This relative importance of housing and financial security on health increases during working life and retirement. In the Whitehall II cohort, the mental health of people who owned their own houses, unlike those who rented their houses, consistently improved over their working life and continued during their retirement. This cumulative impact of housing tenure on the inequalities in mental health in older people increased when we took into account not only this structural factor, but also intervening factors such as housing quality and financial problems. But when we controlled for confounding variables, housing tenure no longer had a significant independent effect, whereas housing quality and financial difficulties retained their explanatory significance.
The finding that the quality of housing and financial security are more important explanatory factors in explaining the mental health of older people than housing tenure is supported by other studies. Experiencing financial difficulties at baseline was the only predictor in new episodes of depression in the General Psychiatric Morbidity Survey [40
] and in the British Household Panel Survey, which adjusted for more objective measures of standard of living, such as occupational level [41
The explanation for how "the social becomes biological" is likely to have many strands [42
].(p.48) Housing is part of the network of health resources that can either promote health over the life-course or increase susceptibility to illness and disease [17
]. However, the quality of housing is particularly important to health at older ages, because susceptibility to low temperature increases with age and older people are exposed more than other age groups to the indoor home environment. Moreover data from the first five waves of the British Household Panel Survey, which looked at residential mobility for those over 55, found relatively few older people moved house [43
]. In what might be another case of the inverse care law, older people on low incomes may also lack the funds to maintain and repair their homes, or afford the co-payments to take up public funds to improve their houses through retrofitted insulation and boilers and to pay their heating bills.
In policy terms, housing remains an important way of improving older people's health. Successive governments have encouraged home ownership through various tax subsidies and Right-To-Buy schemes. However, the relationship between housing tenure, quality, financial status and health may not always be direct. The Whitehall II study enables us to look at the direct and indirect interrelationship between the broad aspects of the determinants of health (housing, employment, financial problems and so on) to show that, after controlling for intervening variables such as employment grade, financial problems and housing quality, housing tenure is no longer a significant explanation of mental health in retirement. These results suggest, that as in the Burrows' study, owning a home in poor condition, without the financial resources to remediate it, may be a health burden for the owner occupier [33
There are however, some methodological caveats to our results: both the independent variables (tenure, housing quality and financial problems) and the dependent variable (psychological well-being) are based on self-report, so that some of the covariance between housing quality and mental health might be created by the overlap in method. While self-perceptions are generally powerful predictors of health, the self-perception of housing quality used in this analysis was non-specific; future studies of the relative impact of housing tenure and housing quality would be strengthened by having independent measures of housing quality. Another limitation is the low proportion of participants living in rented housing, well below the average of around 22% of the population in this age group in the 2001 England and Wales census. This reflects the socioeconomically advantaged nature of the sample, namely those employed in the civil service. However, the Whitehall II study was never designed to be representative of the British population. Instead, its strength lies in discovering aetiological relationships on the social determinants of health. An association between housing problems and health in this advantaged sample suggests that this association would be even stronger in the general population with a greater proportion of socioeconomically disadvantaged groups.
Selection biases due to missing data are a problem inherent in all longitudinal studies, especially so in ageing studies. With over 1300 deaths in the cohort up to the ninth phase of the study, those remaining in the study are healthier on average than non-participants and are also more socioeconomically advantaged. However, this pattern of non-response would only affect the results presented here if the association between housing and mental health differed between those remaining in the study and non-participants at later phases of the study. Previous analysis of non-participation in the Whitehall II cohort has shown that the association between non-response and mortality does not differ by socioeconomic group [44
]. This suggests that the pattern of non-response in this analysis may not have biased the results.
In this paper we have looked at only one health outcome, but it is biologically plausible that a number of other health symptoms, such as respiratory and coronary symptoms could also be affected by housing and the indoor environment, as poor housing has been related to cardiovascular disease [28
]. The London-focus of the Whitehall II study also means that we have smaller variation in housing quality than if it were a national cohort. All these factors may under-estimate the relationship between housing and the mental health of older people.