Conference Paper BHPS-2009 Conference: the 2009 British Household Panel Survey Research Conference, 9-11 July 2009, Colchester, UK
Can CASP scores predict mortality in BHPS sample?
Background: As society ages, the quality of life of its members gain greater salience. At least in Western affluent countries, the ageing has not been accompanied by a reduction in the quality of life and the indications are that the quality of life might actually improve in the early old age. Although, the majority of the research effort is directed towards assessing the impact of longevity on the quality of life, the complement of that research question, whether the quality of life impacts longevity is equally interesting. To date, that question has been investigated only in the context of disease and health related quality of life. Recently, introduction of a generic measure of quality of life, CASP-19 into longitudinal studies offers an opportunity to test the association between quality of life and mortality. In BHPS it was offered in 2001 and 2006. This paper aims to examine whether CASP-19 scores can predict mortality in BHPS sample during 2001-2006.
Methods: We used data from British Household Panel Survey Wave 11 (2001-2) when CASP-19 was first used. We selected all those who were interviewed directly and face to face and who were 40 years or older (N=10,291). We followed all those included in the sample for the next five waves (Wave 12 to Wave 16) and our primary outcome was a binary variable indicating whether the participant has died during the interval. We excluded two items from CASP-19 related to health and financial status to avoid bias. Other co-variates used were age, sex, socio-economic position, household income, selfrated health, limiting long-standing illness, and medical conditions.
Results: Compared to a mortality rate of 12 per thousand in those having average quality of life (CASP score 29.4 to 45.8), those with below average quality of life had more than twice (27 per 1000) and those above average had a third less (8/1000) the mortality rate. This gradient was retained for the most part when age and sex strata were examined separately. Regression models adjusted for covariates confirmed the protective effect of quality of life on mortality. Domain specific analysis showed that only control and self-realisation had this effect. These results are discussed in order to explain how quality of life might influence mortality.