Cross-sectional population studies often suffer from unobserved individual heterogeneity (e.g. genetics or unobserved predisposition to adopt healthy lifestyle) contaminating the estimate of social inequalities on well-being. Accurate estimates of risk factors for impaired quality of life using large epidemiological studies may help to develop services for older people in the community.

We examined the effect of socio-demographic characteristics, lifestyle choices, smoking status and their impact on quality of life in non-institutionalised older people in six-year follow-up.

English Longitudinal Study of Ageing is current and ongoing longitudinal population studies in England over three waves (2002 = 9,953; 2004 = 8,411; 2006 = 8,411) were recruited. The study collected data on the personal, economic, psychological, and social circumstances of aging from a national sample of the non-institutionalised adults aged 50 years or more living in England. Quality of Life was measured using the CASP-19.

Determinants of quality of life using CASP-19 as the dependent variable were: female gender (beta = 1.19, p < 0.0001), younger age (beta = -0.56,p < 0.001), self-reported health (beta = -4.34, p < 0.0001), separation/divorce (beta = 0.81 p < 0.0001), monthly income (beta = -0.91, p = 0.01), accumulated wealth (beta = -0.094, p = 0.01), current smoker (beta = 0.34, p = 0.01 and socioeconomic ladder (beta = -0.115, p = 0.001).

Female gender, younger age, poor self-health reported status, family disruption, active smoking and income were predictors of impaired quality of life. Studies are needed in how to address these socio-economic related factors, social habits and smoking to improve the well-being of older people.