The contextual contribution to individual health

Background Many Countries consider health as a fundamental right for their citizens. In terms of health policy, this means equal opportunity of prevention and access to care guaranteed to all. Thus, a major concern in research has been given to health inequalities and their determinants. A body of evidence has shown that individual characteristics primary affect health, however contextual factors also substantially contribute to the final perception of health and/or mediate the effect of the individual determinants (Kawachi and Subramanian, 2005). Contextual factors could involve geographically the location where the person lives, but also networks to which the individual belongs in terms of culture, socio economic status, family ties. Geographical differences in health can occur as a response to many different characteristics of places, such as environmental conditions, economic background, social context, but also public health resources. Whether health resources have an affect on population health has been much debated (Joumard et al. 2008) and to date there is no conclusive evidence on this issue. At a lower level, the household also acts like a cluster, being the first grouping structure in the society. Family members exhibit similar patterns of morbidity (Johnson et al, 1965; Monden 2007; Merlo et al. 2012), health-related behaviors (Rice et al. 1998), help-seeking behavior (Cardol et al. 2005) and utilization of health services (Sepheri et al. 2008). On the one hand, household’s characteristics, such as socio-economic level of the family, housing conditions, household structure and the burden of care for ill-health member can be hypothesized as determinants of health for the household’s members. On the other hand, health perception itself can operate as a determinant for health status of other members within the same households. Stated differently mutual influences in health perception can be exerted within the households and they result in a high resemblance of health status for people living together. Very limited research have dealt with territorial and familial influences on health perception, and no studies of this kind have been carried out for the Italian context, where paradoxically the magnitude of the family on demographic behaviours is extremely pronounced. This research aims to provide an estimation of the influence that the context (both territorial and relational) has on self-perceived health and to gain a better understanding of the pathways through which this influence is exerted. Methods This is a population-based cross sectional study. Data come from the Italian Health Survey “Condizioni di salute e ricorso ai servizi sanitari” (2004/2005). The survey has a cluster sample design based on households and representative of the population at sub-regional level, with the definition of the so called “Large Areas”, an aggregation of neighboring Local Health Providers (ASL) constituting a unit for health planning. We selected three outcome variables to investigate perceived health: Physical Component Summary – PCS, Mental Component Summary – MCS and poor self-perceived health –poor SPH. The first two measures are a quantitative assessment of physical and mental health conditions as perceived by the respondent through a standardized questionnaire (SF-12), the latter is a binary variable indentifying people reporting “poor” and “very poor” health conditions in the WHO question “how is your health in general?”. Data have a hierarchical structure defined as individuals (level 1) living in different households (level 2), which are, in turns, located in different “Large Areas” (level 3). We adopted a multilevel approach which is entirely coherent with this structure of the data and allows to obtain unbiased . By means of this approach we have been able to capture determinants of perceived health at different levels and to disentangle the proportion of variability due to differences between individuals, households or Large Areas. According to the characteristics of the outcome we run linear and logistic multilevel models, with random intercepts at the household and Large Area level. Results: We documented a very limited, although always significant, impact of area of residence on self perceived health (0.3% for PCS, 0.6% for MCS and 2.3% for poor-SPH, adjusted for individual covariates). This result is partially in contrast with previous works illustrating a health gradient for objective and subjective health in Italy (Costa et al. 2003; Mazzuco 2009). However, researches that adopted a multilevel approach to investigate the Italian Regional/Large Area health heterogeneity came to our same conclusion, recognizing a proportion of variability at Large Area level lower than 3% for poor self-perceived health among the elderly population (Pirani and Salvini 2012b). By contrast, the relevance of household on perceived health was quite substantive. The 15% of variability in PCS is due to household differences, but MCS and poor-SPH show an even greater impact with, respectively, a 33% and 38% of variability at the household level. The characteristics of the household (e.g. economic resources, family structure, size of the municipality of residence) are significantly associated with perceived health, but they explain a very little amount of the overall variability between households. We hypothesized then that mutual influences between family members can have a role in explaining the heterogeneity in household health. We investigated this hypothesis by observing the pattern of health resemblance by family structure and found that the similarity in health status was higher in those circumstances where the link between members were expected to be tighter (2 components, marriage-like link, mono-nucleus families). These results are in line with findings from social psychology, which consistently documented a similarity in mental illness, depressive symptom and distress (Meyler et al. 2007; Monden 2007), and with those from sociology reporting a positive effect of partners interactions on well-being, happiness and life satisfaction (White 1983). By conducting the research by means of three health outcomes we were able to cross-validate the results and shed further light on the different profile of determinants for physical and mental conditions. More particularly, we illustrated how PCS is affected by very concrete agents such as age, education, economic conditions, whereas MCS is particularly linked to socio-relational dimensions at the individual and household level. This result is coherent with the knowledge about determinants of mental health, which include elements as social support, perceived stress and self esteem (Bovier et al. 2004). Conclusions With this research we provided evidence of the existence and the complexity of contextual factors influencing perceived health. Geographical differences seem to be overestimated when a multilevel approach is not taken into account. On the other hand, households are often neglected as a level affecting health perception, whereas they deserves as much attention as individual in the study of health.