Retrospective observation of mental disorders during postpartum period: Results from the Singapore mental health study

Physiological, hormonal, psychological, and social role changes make pregnancy a time of potential vulnerability for various disorders [1‐4]. A systematic review by Fisher et al. (2012) for instance, found the prevalence of pre- and post-natal disorders in low income countries to be 15.6 and 19.8 % respectively [4]; whereas, Gavin et al. (2005) reported the estimated prevalence of perinatal depression to range between 6.5 and 12.9 %, with 19.2 % of women found to have a depressive episode within the first 3 months postpartum [5].

Several risk factors have been suggested for psychiatric conditions during pregnancy and the postpartum period. These include - but are not limited to - lower socio-economic status, lower educational level, younger age, not being married, lack of support and empathy from partner, domestic violence, and previous psychiatric history [2, 4, 6‐12]. Evidence however, has been inconclusive owing to the lack of standardization in survey instruments and the use of different cutoff points for screening tools in determining clinically significant symptomatology [8].

Anxiety and depressive disorders in particular, have received extensive attention in the area of pre- and post-natal research given their high co-morbidity [13] and the adverse effects that these disorders have on both the mother and the infant [1, 4, 6]. With regards to postpartum depression, antenatal depression or anxiety [2, 6, 8] has been consistently implicated as a significant risk factor. In a multicenter prospective study among Japanese women for instance, Kitamura et al. [14] found antenatal depression to be a significant risk factor for postnatal depression. This finding was replicated by Leung et al. [15] among postpartum women in Hong Kong whereby antenatal depression (and perceived stress) significantly predicted an approximate 28 % of the variance in postpartum depression. Research has alluded this to women with antenatal depression being more prone to experiencing a difficult labor and distress about the birth which in turn increases the likelihood of postpartum depression [16].

Antenatal anxiety on the other hand, has been linked to physical defects in the infant, low birth weight, and poor cognitive and neurodevelopment. Anxiety disorders/symptoms are commonly reported during the pre-natal period during which the mother may be concerned about the impending delivery of the child [17] and the associated changes in social roles and responsibilities that this brings forth [7]. In particular, obsessional thoughts pertaining to fear and concern of causing harm to the infant are common [18]. This phenomenon has been postulated to be especially pronounced in an Asian context given the various taboos associated with childbirth and the expectation of Asian women to follow certain traditions to ensure the well-being of the infant [19].

Although pre- and post-natal research has been conducted in the local population, studies are relatively limited. Thiagayson et al. [20] for instance, found the rates of major depression, minor depression, and anxiety disorder to be 11, 7, and 12.5 % respectively among a sample of Singaporean women who were at high obstetric risk. Conversely, in a prospective cohort study by Chee et al. [21], almost one-third of Singaporean women who engaged in the postpartum convalescence period (also referred to as “doing the month” for 30 days after childbirth) quoted it as negative experience which in turn significantly contributed to their depressive episode.

Given the limited research with regards to pre- and post natal psychiatric disorders particularly within the local context, the current paper aims to further build on the understanding of the epidemiology of women’s reproductive mental health by utilizing data from the Singapore Mental Health Study (SMHS) to explore the temporal relationship of pregnancy and child birth on mental disorders (i.e., mood disorders, anxiety disorders, alcohol use disorders). Furthermore, given the mixed evidence with regards to socio-demographic factors as potential risk factors for postnatal mental disorders, the study aims to examine this in the multi-ethnic population of Singapore as this may aid clinicians in identifying individuals who may be at a higher risk of developing psychiatric conditions during this period.

The findings from this study suggest a considerably lower prevalence of mental disorders during pregnancy and postpartum period among women who have children compared to that reported in other studies (e.g. [4, 5]). These rates were low even when compared to estimates reported in Asian countries. Klanin et al. [12] for instance, reported prevalence of postpartum depression to range between 3.5 and 63.3 % in Asian countries.

One reason for the low prevalence of postpartum depression among Singaporean women as compared to other Asian countries could be due to Singapore ranking relatively higher in terms of HDI compared to other countries [23]. The high level of development not only allows individuals to have access to good quality healthcare services but also allows them to seek proper treatment for these psychiatric conditions. This is in contrast to low-income and low-middle income countries whereby poverty not only limits access to proper and timely health care but also restricts their ability to afford these services [31]. Likewise, greater gender equality in terms of access to education and command over economic resources also means that females in Singapore are in a more privileged position to be able to afford and utilize these healthcare services relative to lower income countries whereby gender disparity may limit the autonomy of women in seeking out these services [31].

In contrast to our study, Chee et al. [21] found 18 % of pregnant women in Singapore to have either major or minor depression and 12.5 % to have anxiety disorders, which were relatively higher than our finding of 0.9 % for mood disorders and 0.7 % for anxiety disorders. One possible reason for this discrepant finding between Chee et al. [21] and our finding is likely due to the use of different instruments in assessing mental disorders. While our study utilized the CIDI to establish diagnoses of mental disorders including depression, Chee et al. [21] used the Edinburgh Postnatal Depression Scale (EPDS) designed specifically to measure postpartum depressive symptoms. Though well-validated, the CIDI is not specifically developed to assess psychiatric symptoms associated with disorders during pregnancy and postpartum [32]. For instance, particular symptoms such as changes in appetite or weight, fatigue, sleep disruptions in the CIDI have been reported to be poor in terms of discriminating between clinical depression and ‘normative experience of pregnancy’ given the overlap of symptoms between these two [11, 32]. Thus, while severe cases of postnatal depressions could have been detected, less severe presentations could have been overlooked as natural consequences of childbirth [33]. Furthermore, given some evidence to suggest the use of clinical interviews to yield lower prevalence rates compared to self-report measures, the rates in the current study could have been underestimated to some extent [34].

Conversely, past studies have suggested some difference in symptom expression between the Western and Asian populations. For example, Oates et al. [35] found Asians to have a higher tendency to somatize their symptoms compared to the Western population whereas Hwu et al. [36] postulated Asian cultures to have a ‘relatively high tolerance for or denial of emotional sufferings’ (also known as cultural stoicism). In line with this, Liao et al. [37] reported the prevalence of MDD symptoms to be lowest among non-institutionalized Taiwanese adults followed by Koreans and Americans (ECA and NCS study) [38]. Given that the CIDI has been primarily validated for use in the Western populations [39], the tendency for Asian individuals to repress their emotions as well as somatize symptoms may have resulted in a response bias [37] which may in turn have resulted in lower estimates of mental disorders obtained in this study.

Likewise, the lower prevalence could also have been a result of under-reporting due to a lack of awareness of mental illness, reluctance to report symptoms to a stranger due to shame or stigma [16, 39] as well as recall bias, whereby the women may have been overwhelmed by the tasks of taking care of the infant, and other changes in their life that they have “sealed over” any occurrence of mental health issues.

Despite the low prevalence of mental disorders during pregnancy and the post-partum period, the current findings were in line with past literature which has shown the prevalence of mood and anxiety disorders to be relatively higher during the postpartum period [5, 27]. For instance, though the prevalence of onset of any mental disorder was higher among women in the general population compared to women with children, the prevalence of both mood and anxiety disorders were comparable in both populations.

The onset of anxiety disorders in the current study was found to be associated with the time of pregnancy and particularly in the postpartum, with OCD being the most prevalent mental disorder within 2 years after a woman’s pregnancy. This finding corroborated with past research which have shown anxiety disorders to be common during the postpartum period [18, 27, 40, 41]. Evolutionary theories for instance, posit that intrusive thoughts and obsessional thoughts may to some extent represent adaptive behaviors meant to protect their infants from possible harm [40, 42]. The current finding of high prevalence of OCD was also partially supportive of an earlier study conducted by Chong et al. [39] who found the prevalence rates of mental disorders, with the exception of OCD in the adult Singapore resident population to be considerably lower than those in western countries. Similarly, Subramaniam et al. [43] also found the prevalence of obsessions and compulsions to be higher in the Singapore adult population compared to the US. The relatively higher prevalence in OCD observed in both the general and post-partum female population might suggest a higher risk for this disorder among the local population and warrants further investigation.

With regards to socio-demographic variables, only marital status and employment stood out as significant factors in influencing women’s mental health in the current study. Divorced and separated women were more likely to have an onset of mental disorder after pregnancy, or within the 2 year period following pregnancy, compared to married women. These findings paralleled those of Subramaniam et al. [44], who found single mothers to have significantly higher odds of having mood disorders as compared to those who were married. Previous studies have found the role of social support, particularly support provided by spouses to be a protective factor against mental disorders during this critical period [4, 45]. This finding should however be interpreted with caution, as marital status was obtained at the time when the respondent was being surveyed and as such may not accurately reflect the marital status of respondents at the time of pregnancy and during the postpartum period. Furthermore, it is important to note that the finding implies a correlational rather than a cause-effect relationship and that the onset of mental disorder may predate the change in marital status.

Absence of chronic physical conditions and being employed also served as protective factors for women in the current sample. Women who were married, employed and physically healthy were less likely to have developed any mental disorder than those who were not. This was in line with past studies which have shown low socio-economic status and financial difficulties to be strongly associated with postnatal psychiatric disorders [46, 47].

In contrast to past local studies, ethnicity was not found to be significantly associated with the onset of mental disorders. Chong et al. [48] for instance, in examining the prevalence of depressive symptoms among outpatients with diabetes in Singapore, found Indian ethnicity to be significantly associated with a high risk of depression. Similarly, Chong et al. [39] found the prevalence of depression to be low among adult residents, particularly the Chinese as opposed to other ethnic groups. We are however unable to account for the negative finding in this study.

There are limitations in our study. We were unable to examine the effect of miscarriage, abortion, child death, and gender-based violence on females’ mental health in the current study. Past literature for instance, has suggested a link between gender-based violence (including emotional and physical abuse) during pregnancy and the incidence of psychiatric disorders [4, 49]. Relatively prevalent among socially and economically disadvantaged women, factors such as bias against female babies, role restrictions, as well as multi-generational households in which the daughter-in-law has little autonomy have been identified as possible risk factors for gender-based violence [4]; these factors however were not explored in this study. We were also unable to detect the cases of subclinical levels of depression, and panic disorder as the study did not establish these diagnoses. Estimation of age at first pregnancy was an approximate made from subtracting the age of the first biological child from the respondent’s age and further deducting 1 year (as opposed to 9–10 months to account for pregnancy). Given that not all pregnancies last 9 months, this may have resulted in an over-estimation of the prevalence of mental disorders during the postnatal period. Furthermore, as the CIDI does not contain a specific question targeted at female respondents with regards to reproductive events, we were not able to gauge a more precise pregnancy age. Additionally, we were not able to differentiate the effect of ageing and child birth on respondents’ mental health status, all of which are important factors that can impact women’s mental health [50]. The retrospective design of the current study also limits inference of causality. The low prevalence in the current sample warrants careful interpretation of results given the insufficient statistical power. Future studies can look into utilizing a prospective study design to better understand the impact of pregnancy on women’s mental health.

The prevalence of mental disorders during the two-year period following pregnancy was found to be low in the current sample. However, a peak in estimated hazard rate for anxiety disorders was noted at the time of giving birth. Being married, employed, and the absence of chronic physical conditions (i.e., better physical health) were found to be protective factors for the mental health of women who have children. While the results offer some insight into the impact of pregnancy and child birth on women’s mental health in an Asian population, further studies based on a prospective or follow-up design may help to better understand these associations.