Functional capacity deteriorated in six years among men in oldest cohort and among women in all three cohorts. This study showed SEP disparities, especially disparities by adequacy of income in functional capacity. Poorer adequacy of income was most consistently associated with poorer functional capacity. However, changes in functional capacity over time by SEP remained in most parts the same or even narrowed independent of health behaviours.
Our results indicate steady or partly decreasing SEP disparities in functional capacity. This is partly in accordance with earlier results. In a study from the USA it was found out that poorer women had worse functioning at baseline independent of health-related covariates. However, poverty status was unrelated to lower extremity function decline over three years. Rautio et al. found out that higher income and better education were related to better physical capacity among 75-year old people living in Jyväskylä, Finland, but decline in physical capacity in five and ten year periods was parallel in all socio-economic groups. The association between income and physical capacity remained after adjusting for health behaviour. Similar educational disparities in disability have also been found in other studies. In the present study no clear disparities by education was found. As the number of participants in different age cohorts were rather small, it is possible that there was not enough statistical power to reveal educational disparities.
Our results are in accordance with a study from the UK, where it was found out in longitudinal analyses that self-perceived adequacy of income was a strong predictor of onset of disability. Adequacy of income has not been extensively studied in relation to functional capacity. However, previous studies that have used it in analyses have consistently shown it to be a strong predictor of health and functional capacity[5, 18, 20, 21] Even though adequacy of income as an indicator of SEP has not been comprehensively rationalized in previous literature, it is an indicator which is partly independent of income. Therefore there may be considerable variation in the severity of the financial strains experienced by people having similar incomes. This might be one explanation why increasing number of studies use it when examining socioeconomic disparities in health.
As different studies have various designs, there exist contradictory results. For instance a Danish study concerning 40–50 year old people found out that those with lower social class had a higher onset of mobility disability. Similar results have been suggested from China. In the Danish study, however, people were disability-free at baseline. In the present study it was of interest to examine changes in average scores of functional capacity in different age cohorts by SEP, not only those who are disability-free at baseline.
Results from the present study indicate that adjusting for health behaviours does not have an effect on disparities or changes in functional capacity by SEP. This is in accordance with some previous evidence. It has been concluded previously that health behaviour has a strong impact on functional capacity[2, 26]. Correlation analyses revealed (data not shown) that all the behaviours used in the present study correlated statistically significantly with each other and all the health behaviours excpet alcohol consumption correlated significantly with baseline functional capacity. Even though alcohol consumption was not individually associated with functional capacity, there is previous evidence that health behaviours are interrelated with each other. Especially smoking has shown to have strongest and most consistent associations with other unhealthy behaviours. These interactions between various behaviours might have an impact on the level of functional capacity. It seems that according to population groups examined in this study, health behaviours and their possible interactions may result in the accumulation of advantages or disadvantages in a longer time span.
There are a few points that may explain why health behavior in this study did not had an impact on functional capacity. First a relatively short follow-up time might explain why health behaviour did not turn out to be a stronger predictor of later functional capacity. Second, the health behaviour between various SEP’s did not differ so that it would have had an impact on later functioning. The health behaviour measures used in this study were based on large Finnish nationwide follow-up surveys which have been carried out since 1970s. However, it is possible that the measures were not sensitive enough to reveal the potential impact of health behaviour on later functioning.
This study had several strengths. The study design was longitudinal which enabled the investigation of causal relationships. The data included information on two strong measures of socio-economic position, which made it possible to assess the individual effects of different aspects of social determinants. Another strength was that we had data on health-related covariates which may act as mediators from socio-economic position to ill health and poor functional capacity.
However, our data had some limitations as well. The data was collected from three cohorts born 10 years apart from each other, which made it complicated to combine the data-sets thus giving somewhat small sample sizes for multivariate methods. Also, the follow-up time was relatively short. Especially in the younger cohorts, changes in functional capacity probably occur during a longer period.
The sub-scale of physical functioning produced from the RAND-36 does not allow a comparison of various sum-scores with the quality point of view. Thus it is difficult to interpret what kind of exact differences in difficulties people have if they have for instance scores of 12 or 14 points. Anyhow, high Chronbach’s alpha score of the scale used in the present study indicated very good internal consistency of the scale. Furthermore, many functional capacity scales have a clear hierarchy between the various abilities. It is shown, for instance, that over 90% of older people who have difficulty in dressing and undressing have also difficulty in mobility. In this sense, we can assume that people scoring higher in the present study have somewhat poorer functional capacity.