The objectives of this study were to compare different types of mental ill-health with respect to symptom severity and prevalence of caseness, the latter in both absolute and relative terms, in men and women in different age groups. Data from this large population-based survey (n=3406) showed sex- and age-related differences in mental ill-health, which in several cases were of clinical relevance. Compared to men, women in most age groups had higher levels of symptom severity and caseness prevalence. For example, there was a three-fold increased risk for somatization caseness, and a two-fold increased risk for anxiety and burnout in women in general, compared to men, which is in line with previous studies [49-53].
Interaction effects between age and sex were clearly noted regarding both symptom severity and caseness prevalence. Young women constituted a particular risk group with high levels of severity in all types of mental ill-health, and high prevalence of caseness in both absolute and relative terms, and with a fairly linear decrease with age. The prevalence for anxiety, insomnia and somatization for the youngest women was higher than for the youngest men. This pattern has also been reported from prior studies [54-56]. In the oldest age group there were only very small differences between men and women regarding both severity and prevalence.
In men, the lowest symptom severity and prevalence was, in general, found in the age group 60-69 years, whereas the oldest age group had significantly higher prevalence of symptoms. It is reasonable to expect that failing physical health, social isolation, bereavement, loss of status, loss of friends and reduced income are more common among older people, thus each condition being a risk factor for mental ill-health [57,58]. A tentative interpretation of the higher prevalence of symptoms among the oldest men, regarding both somatization and depression, can be explained by overlap between certain depressive symptoms (e.g. fatigue, diminished appetite, weight loss) and common somatization symptoms . Men aged 30-39 years had the highest prevalence of insomnia compared to other men, and men 40-49 years had the highest prevalence of depression and burnout caseness in both absolute and relative terms. The age 40-49 years represents a consolidating career stage where work-life balance, family responsibilities, ageing parents all are part of a life puzzle that have to be solved [60,61].
With focus shifted to specific types of mental ill-health, the youngest women had very high symptom severity and prevalence of caseness for anxiety, with a 7-fold increased risk for caseness compared to men aged 60-60 years. These measures of extent were lower in the oldest age group, the differences between men and women were rather small. Findings of similar kind have been reported in the past   [64,10,11,55].
An unexpected result is the relatively small difference between men and women in depression. Earlier studies indicate that women experience depressive states more frequently than men, with the ratio of about 2:1 [65-67]. In this study the prevalence rate for depression was 5.5% for men and 7.1% for women, and there was no significant interaction between sex and age regarding depression, neither in severity nor prevalence. The results show a tendency of symptom severity being highest in men 70-79 years. However, the prevalence of depression in absolute terms shows large overlap between men and women in all age groups, making the result somewhat different to other cross-sectional studies [68,62]. Interestingly, similar prevalence studies in other Nordic countries using HADS show higher prevalence rate in men than in women in certain age groups [69,70]. Whether these outcomes is a result of specific environmental factors in the Nordic countries or use of the HADS vs other instruments may be explored in future studies. It is possible that men and women simply are rather similar in this respect in the studied population. A tentative interpretation of the results in the clinical context may be that general practitioners in healthcare settings need to be more attentive to depression in men, since we know that men seek healthcare to a lesser extent, and when seeking healthcare they present to a greater extent other than psychological concerns .
Insomnia, was a common condition in the sample. The overall prevalence rate was 28.6% when using the sleep index , and was particularly high among the young and middle-aged. The prevalence was high compared to other studies, which may be due to different ways of defining poor sleep [16,15,18,19,17]. Men aged 30-39 years peaked in symptom severity and prevalence, with higher prevalence than in women of the same age group, but with substantial overlap for prevalence. This sex difference is in line with other studies [72-74]. Moderate sized correlation coefficients (0.40-0.53) have been reported between sleep quality and anxiety, depression, stress, and mental/physical exhaustion, which is expected since sleep quality is a common complaint in these conditions, and regarded as a transdiagnostic process in many psychiatric disorders [40,75].
The overall prevalence of burnout (17.3%) was higher in this study than in other (7.1%-12.9%) [21,64], which partly may be explained by the relatively low cut-off score (4.0) used in this study. The severity and prevalence was higher in women than in men, and decreased with age. This is coherent with mainstream theory about burnout . The prevalence for burnout in men seems to have an age-span pattern that is similar to that for anxiety and depression, with a peak at 40-49 years and in the oldest age group. In contrast to Maslach’s early definition of burnout and corresponding questionnaire instruments with focus on working life, SMBQ measures burnout due to general aspects in life, and is appropriate also for older populations. Accordingly, and in line with other studies, the present data suggest a non-linear association between age and burnout .
The prevalence rate for somatization was higher in this study (17.4%) than in a population study by Koncalevent et al.  (9.3%) that also used the PHQ-15 [28,2,29]. The prevalence rates in this study are more in line with those from healthcare settings outside Sweden . The differences across age and sex for severity was very similar to that for prevalence. Similar to other studies, women generally reported more bodily distress and more frequent somatization symptoms than men . None of the age groups had an overlap in 95% confidence intervals for prevalence, except for the oldest age group, in which men had significantly higher prevalence than other age groups.
A particular strength of this study lies in its broad approach to investigate both symptom severity and caseness prevalence in both absolute and relative terms of various types of mental ill-health in different age groups and in men and women. This enables direct comparisons between the two aspects of extent, type of mental ill-health, age groups and sexes as well as their interactions in a general adult population. Other strengths of this study include being population-based, stratifying the sample for age and sex, having a large sample size, and that the study population had an age and sex distribution that is very similar to that of Sweden in general [79,35]. The present sample (Table 2) is similar to Swedish adults in general according to national data: smoking 13%, risk alcohol consumption 17%, perceived stress 13%, regular exercise 65% and perceived health 73% .
There are also methodological limitations to be considered. The data was collected in 2010, which limits the generalizability of mental ill-health to today. Only 40% of the sample responded to the questionnaire, compromising the representativeness. This refers, in particular, to young men, with consequences for the overall mean severity scores and prevalence rates, which are likely to be somewhat overestimated, making the interpretation of results from the youngest male cohorts (18-39 years) uncertain . The sample size was large, resulting in statistically significant group differences even for small absolute differences, which highlights the importance of the effect sizes. In this study eta-square values were small, implying that sex and age do not show substantial differences in mental ill-health in a general population. This sums up that even though sex and age to some extent can explain differences in mental ill-health, complementary studies of other mediating and moderating factors are necessary. Diagnostic interviews may have provided more reliable results than presently used questionnaire instruments. However, as reviewed in the methods section, the reliability and validity of the instruments used are in general good.
Cross-sectional studies of this kind cannot answer the question as to whether an age-related variation in health status is due to environmental and societal change, or whether it is due to transition phases, e.g. developmental changes, thus seasons of a person’s life . For this reason the findings may not necessarily mean that ageing itself brings about a diminution in symptoms. However, the WHO argues for higher prevalence rates for anxiety and mood disorders in more recent cohorts in many countries (Andrade et al., 2000). Additional longitudinal studies covering the adult life span are needed to differentiate ageing from cohort effects .
A clinical implication of the present results is that there is a gray zone of individuals outside the healthcare system who do not seek help. In this study the ratio in prevalence between those with caseness of a certain type of mental ill-health and a physician-based diagnosis (Table 2) varied from 4.17 to 6.85, indicating that these conditions are undertreated in the general Swedish population. The prevalence rates obtained in the current study are in line with prior prevalence studies in Sweden [11,81].
Many patients with mental ill-health present to their general practitioner with common somatic symptoms that may result in their mental ill-health being missed in the assessment [82,83,71]. The high prevalence rates found in the current study calls for further efforts regarding development and validation of good prevention interventions and treatments in primary healthcare settings . Such procedures may be particularly valuable for men since they seek healthcare to a lesser extent than women , and are more likely to commit suicide [86-88].