摘要：Public policies aim to promote well-being, and ultimately the quality of later life. Positive perspectives of ageing are underpinned by a range of appraoches to successful ageing. This study aimed to investigate whether baseline biological, psychological and social aproaches to successful ageing predicted future QoL. Postal follow-up in 2007/8 of a national random sample of 999 people aged 65 and over in 1999/2000. Of 496 valid addresses of survivors at follow-up, the follow-up response rate was 58% (287). Measures of the different concepts of successful ageing were constructed using baseline indicators. They were assessed for their ability to independently predict quality of life at follow-up. Few respondents achieved all good scores within each of the approaches to successful ageing. Each approach was associated with follow-up QoL when their scores were analysed continuously. The biomedical (health) approach failed to achieve significance when the traditional dichotomous cut-off point for successfully aged (full health), or not (less than full health), was used. In multiple regression analyses of the relative predictive ability of each approach, only the psychological approach (perceived self-efficacy and optimism) retained significance. Only the psychological approach to successful ageing independently predicted QoL at follow-up. Successful ageing is not only about the maintenance of health, but about maximising one's psychological resources, namely self-efficacy and resilience. Increasing use of preventive care, better medical management of morbidity, and changing lifestyles in older people may have beneficial effects on health and longevity, but may not improve their QoL. Adding years to life and life to years may require two distinct and different approaches, one physical and the other psychological. Follow-up health status, number of supporters and social activities, and self-rated active ageing also significantly predicted QoL at follow-up. The longitudinal sample bias towards healthy survivors is likely to underestimate these results.
