Effect of social mobility in family financial situation and housing tenure on mental health conditions among South Australian adults: results from a population health surveillance system, 2009 to 2011

This study used representative cross-sectional population data from a surveillance system and demonstrated that people with low SEP in adulthood were more likely to experience poor mental health outcomes regardless of their family financial situation or housing tenure in childhood. Even when controlling for the effects of family structure at age 10 the association between low SEP in adulthood (downward mobility and stable low SEP) and mental health illnesses decreased by only 1 to 9 %.

These findings imply that policies to improve housing and providing financial support could have a positive influence on mental health outcomes in the population, to which there is no simple singular answer. To improve housing for both renters and owners is a complex undertaking which has been shaped over many decades through various interactions between legislation or regulation, government policies (state and federal), local and global market forces (for example the 2008 Global Financial Crisis) and family needs in terms of physical requirements and psychosocial needs such as privacy, safety and security [33]. Continual housing assistance programs including public housing and rent assistance for families has long-term benefits in terms of improved access to education, employment and income levels [34, 35]. Other studies have suggested better mechanism for first home buyers and policies to adapt to the ongoing risks (unemployment, reduced work hours, rise in interest rates, unexpected illness) faced by home owners with a mortgage who are financially stressed [33, 36]. Even though financial stress or the difficulties regarding housing tenure cannot be eliminated, programs to develop resilience and coping skills can be taught and possibly reduce mental illness later in life.

The association of downward social mobility with poorer mental health in this study is similar to a previous study of Norwegian men, where suicide and schizophrenia disability were associated with a combination of high parental and low own education level [11]. In a longitudinal study of older Americans, people experiencing upward social mobility had similar levels of depressive symptoms to those experiencing relatively high socioeconomic circumstances across the life course [12]. Cumulative effects of adult SEP on mental disorders were also not influenced by adjustment for childhood SEP in analyses of the 1958 British Birth Cohort [37]. Even taking into account education, which has been shown to influence SEP in adulthood [37, 38], poor mental health was more prevalent among those with low SEP in adulthood.

To our knowledge, this is the first paper of its kind in the Australian context and is unique in having a large sample on the general population so that the conclusions are likely to be generalizable to the Australian population. The prevalence of each mental health condition in this study was comparable to other Australian population estimates [2, 39–41]. The large sample size enable us to examine associations of SEP with suicidal ideation, an outcome of low prevalence. The design of this study is reliant on retrospective self-reported SEP data, and the memory of participants could result in over-or under-reporting of early life circumstances, with a high proportion unable to recall these circumstances. As a result, this could weaken the observed associations. The choice of these SEP variables at age 10 were based on previous studies that measured performance in recall; including how easily and accurately these circumstances were recalled, whether the indicators have the same meaning over time and for different population groups such as migrants and older people, and indicators that minimised socially desirable responses (for example embarrassment of early childhood circumstances) [13, 24, 42]. This study demonstrated that having missing responses for childhood housing tenure and family financial situation was associated with a few variables [24, 42], making recall suitable for inclusion in a surveillance system like SAMSS. However, these studies did not examine the accuracy of recall of childhood housing tenure and family financial. Other measures of childhood SEP, including parental education and occupation, have been shown to be less well recalled by adults [24] and are not included in this surveillance system. While 17 % of the response sample was excluded from analyses because of missing responses, few differences existed between the response and analysis samples. Missing data were mainly due to those who could not recall childhood circumstances: 2.2 % could not recall their housing situation and 11.4 % did not state or could not recall their family financial situation when they were 10 years of age. A small proportion (0.7 %) were excluded because their childhood circumstances were atypical, for example, in an orphanage, concentration camp, institution, displaced, refugee.

Another limitation of this study using self-reported data, is that the respondents may provide socially desirable responses, which may result in the under-reporting of having a mental health condition. Clinical assessment of mental health outcomes would be ideal but is costly and resource intensive, especially in a population surveillance system such as SAMSS. Using a mental health service or professional as an indicator is problematic as middle aged people are higher service users compared to the younger age group which has the highest prevalence [6]. Hence, our study is reliant on established, validated instruments such as Kessler 10 for psychological distress or GHQ-28 for suicidal ideation that are suitable for assessment via the telephone. Furthermore, due to the cross-sectional nature of the study, the mental health instruments focus on the most recent four week prior to the interview, or current treatment, therefore, they only measure current mental health status; not past mental health episodes. Again, this could result in under-estimation of the strength of the association. Our study was unable to adjust for factors (were not included in SAMSS) which have been shown to contribute to mental illnesses in adulthood. These include family history of mental illness, family relationships or social networks during childhood, childhood mental illness [43] and factors such violence, abuse, neglect, lack of family cohesion [38, 44]. Possibly, those with poor physical and mental health are more likely to experience downward social mobility or a barrier to upward mobility [45]. Possible bias from using only listed telephone numbers can be considered a weakness of this study. Although telephones are connected to a large number of Australian households, not all are listed (mobile-only households and silent numbers) whereby 9 % of South Australian households are mobile-only and 69 % of households do not have their landline or mobile number listed [46]. Previous work has shown that inclusion of unlisted landline numbers in the sample did not impact on the health estimates [47]. The proportion of mobile-only households have increased in South Australia to 21 % as of 2011 [48, 49] with a very small proportion (4 %) listed in the telephone directory [46]. The impact on health estimates by excluding this group from the sample in Australia is unknown to date. However, the characteristics of people living in mobile-only households have been found to be distinctly different. Hence, potential bias may exist in the estimates obtained in this study.