability. Although gait speed is useful in frailty risk assessment, further research of more detailed gait analysis will enhance the sensitivity and specificity of using gait markers to predict the presence of frailty (Schwenk et al, 2014). Analysis of gait technique, quality and the assessment of more demanding tasks, such as fast walking or a cognitive task, would be useful.
Since frailty is a dynamic process influenced by multiple factors (Berrut et al, 2013), it is unlikely that one measure alone will adequately identify all the indicators of frailty. The CSHA Clinical Frailty Scale (CFS) (Rockwood et al, 2005) is easy to administer; the clinician makes a judgement based on their holistic assessment of the patient. The 7-point scale (revised in 2007 as a 9-point scale) was developed using data from the CSHA (n=2305). The authors concluded that this scale measured well as a predictor of frailty against previous rule-based methods, and it appears to address the complexity of frailty better (Rockwood et al, 2005). The CFS was highly correlated with the FI, with each point on the 9-point scale increasing the medium-term risk of death, with 95% confidence intervals. However, it is not well validated in clinical settings, and its inter-rater reliability was not examined in the original research.
All clinical staff providing primary care to patients over the age of 65 years are required to have skills in recognising frailty (NHS Employers, 2017), and the CFS has been recommended by NHS England as a standardised opportunistic screening tool (Moody et al, 2017). It is useful when implemented alongside the eFI tool, in order to confirm clinical correlation. The NHS Acute Frailty Network, hosted by NHS Elect, also recommends the CFS as a logical, first-choice opportunistic screening tool for all clinicians to adopt (Acute Frailty Network, 2019). The tool is recommended to be applied against the individual and their baseline level of function 2 weeks prior to the illness or injury for which they are presenting to health services. This is to avoid any bias that may arise from acute deteriorations in function; understanding the patient's baseline level of function and, therefore, baseline level of frailty and risk is important for correct screening. Adopting a standard tool will support a common language when discussing frailty in clinical, research and educational contexts.
Social frailty
As the determinants of frailty are often multifaceted (Mulasso et al, 2016), Fairhill et al (2011) proposed treating frailty according to the evidence base for each issue identified, but some measures like those based on the frailty phenotype neglect cognitive, social and environmental factors (Bessa et al, 2018). Limited evidence exists regarding the social component of frailty and what elements should be assessed to determine its presence, that is, engagement in social activities, social support, perceived loneliness, social networks or if living alone (Bessa et al, 2018). Social isolation and loneliness are possible contributors to developing frailty, and increased social support is associated with lower frailty (Woo et al, 2005; de Labra et al, 2018). Makizako et al (2018) found that social frailty may precede physical frailty, but other studies indicate that social isolation is a consequence of increased frailty (Gale et al, 2018). Social factors can predict the likelihood of mortality as effectively as factors of physical frailty (Garre-Olmo et al, 2013), and both isolation and loneliness are related to reduced quality of life (Windle et al, 2011; Gale et al, 2018).
Social isolation refers to reduced social contact or involvement (Gale et al, 2018), which could be imposed due to ill health or disability (Windle et al, 2011). Some 17% of older people have contact with family, friends or neighbours less than weekly (Davidson and Rossall, 2015:2). Loneliness relates to a subjective emotional state regardless of regular contact (Windle et al, 2011; Davidson and Rossall, 2015). Related is social detachment, occurring when a person fails to participate in social activities. It is associated with living alone, lower income, environmental barriers and ill health and can be confounded by maladaptive social cognition, where lonely people present themselves in a manner that increases the probability of further isolation (Davidson and Rossall, 2015).
The effects of social isolation can increase cost and demand on health and social care (Windle et al, 2011), and loneliness affects wellbeing, leading to higher rates of cardiovascular problems and low mood (Contact the Elderly and Social Care Institute for Excellence (SCIE), 2012). Lonely and isolated older people may experience nutritional problems (Henderson, 2016) and reduced mobility due to sarcopenia, causing functional dependence and physical frailty (Gale et al, 2018). Skeletal muscle decline is progressed by further physical inactivity and poor nutrition, leading to negative outcomes for health (Stott and Quinn, 2017). Maladaptive midlife behaviours, such as excess alcohol consumption, are linked to frailty and poorer outcomes with mobility and independence in later life (Gobbens, 2016).
Preventative measures need to be adopted to ensure people not only live longer, but live healthier lives (Charles and Rashbrook, 2016). Evidence encourages timely midlife approaches to maintain wellbeing and independence and reduce the effects of sarcopenia and frailty by eating healthily, being physically active and refraining from alcohol and smoking (Charles and Rashbrook, 2016; Landi et al, 2016). These issues could be addressed alongside social connectivity. Further research is required to determine whether social isolation or loneliness is pivotal in identifying frailty (Bessa et al, 2018). Positive evidence to link social factors could inform preventative strategies for frailty and support ageing well (de Labra et al, 2018).
Isolation and loneliness
As with frailty, isolation and loneliness can be improved to prolong quality of life (Contact the Elderly and SCIE, 2012). Interventions may increase the quantity and quality of interactions, address maladaptive social cognition and improve social skills (Davidson and Rossall, 2015). A high proportion of services in the community are delivered by the voluntary sector (Contact the Elderly and SCIE, 2012), reducing the strain on health and social care. A community asset-based model values the strengths of local resources, connects communities (Charles and Rashbrook, 2016) and builds local community resilience (Fox, 2016) to improve health and wellbeing and empowers self-management (Health Foundation, 2015). The SCIE's (2017) asset-based places model for development identifies and maps local resources and then connects assets to build social networks. Individuals are guided to relevant services by social prescribing, community navigators or peer support. The SCIE showcases a social prescribing scheme that improves wellbeing for local residents with complex long-term conditions. Referral to community activities based on individual needs resulted in reduced emergency contacts and hospital admissions and yielded social care savings (SCIE, 2017).
One-to-one contact with befriending schemes provides essential companionship, and community navigators effectively reduce social isolation by introducing individuals to appropriate social outlets (Windle et al, 2011; Contact the Elderly 2012; Davidson and Rossall, 2015). Evidence indicates that social activity groups effectively reduce loneliness and social isolation (Contact the Elderly and SCIE, 2012). Luncheon clubs, coffee mornings and reminiscence groups improve wellbeing by increasing confidence and general happiness (Ipsos Mori, 2014; Henderson, 2016). Older men are less likely to engage in social groups than women; however, schemes such as ‘Men in Sheds’ providing practical tasks within a social context have had worldwide success. Most research available in this field relates to small, localised initiatives (Health Foundation, 2015) that may not be generalisable to other areas. In some regions, Age UK in partnership with hospital trusts offers a home-from-hospital service to facilitate safe discharge for older people living alone. In addition to navigators encouraging participation in community activities, they assist with food shopping and medication, support to attend medical appointments and ensuring that homes are safe and warm (Age UK, 2018). Low income and progressing frailty both present a challenge to maintaining a safe environment (Age UK, 2015). A frail person returning home on discharge from hospital may not thrive in a cold home, leading to subsequent readmission (Stewart and Dhesi, 2016).
Self-neglect: signs of frailty
The presence of personal disorders and hoarding behaviour may indicate frailty (Lee et al, 2016) and has been found to be correlated with low social engagement and living alone (Abrams et al, 2002). A longitudinal study conducted in the US by Schafer et al (2018) examined whether disorderly household conditions and bodily self-presentation predict mortality. The study methods included naturalistic observation of 2995 older adults, and a wide range of additional co-variates significantly associated with mortality were analysed. This study concluded with 95% confidence intervals that older people living in disorderly households are at greater risk of mortality, because it accelerates frailty. Disorderly body self-presentation in the form of a dishevelled appearance or poor hygiene is associated with mortality, alongside demographic factors, health conditions, poor social interaction and a level of frailty. Self-neglect has been linked to poor social support, reduced nutritional intake and physical function (Schafer et al, 2018), resulting in poor quality of life and increased falls, admissions and mortality (Papaioannou et al, 2012; Dong, 2014). An individual with frailty is more likely to live in an unkempt environment (Dong et al, 2012), and self-neglect is further compounded by the individual's inability to complete their daily activities, poor health, cognitive deficits or lack of care support (Dong et al, 2012). Timely preventative action could prevent this deterioration.
Environment
The Bristol Centre for Public Health and Wellbeing in collaboration with the Building Research Establishment (BRE) conducted a comprehensive review of peer-reviewed literature and practitioner-driven grey literature to examine the role of home adaptations in improving later life (Powell et al, 2017). Their review focused on exploring the efficacy of adaptations, cost efficiency and the impact of home adaptations on health and wellbeing outcomes of community-dwelling adults aged over 65 years. Some 60 eligible documents were included in the systematic review with a wide range of study designs, including randomised controlled trials (11), cross-sectional studies (10) and qualitative studies (17). Ten studies rated against specific criteria were found to be of high quality and rigour, 26 were of medium quality and 24 were of fair quality. Evidence was reviewed from 13 high-income countries including the UK, all experiencing similar health and social care challenges associated with ageing, so the results can inform decisions for the UK population. Powell et al (2017) concluded that strong evidence exists that adaptations are effective in reducing falls and injuries and, therefore, admissions. Improvements in functional ability are more significant when undertaken in conjunction with other improvements, such as removing hazards and improving lighting. Powell et al's (2017) systematic review indicates that the greatest outcomes are achieved when individuals, families and carers are all involved in the decision-making process. There is, however, insufficient evidence to support the cost-effectiveness of home adaptations or potential for return on investment (Powell et al, 2017).
Kazuko and Shimanouchi (2014) advised that home modifications prevent the progression of frailty and increase quality of life and independence in daily tasks such as bathing. Compensatory home improvements and adaptations to optimise function (Renaut et al, 2015) encompass the philosophy of applying proactive and preventative measures to prevent and reverse frailty (BGS, 2014). Small changes can be cost effective and improve the lives of older people (House of Lords, 2013). They can empower the individual to plan for their future capabilities and encourage continued independence (Burgess and Morrison, 2016). Yet, Mackenzie et al (2015) reported that individuals often modify their behaviours rather than their environment. Housing services are instrumental in the provision of a safe environment that maintains independence (Wood and Salter, 2012). Burgess and Morrison (2016) suggested that a proportion of older people wishing to ‘optimise their environmental richness' (Burgess and Morrison, 2016:209) are unable to make positive changes to their environment without external direction and support. These individuals are found to simply endure their poor environmental conditions.
Conclusion
Frailty poses many operational challenges in a clinical setting, with no universally accepted clinical definition and multiple theoretical models and screening methods available. What is widely accepted is that frailty, and its many contributing factors, have the potential to predict poorer health outcomes and mortality. Understanding the link between frailty, and patients' health and social care outcomes is very relevant in all clinical settings, but especially in primary care and community care settings. Health professionals working in these areas of care are ideally placed to recognise or prevent frailty, as the patient's own dwelling or living environment may also be the setting in which the care is delivered. Recognising frailty early and recognising patient groups at elevated risk are priorities for enhancing care and timely intervention in the future.
The review of the literature presented here indicates that physical frailty and reduced function can certainly be accelerated by social frailty. Frailty is complex and should be considered from a multidimensional perspective considering social and environmental elements in addition to physical factors. A shared approach between health, social care, housing and community organisations can make a difference.