Data

For this study the output data-set of a microsimulation model, Health&WealthMOD, was analysed. Health&WealthMOD was specifically designed to measure the economic impact of ill health on Australian workers aged 45 to 64 years. It is Australia's first microsimulation model of health and its associated impact on labour force participation, personal income and wealth and government revenue. Health&WealthMOD was built on the unit record data of the 2003 Survey of Disability, Ageing and Carers conducted by the Australian Bureau of Statistics ¹³ and the output data of a microsimulation model of income and wealth, STINMOD, developed by the National Centre for Social and Economic Modelling for the Australian Government. ^{Reference Lambert, Percival, Schofield and Paul14}

The Survey of Disability, Ageing and Carers provides detailed self-reported data on sociodemographic status, labour force participation and health and disability status (such as chronic conditions) for each individual in Health&WealthMOD. Respondents’ health conditions were classified by the Australian Bureau of Statistics using ICD–10 ¹⁵ codes. People who reported their main long-term health condition as depression/mood/affective disorders (excluding postnatal depression) (ICD code F30–39) were considered to have ‘depression’. Those who reported their main long-term health condition as mental and behavioural disorders, dementia, schizophrenia, phobic and anxiety disorders, nervous tension/stress, attention-deficit disorder/hyperactivity, and other mental and behavioural disorders (ICD codes F00–29, F40–99) were categorised as ‘other mental illness’ for this study. Depression was analysed separately from other mental illnesses owing to its high prevalence in the population.

In this study those who reported themselves as being out of the labour force because of their illness and listed depression as their main condition were considered to be out of the labour force because of depression. Those who reported being out of the labour force because of their illness and listed one of the other mental illness conditions as their main condition were deemed to be out of the labour force as a result of other mental illnesses. All people who are out of the labour force, regardless of the reason for it, are assumed to be permanently retired.

Economic information such as income and wealth were imputed from STINMOD. The imputation was carried out by synthetically matching person records with similar sociodemographic and labour force participation characteristics in the two data-sets. Synthetic matching was based on nine matching variables that were common to both data-sets and strongly related to income: gender, income unit type, type of government pension/support, income quintile, age group, labour force status, hours worked per week, highest educational qualification and home ownership. The data were reweighted to represent the 2009 population and up-rated to reflect the demographic, labour force, income and wealth, and other changes that had occurred between 2003 and 2009. The development of Health&WealthMOD is explained in detail elsewhere (details available from the author on request).

Statistical methods

Logistic regression models were used to compare the odds of owning wealth by those who reported being out of the labour force owing to depression and other mental illness (that is, people who were out of the labour force owing to ill health, and identified depression or mental illness as their main health condition) with those who were in full-time work and had no chronic condition. Multiple linear regression models of the log of the value of wealth were used to analyse the differences between the wealth of people working full time with no chronic condition, people working part time with no chronic condition and people not in the labour force because of depression and other mental illness. Full-time work with no chronic condition was used as the reference group. Four different classes of wealth were included – cash, superannuation, value of owner-occupied home and value of other financial investments – as well as the total value of an individual's wealth, which was calculated as the sum of all the above four classes of wealth, and analysed in this paper.

Analyses were separately conducted for all four different classes of wealth and total wealth in both the logistic regression and multiple regression models, adjusted for age group, gender and highest education. Multiple regression analyses were undertaken on log-transformed data in order to satisfy the assumptions of linear regression analysis, and regression diagnostics confirmed that the assumptions were reasonably satisfied. A conservative measure of the impact of retirement as a result of depression or other mental illness on the wealth of individuals was used: only including individuals with depression or another mental illness and who stated that their reason for being out of the labour force was their ill health. A sensitivity analysis was undertaken by including all individuals who had depression or other mental illness and were not in the labour force – regardless of their stated reason for not being in the labour force. The results for assessing the value of total wealth were repeated using this measure.

In order to estimate the results for the entire Australian population in the 45–64 years age group, we performed weighted analysis using weights that represented the number of similar individuals in the Australian population. All analyses were undertaken using SAS V9.2 for Windows. All statistical tests were two sided with the significance level set at 5%. All results are presented with their 95% confidence intervals.

Main finding

This study has quantified the amount of wealth held by people who have retired from the labour force as a result of depression or other mental illness and also the difference in this amount compared with those who are in the labour force with no chronic ill health. Those out of the labour force because of depression or other mental illness have significantly less wealth. These individuals are also more likely to have the wealth that they do have in cash assets, rather than higher-growth assets such as superannuation, home equity and other financial investments.

The comparatively few individuals out of the labour force because of depression and other mental illness who have any wealth in superannuation is a reflection of the impact of the lower labour force participation rates for these individuals. The impact of depression and other mental illness on labour force participation is significant – those with depression are 6.71 times more likely to be out of the labour force, and those with other mental illness are 5.71 times more likely to be out of the labour force than those with no health condition. ^{Reference Schofield, Shrestha, Passey, Earnest and Fletcher5} (Although depressive episodes do fluctuate, it has been found that 44% of people with depression have reoccurring episodes, and individuals with depression spend as much as 19% of their time in a depressive state ^{Reference Vos, Haby, Barendregt, Kruijshaar, Corry and Andrews17} – which indicates why depression, although a fluctuating condition, does cause labour force disruption.) Being absent from the labour force because of depression results in 43% lower income than being in the labour force full time; and being absent from the labour force owing to other mental disorders results in an income 37% lower (details available from the author on request). Thus those out of the labour force owing to depression and other mental health conditions do not have the same income with which to accumulate wealth assets; they may even be drawing upon these assets to supplement lost income. However, because of data limitations it is not clear how long the individuals in this study have been out of the labour force.

The full cost of mental illness

The findings add further support to the arguments that highlight the full extent of the cost of depression and other mental illness. It has been noted that mental illness alone is responsible for 13.3% of the burden of disease in Australia, only falling behind cancer and cardiovascular disease. ^{Reference Begg, Vos, Barker, Stevenson, Stanley and Lopez18} This highlights the cost of the disease to the health of the Australian population. However, the economic costs are also significant. The cost of treating mental health conditions was estimated to be AU$3741 million (£2058 million) in 2000–01. This estimate covered health expenditure in hospitals, aged care homes, non-hospital medical services, pharmaceuticals, research and community and public health. ¹⁹ However, it should be noted that only 62% of those with a mental illness seek medical help in Australia ^{Reference Andrews, Hall, Teesson and Henderson20} and thus the potential direct medical costs may be higher if adequate services were available.

Fig. 1 Flow chart of labour force status for those with depression and other mental illness within the Australian population aged 46–64 years.

In addition to the direct health system costs, mental illness also has significant additional costs to society, which are incurred indirectly through lost labour force participation for example. This study has quantified the indirect costs of mental illness through early retirement and the effect this has on individuals’ wealth. However, there are additional costs because of early retirement. In previous studies it has been found that those retiring early as a result of depression pay 96% less tax (tax on income from all sources), and those with other mental illnesses pay 91% less tax than those who are healthy and remain in the labour force (details available from the author on request). Further, in 2006, 101 936 people with psychological/psychiatric disorders who were over the age of 45 claimed a Disability Support Pension, adding to the cost of mental illness to the state. ²¹ (The Disability Support Pension is AU$644.20 (£354.31) per week for single people and AU$485.60 (£267.08) each for couples. In total, AU$1468 million in Disability Support Pension payments are made to Australians aged 45–54 years of age per annum).

Analyses of the numbers needed to treat (NNT) to achieve good clinical outcomes for people with depression have demonstrated high efficacy: most antidepressants an NNT<5. ^{Reference Katona and Livingston22} Given the low cost and cost-effectiveness of these drugs ^{Reference Barretta, Byforda and Knappa23–Reference Doyle, Casciano, Arikian, Tarride, Gonzalez and Casciano25} the cost to the government of appropriate treatment is likely to be more than offset by reductions in the Disability Support Pension payments.

The Australian universal healthcare system and more generous social security system may to some extent help to protect individuals with depression and other mental illness from loss of all their savings and associated deterioration in living standards. Nonetheless, the savings of those with depression and other mental illness are still significantly compromised relative to those who are well. This study showed that if people out of the labour force because of depression or other mental illness do have any wealth, on average they have AU$236 727 (£130 200) and AU$148 771 (£81 824) in accumulated wealth respectively. This is far lower than the amount held by those in full- and part-time employment without ill health.

This lower amount of accumulated wealth is likely to result in lower living standards for these individuals. Wealth is generally used to protect against unexpected events (such as a deterioration in health), finance large financial commitments (such as children's education), and also allow for general daily consumption. ^{Reference Cagetti26–Reference Dvornak and Kohler28} Indeed Dvornak & Kohler found that a $1 increase in the value of wealth stored in housing assets is likely to result in a 3% increase in consumption spending. ^{Reference Dvornak and Kohler28}

Findings from other studies

The quantification of lost savings and wealth because of mental illness has not been reported in other studies. Furthermore, there have been few studies on the impact of ill health in general on savings or wealth with which to compare these findings. However, one study conducted in the USA by Covinsky et al found that 38% of patients aged 45–64 lost all or most of their savings owing to illness. ^{Reference Covinsky, Goldman, Cook, Oye, Desbiens and Reding29} More recently, another USA study by Swoboda & Lipsett found similar results, with 36.7% of families losing most of their savings as the result of having a family member critically ill for 12 months or more. ^{Reference Swoboda and Lipsett30} However, these studies only considered the impact on savings, and did not consider other wealth assets available to individuals such as home equity or investment properties, nor did they quantify the value of wealth held by individuals, as our study did.

Implications

Although comorbidities may contribute to the reduced labour force participation, and hence income and wealth accumulation, of people with mental illness, the analysis undertaken focused on people who identified their mental illness as their main health condition, and their health condition as being the reason for retirement. Other people who identified other illnesses as their main condition may have a concurrent mental illness that also contributes to their non-participation in the labour force. Thus, it is likely that our analysis has underestimated the total impact of mental illness on individuals’ wealth accumulation.

The Council of Australian Government has developed a National Action Plan on Mental Health 2006–2011, which is based upon a commitment to the prevention of mental illness and early intervention (depression was included under the broader term of ‘mental illness’ in the plan). ³¹ Prevention and early intervention is anticipated to reduce the severity, length and reoccurrence of mental illness, and will focus on increasing access to health services. ³¹ New Horizons: A Shared Vision for Mental Health, published by the UK Department of Health, explores the need for prevention and early treatment of mental illness. ³² It recognises that doing this will have economic benefits because of the large financial burden of mental illness on the state.

Preventing mental illness is likely to be a cost-effective activity for governments, and in the long term save on the costs of financially supporting those who are forced to exit the workforce early because of their ailment and thus are unable to support themselves financially. This paper has shown that those who have retired early as a result of mental ill health will not only have a loss of immediate income from employment but also have very low value of wealth assets and savings – thereby drastically reducing their ability for financial self-sufficiency. Thus the responsibility of providing financial support for these individuals falls to the state – a hefty financial burden that may be avoided through actions to prevent the occurrence of mental illness.

The ultimate aim of the Australian National Action Plan on Mental Health 2006–2011 is to allow individuals to live ‘independent and fulfilling lives’, which includes increasing capacity for labour force participation. ³¹ Although this study looked at the impact of depression and other mental illness separately, and showed that each condition has different effects on wealth accumulation, the findings of our study highlight the need for such overall preventive action, in light of the costs to both individuals and to governments as the result of early retirement because of ill health.