Due to hormonal and anatomical changes and other factors including heavier care burdens, pregnant and postnatal women are more likely to suffer from psychological disturbances, particularly depression [25]. In this study, the prevalence of depression was 7.45% (95%CI: 5.87–9.04%) among pregnant and postpartum women, a rate that was lower than estimates from recent China-based meta-analyses [11, 12]. Notably, however, rates have varied between reviews. For example, an older meta-analysis [26] found that the pooled point prevalence of perinatal depression ranged from 6.5–12.9% from the start of pregnancy to the first postpartum year; 19.2% of the depressed subgroup reported depression during the first 3 months after delivery. Conversely, another recent meta-analysis found the overall prevalence of depression in pregnant and postnatal women was 11.9% (95% CI: 11.4%-12.5%) [5]. Discrepant rates between studies could be partly explained by different study samples, sampling methods, sample sizes, measurement tools for depression, socioeconomic backgrounds and clinical status [5, 6, 26].
Previous studies have indicated that depression is a strong predictor of suicidality [27–29]. In this study, a relatively small proportion of women reported suicidal ideation (2.5%), plan (0.38%) or attempt (0.47%) but the rate was substantially higher than those of other studies. For example, a large retrospective study from the US found that the prevalence of suicide attempts during pregnancy was 0.04% [30]. Another study from Canada reported that the prevalence of suicide attempts was 0.03% in pregnant women and 0.06% in postnatal women [31]. A study conducted in mainland China found that the prevalence of suicide attempt during pregnancy was 0.21% [32]. Reasons for the comparatively higher rate in this research are not known. However, due to differences in timeframes and measures of suicidality between studies, direct comparisons should be made with caution. The growing awareness of mental health concerns in pregnant and postnatal women, better social support networks, increased stigma related to disclosing distress, and accessibility to mental health services in major cities of China where the participating hospitals are located could contribute to the low overall prevalence of suicidality in this population. Nonetheless, we found that pregnant and postnatal women with depression were more likely to have suicidality compared with those without, which is consistent with previous findings [1, 33].
Additionally, our study found that participants with physical comorbidities were more likely to report depression, which also dovetails with previous findings [34, 35]. Physical comorbidities and adverse effects of treatments in pregnant and postnatal women were associated with more severe physical discomfort and impaired daily functioning, which could increase the risk of depression [34, 35].
Similar to previous findings [36], depression was significantly associated with lower QOL in all domains. The poorer QOL of depressed women could be explained by the distress/protection QOL model, in which QOL is determined by a range of protective and distressing factors [37]. QOL tends to be lower if distressing factors (e.g., frequent sleep disturbances, fatigue, and physical discomfort caused by depression) predominate overprotective factors (e.g., better social support from social networks). Depressed women often present with psychological and physical symptoms such as sadness, helplessness, cognitive impairments, body pain, insomnia, and digestive problems [38], all of which are related to lower QOL in depressed pregnant and postnatal women.
The merits of this study included its multicenter study design, large sample size, and use of a depression measure validated specifically for pregnant and postnatal depression. The main limitations should also be acknowledged. First, this was a cross-sectional study, so causal associations and changes between variables over time could not be examined. Second, all results were based on self-reported data; therefore, we cannot rule out biases in recall or social desirability as influences on the data. Third, to ensure response burdens were not excessive, variables that may be related to depression, such as social support, were not included in this study. Fourth, although a relatively large sample size and multicenter design based on different geographic regions of China were used, the needed number of eligible perinatal women was not calculated separately for each participating hospital. In addition, those with pre-existing psychiatric disorders and potential comprehension difficulties were excluded. Therefore, selection biases cannot be ruled out as influences on the data. To increase population representativeness, more costly national epidemiological surveys using multistage designs and random sampling should be conducted in China.