Does the millennial generation of women experience more mental illness than their mothers?

In the context of societies experiencing rapid social change, there are concerns that the mental health of more recent generations is declining. There is ample evidence of recent social and structural changes in behaviour and the lifestyles of younger generations. These changes are in four broad categories and encompass almost every aspect of life. Firstly, there have been fundamental demographic changes in family and work life that could have a negative impact on mental health [1‐3]. Societies are more urbanised than ever before, with higher density living, reduced opportunities for physical activity and increases in obesity [4]. Secondly, it is suggested there has been a shift from communal aspirations towards individual goals emphasising wealth, appearance, status and personal attainment [5, 6]. There is some evidence that younger generations are more narcissistic than previous generations – a shift to generation “me” from generation “we” [3]. Thirdly, more recent generations live in a media environment which dominates their day-to-day activities. More recent generations are more (electronically) connected, but possibly more physically isolated from social networks than ever before [7]. Fourthly, young people are delaying attaining what were previously important developmental milestones such as employment, marriage, having children and buying a home [8, 9].

Arguably these four types of societal changes could negatively influence mental health at a population level. However, it is not at all clear that current generations are experiencing more rapid or substantial change than did previous generations, nor that these changes lead to increased rates of mental illness. Arguably, all generations experience what are for that generation, major social change.

In the context of ongoing rapid social change, some suggest there have been generational increases in rates of mental illness [6, 10‐13]. A meta-analysis of 131 studies has suggested that there may have been increases in mental illness over time in North America but not in some European populations [14]. The above findings contrast with a body of research suggesting an absence of change in psychopathology over time or across generations. A systematic review of studies using similar measures of mental illness pointed to inconsistent results but with the possibility that there had been an increase in internalizing symptoms experienced by young females [15]. A meta-analysis of 26 studies, restricted to using concurrently administered measures of depression, found no change over time for child or adolescent psychopathology [16]. Other studies have found no changes [17] or no consistent changes over time in the prevalence of depressive symptoms or disorders [4].

The conflicting findings should be interpreted in context. There is no doubt that in economically advanced countries, there has been a substantial increase in the availability and use of treatment services for those with mental illness [18‐20]. It is unclear whether this increase is related to the incidence and/or prevalence of mental illness in a population or rather a consequence of increased level of service provision and/or mental health literacy in high income countries [21]. Much of the available data which is used to test the case for changes in mental illness over time involve comparisons of cross sectional surveys either using student [6] or population based samples [14]. In these circumstances it is not clear that the samples are effectively comparable.

Mothers in this study are primarily of the Baby Boomer generation (born 1946–63), with a small number early Generation X (born 1964–81). These are women who were reared in the context of the protest movements common in 1970s generation and the threat of a potentially catastrophic nuclear war. Women in this generation joined the paid workforce in increasing numbers as their children left home. This generation lived a relatively stable lifestyle coming from families where marriages rarely led to divorce, but were the first generation to divorce their partners in substantial numbers [22].

Daughters in this study were of the millennial generation, known as generation Y (born 1977–2001). In Australia this generation experienced long periods of almost uninterrupted economic prosperity. Historically this is the most educated generation of women ever. Compared to previous generations, they have delayed marriage and having children and more frequently adopt single marital status and single parenthood as a life choice. Most of the women in this generation are in the workforce and aspire to achievement [23].

In this study we aim to test the proposition that the different life courses experienced by the mothers and daughters in this study may have contributed to differences in their life histories of mental illness. We compare mothers and their daughters across four lifetime ever diagnoses: major depressive disorder, generalised anxiety disorder, panic disorder and post-traumatic stress disorder (PTSD). We hypothesise that there is a generational change in the rates of mental illness with daughters having higher rates of mental disorders than their mothers.

Table 1 compares key sociodemographic characteristics of mothers and daughters. With the exception of education, mothers’ sociodemographic characteristics are taken from the 5 year follow-up when mothers were a mean age of about 30 years. Mothers described themselves primarily as housewives, with a partial or complete secondary education, working overwhelmingly in clerical/service/sales positions (when employed) and of married status. Their daughters at 30 years are generally employed, the majority with a tertiary education, working either in clerical/service/sales or in professional and managerial positions. By 30 years of age 83.1% of mothers and 49.7% of daughters were married.

For daughters there are a strong and consistent set of associations between their sociodemographic characteristics and their lifetime ever experiences of most CIDI diagnoses (Table 2). Unemployed daughters have higher rates of major depression, generalised anxiety disorder, panic disorder, and PTSD. Daughters working manual occupations have higher lifetime ever rates of major depression, generalised anxiety disorder and PTSD. Daughters who are separated/widowed/divorced have higher rates of major depression.

For mothers the associations between their sociodemographic characteristics and DSM-IV lifetime ever diagnoses appear to be neither strong nor consistent (Table 3). Employed mothers have higher rates of lifetime ever generalised anxiety disorder, and unemployed mothers have higher rates of PTSD. Separated/widowed/divorced mothers have higher rates of lifetime ever major depression and PTSD. Women who are married have lower rates of major depression and PTSD.

In Table 4 we compare daughter-mother dyads and have only considered mothers’ diagnostic events if the age-of-onset was 30 years or less. Additional file 1 provides details of the age-of-onset excluded cases with about 50% of most mother age-of-onsets occurring after 30 years of age.

Comparing daughter-mother pairs, daughters have higher rates of lifetime ever major depression, generalised anxiety disorder and panic disorder. Daughters appear to experience almost three times the rate of major depression compared to their mothers, and almost twice the rate of generalised anxiety disorder and panic disorder. Adjustment for daughter and mother sociodemographic characteristics appears to have little effect on these findings. We have repeated this analysis in Additional file 1 (mothers and female children but without pairing mothers and daughters). The findings are generally similar to those observed in Table 4.

The sample was then stratified by illness duration. Comparing daughters and mothers with age-of-onset 30 years or less we note that the data are very similar suggesting there has been little or no change between generations in the age-of-onset of their disorders (Additional file 1). Considering persistent disorders (12-months or longer), generalised anxiety disorders and panic disorder, daughters have higher rates than do their mothers (Table 5). For persistent disorders there was no difference between the generations in rates of major depression and PTSD.

For less persistent disorders, daughters report substantially higher rates than their mothers for major depression, generalised anxiety and PTSD (almost all panic disorders are of persistent duration). These shorter duration episodes are arguably more likely to be subject to recall bias.

In Table 5 we further test the possibility that poorer maternal recall of past episodes may account for our findings. For both persistent and less persistent conditions we inflate maternal reports (cases) by 10, 20 and 30% and recalculate daughter-mother ratios of DSM-IV mental disorder. For the persistent disorders only Generalised Anxiety Disorder remains significantly higher for daughters compared to their mothers. For the less persistent disorders inflating maternal reports even by 30% does not substantially alter the finding that daughters report substantially higher levels of less persistent symptoms and disorders than do their mothers.

Despite some evidence that rates of mental illness in more recent cohorts have greatly increased [6, 12] and that rates of treatment for mental illnesses have been rapidly escalating [18‐20], doubts remain about the extent to which reported changes may reflect a greater willingness to report symptoms, an increased sensitivity to symptoms and/or the increased availability of treatment services [40]. There is also some evidence that rates of depression may have increased over time, particularly for younger women [41].

In the current study we have restricted our comparisons to a mother-daughter dyad and used a structured clinical diagnostic assessment based upon DSM-IV criteria, to determine whether there has been a generational change in four categories of mental illness.

Comparing mother-daughter dyads we find, as anticipated, that the daughters are mostly employed or otherwise in the workforce (a minority of mothers were employed), much more likely to have a university level education and much more often employed as professionals or managers. Some 46.6% of unemployed daughters meet the CIDI criteria for a lifetime ever major depression compared to 28.8% of unemployed mothers who report ever experiencing depression. Of the daughters who are employed in manual/trades groups, 41.1% report ever experiencing major depression compared to 23.0% of their mothers in this occupational category. The pattern of results suggests that the employment and occupational circumstances of daughters are more strongly related to their mental health than is the case for their mothers. In interpreting this finding, we speculate that the mental health of daughters may be more closely tied to their employment and workforce participation than was the case for their mothers. While there are marital status differences comparing mothers and daughters, the marital status of women with the highest rates of mental illness are those who are separated/widowed or divorced. For this category of marital status there appear to be (slightly) fewer daughters (6.1%) compared to mothers (8.3%). While there are changes in the characteristics of mothers and daughters than might lead to decreased levels of mental illness (fewer unemployed daughters than mothers, fewer separated/widowed/divorced daughters than mothers) there may also be changes that might lead to increased levels of mental illness (more daughters are single compared to their mothers). Overall our findings do not point to social and demographic changes in the life circumstances of mothers and their daughters that are likely to have a substantial impact when comparing the mental health of these two generations of women.

The direct comparison of the lifetime ever experiences of mental illnesses confirms that adjustment for the changes in the life circumstances of these two generations of women has no impact on the estimates of the extent to which daughters experience substantially higher levels of lifetime ever major depressive disorder, generalised anxiety disorder and panic disorder. However disaggregating the comparison groups into disorders of greater and lesser persistence, and further inflating maternal rates to provide for the lower levels of recall of past episodes, a possible consequence of the mothers’ older age at interview, suggests that higher rates of major depression and PTSD are of less persistent conditions. For generalised anxiety disorders somewhat higher rates are observed for daughters for both persistent and less persistent episodes.

We find that daughters are much more likely to report a wide range of less persistent mood disorders than did their mothers. Overall we conclude that the more recent generation of women only experience somewhat higher rates of persistent anxiety disorders. The most recent generation of women in this study also appear to experience higher rates of a number of more transient mood and anxiety disorders.

Three key issues arise from our findings. The first concerns whether the observed increase in mental disorders experienced by daughters might be attributed to underlying societal changes. Were this the case and given that societal changes are likely to be ongoing, treatment for mental illness might be characterised as a strategy for mitigating the harmful consequence of societal change. In the event our findings do not point to societal changes that may have contributed to the increase in episodes of mental illness. Second, if our findings are correct then the increase in mental illnesses over a generation is primarily in shorter term conditions which appear to remit over time. There is a need to consider the wisdom and value of increasing treatment services for conditions which are more transient and remitting. Third, we have noted the rapid increase in the use of treatment services by those with a mental illness. It is important to consider whether treatments for mental illnesses of shorter duration have different benefits than treatments provided to those with more chronic conditions.