The longitudinal relationship between loneliness, social isolation, and frailty in older adults in England: a prospective analysis

Research output: Contribution to journalArticlepeer-review


Background It is estimated that about 10% of people aged 65 and older are frail. Loneliness and social isolation are linked to increased mortality and poorer functional capacity. We assessed trends in frailty status associated with loneliness and social isolation over 14 years in a representative sample of English older adults. Methods In this longitudinal study, we used data from the English Longitudinal Study of Ageing (ELSA), which was designed to recruit a representative sample of adults aged 50 years and older living in private households in England. We analysed Waves 2–8 (covering June, 2004, to June, 2017). Frailty was defined using the frailty index, analysed continuously and as pre-specified categories, to categorise individuals as being non-frail (≤0·08), pre-frail (>0·08 to <0·25), or frail (≥0·25 to 1·00). Loneliness was measured using the UCLA 3-item Loneliness Scale and social isolation was measured following a previous ELSA approach, and both sets of scores were categorised into low, medium, or high. Linear mixed methods and Cox proportional hazard modelling were used, adjusted for confounders. Findings The study sample consisted of 9171 participants at the baseline of Wave 2 (4083 male and 5088 female), with similar numbers in subsequent waves. In the fixed effect model, adjusted for marital status, age, gender, wealth, and smoking status, respondents with higher levels of loneliness had a higher frailty index score (β coefficient 0·006, 95% CI 0·006 to 0·007; p<0·0001), as did those with a higher level of social isolation (β 0·002, <0·001 to 0·002; p<0·0001). Increasing age was associated with an increased frailty index, adjusted for loneliness and social isolation independently. Compared with a low level of loneliness, there was a higher risk of developing frailty with medium loneliness (hazard ratio [HR] 1·57, 95% CI 1·49 to 1·65; p<0·0001) and high loneliness (HR 2·62, 2·49 to 2·76; p<0·0001). Compared with a low level of social isolation, there was a higher risk of developing frailty with medium social isolation (HR 1·12, 1·05 to 1·20; p<0·0001) and high social isolation (HR 1·32, 1·22 to 1·43; p<0·0001). Interpretation Both loneliness and social isolation increase the risk of developing frailty. Understanding these mechanisms might offer opportunities to attenuate this risk.

Bibliographical metadata

Original languageEnglish
Pages (from-to)1-8
Number of pages8
JournalThe Lancet Healthy Longevity
Publication statusPublished - 14 Jan 2021