High levels of depressive symptoms and low quality of life are reported during pregnancy in Cape Coast, Ghana; a longitudinal study – BMC Public Health

Depressive symptoms during pregnancy

In line with our hypothesis, depressive symptoms were found to be highly prevalent throughout pregnancy in Cape Coast, Ghana; however, counter to our hypothesis, this prevalence estimate decreased during the course of pregnancy. Even so, the high prevalence of depressive symptoms in this population was of public health significance during all trimesters of pregnancy. In Ghana, like many developing countries, depressive symptoms are not usually assessed during pregnancy due to low priority of mental health, lack of mental health facilities, insufficient routine data collection on mental health and lack of mental health data for planning [44]. Additional cultural reasons for this neglect include the stigmatization of depression, leading women who are depressed to not seek psychiatric treatment due to fear of being labeled with psychosis [3]. Instead, many opt to seek treatment and counselling from traditional and religious healers, due to cultural acceptance [60]. As such, the magnitude of the problem is not realized and diagnosed in clinical settings. We have shown that a high prevalence estimate of depressive symptoms exist during pregnancy in this population, indicating that policies should be put in place to prioritize the assessment of depressive symptoms during pregnancy to avoid or diminish its effects on mothers and their fetus.

The prevalence estimates of depressive symptoms in our study are much higher than rates found in most studies conducted in developed countries. Schmied et al. [30] observed depressive symptom rates of 8.7% in Australia and New Zealand while Underwood et al. [61] observed 17% prevalence of antenatal depression across the entire pregnancy from a review involving twelve developed countries. Bennett et al. [16] observed rates of 7.4%, 12.8% and 12% in the 1st, 2nd and 3rd trimesters, respectively, in a review covering 21 developed countries. In Europe, prevalence rates of 12% and 14% were found in the 2nd and 3rd trimesters of pregnancy, respectively [62]. A study in rural US conducted during pregnancy observed rates similar to ours, 33% [63]. In developing countries, most reported rates are higher but still not as high as what we found. A review by Gelaye et al. [31] in low-income countries observed a pooled prevalence of 25%. Previous studies in Ghana found prevalence rates of 26% in the 3rd trimester of pregnancy [41, 42] a rate comparable to what we found in our 3rd trimester (29%) using the Western cut off (≥ 16). Our factor proportional cut off (≥ 14) however, gave a higher prevalence estimate of 34% in the 3rd trimester. In the few studies that have measured depressive symptoms longitudinally during pregnancy, some reported increasing rates [16, 30], but we found decreasing rates in our population. The previous longitudinal studies were conducted in developed countries and, even though the findings indicate increasing rates throughout pregnancy, the rates reported in the 3rd trimester are still not as high as rates reported during each of the trimesters in this study. While the use of different instruments and cutoffs might explain some of the differences, (most studies have used the Beck Depression Inventory (BDI) or the Edinburgh Postnatal Depression Scale (EPDS), while we used the CES-D), there is the possibility that the construct of depression is conceptualized differently in Ghana and there may be the need for better instruments that truly capture depressive symptoms in this setting. Despite the high prevalence of depressive symptoms observed in this population, depressive symptoms decreased with gestational age even after controlling for sociodemographic characteristics. Similar to previous findings, parity and food insecurity were significant predictors of depressive symptoms [19, 22]. This finding emphasizes the importance of addressing the issue of food insecurity not only to solve the physical needs of the pregnant woman but also the psychological needs.

Anxiety during pregnancy

Our findings suggest that estimated prevalence of anxiety symptoms is high in the 1st trimester but low in the 2nd and 3rd trimesters of pregnancy. A review by Schmied et al. [30] observed anxiety symptom rates of between 8 to 10% during the entire pregnancy. Our rates are much higher in the 1st trimester (34%) than rates found in most studies. Despite this, our prevalence rates during the 2nd and 3rd trimesters are comparable to other studies. A study in Nigeria found the prevalence of anxiety symptoms to be 13%, 4% and 6% in the 1st, 2nd and 3rd trimesters, respectively [17]. A study in Kumasi, Ghana, observed a higher prevalence rate of 11% in the 3rd trimester, compared to what we observed [41]. One reason for the discrepant findings could be the different instruments used to assess anxiety symptoms; our study used the BAI while the study in Kumasi used the 7-item Anxiety Scale (GAD-7).

We were surprised by the finding of a high estimated prevalence of anxiety in the 1st trimester but not the 2nd or 3rd trimesters. When we considered gestational age, we found total anxiety, fear, nervous, panic and somatic symptoms decreased with increasing gestational age. There may be cultural reasons that explain the prevalence of anxiety in this population. One might be the fear of pregnancy, especially during the first trimester. In Ghanaian settings, most women are first informed of their pregnancy status when they visit the clinic. For instance, a woman may present with symptoms that resemble malaria, and may have been treated for malaria over a period of time but the symptoms did not improve. She may then report to the clinic, only to be told that she is pregnant. Thus, the news of her pregnancy may come as a surprise. This may cause a woman to be anxious, especially during the first trimester [64,65,66]. Additionally, a woman might be concerned about her husband/partner accepting a new baby [67]. If the woman is not married, it poses multiple challenges including who this baby belongs to, whether the man responsible will accept it or not, and how the community will handle her pregnancy since being pregnant outside of marriage is frowned upon [68, 69]. Another cultural reason that may add to a woman’s anxiety is the fact that a woman in her first trimester will tend to hide her pregnancy and not share her news until she is visibly pregnant. This is due to the belief that if people get to know of her pregnancy she might be “bewitched” or “something bad will happen” and she might lose her baby [70]. This belief is borne out of the high rates of miscarriage observed during the first trimester of pregnancy [71]. Food insecurity may also be an issue. If a woman already has a child/children and she is not prepared for another pregnancy, this may be a cause of worry. In our study, we found the prevalence of food insecurity to be 50%, 30% and 25% for the 1st, 2nd and 3rd trimesters, respectively and food insecurity and gestational age were significant predictors of anxiety symptoms; one unit increase in food insecurity increased total anxiety symptoms by 0.619 units. Previous research indicates that food insecurity is an issue in this population and this can be a cause of anxiety and depression [72, 73]. As far as changes in the prevalence of anxiety over the course of pregnancy, once the news of the pregnancy is announced, if the husbands/partners and family members are happy with the pregnancy and are in support of the woman being pregnant, the woman’s worry, fear and panic may decrease, thus reducing anxiety as the pregnancy advances. This may account for the decreased prevalence estimate of anxiety symptoms seen during the 2nd trimester. By the 3rd trimester, generally, the Ghanaian family and society is happy to receive a new baby and this may lead even anxious women to become less anxious by the end of the pregnancy [16]. The high dropout rate observed between the 1st and 2nd trimesters could also account for a decrease in anxiety symptoms, assuming women who were anxious were those who dropped out. However, there were no significant differences in anxiety symptoms between women who dropped out and those who did not.

The instruments used to assess psychosocial health could also account for the low estimated prevalence of anxiety yet high depressive symptoms observed in the population. The BAI and CES-D, even though widely used by clinicians and researchers to determine anxiety and depressive symptoms, respectively, may not be as appropriate among pregnant women in the Ghanaian culture as they are in Western cultures. For instance, items may be interpreted differently among Ghanaian women than Western women. Even though the psychometrics were run to determine cultural appropriateness, we may have missed certain constructs that may describe anxiety or depressive symptoms in this population as factors that may determine anxiety and depression in the Ghanaian culture may be different from factors in Western populations. This research highlights the importance of developing valid cultural psychosocial measures that consider and understand how people from different cultures think about mental health and mental health problems. Research by De-Graft Aikins and Ofori-Atta [74] found symptoms of mental illness in Ghana to be characterized by excessive thinking, worry, persistent physical symptoms such as headaches, bodily pain, stresses arising from multiple responsibilities from family and work, and financial hardship. These symptoms are not listed on either the BAI or CES-D scales. It may be important to capture some of these factors in assessing anxiety or depression in the Ghanaian population.

QoL during pregnancy

We hypothesized that a significant number of the women would have low QoL and that this estimated prevalence would increase over time. QoL did change significantly over time with women in the 3rd trimester having a significantly lower mean QoL, lower physical health and lower role physical scores compared with the 1st and 2nd trimesters. Our study also found that GHV scores were lowest in pregnant women in their 1st trimester, and scores did not differ significantly between the 2nd and 3rd trimesters. Our results are similar to those of Chang et al. who found that pregnant Taiwanese women increased in GHV across trimesters with a significant difference between early and mid-pregnancy but no difference between mid and late pregnancy [33]. We found that pregnant women in their 3rd trimester had the lowest role emotional scores, but this did not differ from those during their 1st trimester. Our results agree with findings from two earlier studies which found that role physical scores decreased from early to late pregnancy but there was no difference between early to mid-pregnancy role physical scores [32, 33]. Chang et al. also found that role emotional scores were stable throughout pregnancy, which was similar to our results. Our findings may be related to common symptoms in early pregnancy such as feeling weak, low energy, nausea and vomiting [75] which are unwanted side effects experienced as a result of the hormonal changes that occur. These symptoms not only affect the physical health of pregnant women but can also negatively impact their psychological function [37]. In the 2nd and 3rd trimesters however, these symptoms may disappear, and women may gain more energy, thus improving GHV.

We found that gestational age was a negative predictor for total QoL, even after controlling for parity, marital status and food insecurity. Our findings agree with a study in France by Morin et al. [34], who assessed QoL at each month during pregnancy and found that QoL decreased significantly over time during pregnancy and decreased further between the 4th and 8th months. Similar to our study, Hueston and Kasik-Miller also found that physical health and role physical domains decreased with gestational age [32]. It is important to note that food insecurity was a significant predictor of total QoL and this emphasizes the need to address food insecurity particularly among pregnant women to improve QoL during pregnancy.

The main limitation of this study was sample size due to a dropout rate of 37.8% from the 1st to 2nd trimesters. Reasons such as miscarriages, husband refusing participation of their spouse in research, unanswered phone calls, phone switched off, and relocation accounted for the high dropout rate in this study. Another limitation could potentially be the new cut off developed for this population which must be validated in other studies, even though the prevalence estimate of psychosocial outcomes between the two cut offs was not different. Strengths of this study include the longitudinal nature which assessed psychosocial outcomes throughout pregnancy and including a population and culture which has been understudied. Another strength is the fact that we ran psychometric analyses on the psychosocial outcome scales and thus, we are confident of the results of our analyses.