Background
The birth of a newborn brings happiness, a sense of fulfillment and euphoria in women. However, many women and infants die in the first six weeks after delivery [
1]. This period is also characterized by abrupt changes and stressful life events such as depression and anxiety [
2,
3]. Depression and anxiety are common and debilitating complications of postpartum women [
4,
5]. A meta-analysis conducted by O’Hara and Swain [
3] showed that around 13% of childbearing women suffer from postnatal depression (PND). Evidence suggests that undetected or untreated depression pose a risk of adverse health outcomes for mothers, their babies, and family members [
4,
6,
7]. Studies have suggested that PND has received considerable research attention while postnatal anxiety (PNA) has been relatively neglected [
2,
8]. It has been established that symptoms of depression and anxiety co-occur and that this comorbidity may be an indicator of the severity of psychological distress [
9].
However, evidence suggests that women also experience postnatal anxiety without depression. For example, a systematic review conducted on pre- and postnatal psychological wellbeing in African women living in Africa showed that the prevalence rate of postnatal anxiety was 14.0% [
10]. Studies have pointed out the importance of distinguishing between depression and anxiety in the postnatal period in order to provide appropriate interventions that specifically target the symptoms and etiology of anxiety [
4,
11,
12]. Furthermore, studies with non-postpartum populations have shown that comorbidity (simultaneous appearance of multiple illnesses) of anxiety and depression (i.e., anxious depression) manifests more severe symptoms [
13]. Co-occurrence of depressive and anxiety symptoms in postnatal period is more difficult to treat than each disorder alone [
14] and linked with adverse maternal and child health outcomes such as poor cognitive development and social-emotional development of the child [
4,
6], increases the risk for suicide, and requires specific treatment strategies for both sets of symptoms [
14,
15]. Miller et al. [
2] observed that the focus on depression can result in cases of anxiety (without depression) being undetected, and untreated. The authors further averred that in cases where depression and anxiety co-occur (anxious-depression), there is a risk that treatment strategies focus on the depressive symptoms, at the preclusion of specific treatments for the symptoms of anxiety.
Few studies have reported the prevalence of depressive and anxiety symptoms in postnatal women in Nigeria [
16‐
18]. For example, Uwakwe and Okonkwo [
16] reported depression in 10% of perinatal women in Eastern Nigeria, Adewuya et al. [
17] reported depression in 14.6% of postpartum women and 6.3% of general women population in Western Nigeria, while another study [
18] reported depression in 22.9% of postnatal women suffered in Enugu, Eastern Nigeria. The prevalence of PND in the later studies was higher than 13% of women reported by O’Hara and Swain [
4]. The higher prevalence of PND in Nigerian women may be attributed to many factors. Although maternal health service delivery in Nigeria has improved greatly in the past few decades, many challenges still undermine the quality and uptake of maternal health services by Nigerian women. For instance, poor use of maternal health services in Nigeria is a key factor contributing to high levels of maternal morbidity and mortality, as only 51.1% of women completed four or more antenatal care visits and only 36% of births took place in a health facility in 2013 [
19]. Poor perceived quality of care at facilities is a critical barrier [
20‐
22], and poor health worker attitudes contribute to a woman’s choice of using a facility or traditional provider [
22].
These apparent inadequacies of health facilities, maternal health service delivery and healthcare manpower in Nigeria may suggest that many cases of PND and PNA are undetected in women or when detected may be poorly treated. The connection between depression, anxiety symptoms and maternal factors has been explored in many studies, which have, however, produced inconsistent results [
8,
14,
23]. For instance, a study found 13 significant predictors of postpartum depression [
23]. Some maternal factors that have been linked with PND and PNA include self-esteem, childcare stress, life stress, social support, history/previous depression, infant temperament, maternity blues, socioeconomic status, and unplanned or unwanted pregnancy, advanced maternal age, low education, low monthly income, financial difficulties, poor family support, dissatisfaction in marital life, and poor marital relationship [
5,
8,
14,
23‐
26]. Only one Nigerian study [
27] has examined the prevalence of depressive and anxiety symptoms in postpartum women.
Discussion
This study aimed to determine depressive symptoms, anxiety and associated factors among postpartum women. In this study, the prevalence of PND in our sample was relatively high compared to previous findings in Nigeria. The prevalence of PND was higher than the previously reported prevalence of 13% in a meta-analysis [
4] 10% in Eastern Nigeria [
16], 14.6% in western Nigeria [
17], and 22.9% in Enugu metropolis [
18]. The similar findings consistently suggest that women suffer from depression during the postpartum period. The finding also corroborates the evidence that PND is very contextual and its prevalence varies between different cultures and communities. The higher prevalence of PND in our study may be attributed to the recent economic recession and challenges facing women, especially in the rural communities. Another reason for the higher prevalence of depressive symptoms in our sample may be attributed to the influx of women with traumatic experiences of communal or herdsmen-and-farmers-related conflicts who registered for ANC services in the study area. The finding suggests the need for protection of women from exposure to conflicts or traumatic experiences, context-specific mental health interventions (e.g., cognitive behavior therapy) for postpartum women to promote mental wellbeing, prevent maternal and child adverse health outcomes. Our study further shows that the prevalence of PND was higher at the 12 weeks postpartum. This probably indicates women’s prolonged exposure to stressful events during this period. Consistent with previous findings, the 12-week postpartum has been reported as one of the peak periods of PND [
4].
The point prevalence of anxiety symptoms in our sample was high. The prevalence of PNA was higher than the previously reported prevalence of 13% in Australian postpartum women [
2], 12.7% in Australian women [
3], and 13.1% in Arab postpartum women [
8]. This finding indicates that a considerable proportion of women suffers from anxiety during the postnatal period. The disparities in the prevalence of anxiety symptoms in postpartum women in the previous studies and present study may be due to variations in cultural settings, use of different screening tools for psychiatric morbidity and cut-off points in the identification of women with anxiety symptoms. For instance, Bener et al. [
8] investigated the prevalence of psychiatric morbidity in Arab women using DASS-21 with a cut-off ≥8 to identify women with anxiety symptoms. In contrast, we used HADS-A with a cut-off + 8 to identify women with anxiety symptoms. Regardless of the screening tools and cut-off used, identification of women with psychiatric morbidity in postpartum period requires proper and quality mental health intervention. The higher prevalence of anxiety symptoms in our sample may be due to the fear of newborn, displacement or detachment of some postpartum women from their support groups as a result of communal violence and the changing family structure in many Nigerian societies from extended to isolated nuclear families [
17], which deprives many young women of support from their family members especially older experienced women. This situation might have heightened anxiety levels in the women. In addition, our study shows that the point prevalence of anxiety symptoms was at its peak during the 4 weeks postpartum. This finding is consistent with the previous study who reported that 30% of new postpartum women were emotionally distressed [
51]. The higher prevalence of anxiety symptoms in this study is a call for postnatal care that is culturally sensitive, patient-centered, accessible and affordable by women, most importantly rural women. Since evidence suggests that poorer, less educated, and rural women have lower coverage of postnatal care [
1]. Furthermore, fragile family ties which have characterized modern family structures should be strengthened to galvanize support for postpartum women. Besides, stakeholders in the integrated maternal, newborn and child health (IMNCH) and the government at all levels should strengthen the health systems for improved quality of care for mothers and babies.
Interestingly in this study, a considerable proportion of women were
anxious-depressed. A lower prevalence of 7% was found in a study [
2] that used DASS-21 to assess comorbidity of anxiety and depressive symptoms in postpartum women. The higher point prevalence of comorbidity of anxiety and depressive symptoms in our sample may be attributed to a myriad of factors such as prevailing economic situation, poor quality of maternal health services, mistreatment and disrespectful care by health workers at the health facility, exposure to traumatic experiences such as communal violence [
19‐
22]. Evidence showed that women in low-and-middle-income countries (LMICs) such as Nigeria lack access to quality and adequate postnatal care coverage [
1,
52,
53].
In our study, the following factors were associated with the occurrence of depressive symptoms in women: maternal age, and health facility type. No significant association was found between depressive symptoms and a history of PND symptoms. However, in this study, about half of the women had a history of depression. This finding contradicts existing evidence which shows that a personal history of depression is a strong risk factor for developing depressive symptoms during pregnancy and postnatal period [
2‐
5,
8,
14,
23,
54]. The non-existence of association between a personal history of depression and PND in our sample may be due to some contextual factors such as cultural influence, study setting and method adopted. For instance, some cultural factors such as domestic violence (e.g., wife beating) and male child preference might have heightened anxiety symptoms in women especially women who with female babies whose husbands or family members expected to have male children. Such women received little or no family support and care from their husbands and family members. This finding contradicts the finding of a previous study which reported that women with a personal history of a major depressive episode are known to be at a 30% higher risk of developing PND [
23].
In the bivariate analysis, mothers aged 45 years and above were 2 times more likely to develop depressive symptoms as compared to mothers aged 15–29 years and attendance at a secondary health facility was significantly associated with PND in women. Maternal level of education was not significant. The finding is inconsistent with the finding of a study [
8] which reported more depressive symptoms in women under 30 years old. The association between health facility attended during pregnancy and postnatal period and development of PND probably be due to maltreatment of women by health workers during childbirth [
55]. This finding is consistent with studies that have identified associations between sociodemographic factors and PND in women [
23,
52,
56]. Evidence showed that being a younger mother is a risk factor for the development of PND [
8] and attendance at a health facility characterized by the maltreatment of women by health workers during childbirth is a risk factor for adverse maternal outcomes [
55]. Therefore, our interaction terms result showed that being a mother aged 30–44 years and attendance at a secondary health facility characterized by maltreatment of women by health workers during childbirth or postnatal period had a profound significant effect on the development of depressive symptoms among mothers. This finding advocates the importance of providing age-specific mental health interventions for women during prenatal and postnatal periods in Nigeria. This may reduce the prevalence of PND among women especially younger women. Furthermore, implementation and enforcement of valid measures targeted at preventing mistreatment/abuse of women during the childbirth and postnatal period by health workers should be encouraged at all the levels of health facility in Nigeria. This will reduce to very large extent cases of maternal mistreatment and in turn prevent or reduce significantly associated adverse outcomes.
The logistic regression analysis showed that being a mother under 44 years old, and having monthly income higher than #20,000.00 were significantly associated with PNA. Mothers with monthly income ≥ #20,000.00 were 3 times likely to develop anxiety symptoms as compared to mothers with monthly income < #10,000.00. Many studies have reported associations between low socioeconomic factors, anxiety and depressive symptoms in postnatal women [
5,
8,
23,
52,
57,
58]. For instance, Segre et al. [
58] reported that women who earned the higher income ($70,000 annually) were at 4 times lower risk of developing PND than with women with low incomes ($10,000 annually). In Nigeria, women with #10,000 or #20,000 monthly income, earn #120,000 or #240,000 annually respectively. This is equal to $333 or $667 annually respectively (current exchange rate = #360/$1). Consequently, women with low incomes experience hardship in raising their infants in Nigeria. Poor or limited financial resources might have increased anxiety symptoms in postnatal women. The gloomy economic reality compounded this situation and portends danger for women with low incomes who cannot afford quality postnatal care. This poses serious threats to maternal and child survival in Nigeria. Although not significant, mothers with primary education were 2 times more likely to develop anxiety symptoms compared to mothers with NFE. This finding was also observed in previous studies that low level of education is a strong risk factor for PND in women [
58‐
60].
In this study, we found a significant association between attendance at postnatal care clinics in a secondary health facility and co-occurrence of anxiety and depressive symptoms in our sample.
Studies have reported significant association between low socioeconomic factors and psychiatric conditions in women population [
23,
52,
57,
58], however, there seems to be a dearth of studies on relationship between the health facility type attended during perinatal or postnatal period and comorbid depressive and anxiety symptomatology in Nigerian postpartum women. This association may be attributed to maltreatment of women by health workers during childbirth and barriers to quality maternal care at the secondary health facilities especially in Nigerian rural communities [
20‐
22,
55]. More research is needed to examine their relationship to one another in Nigerian context. Such research could substantially improve our understanding of this relationship with the aim of improving the conditions of maternal care services across the levels of health facilities in Nigeria. Having no history of depression was significantly associated with the comorbidity of anxiety and depressive symptoms. This finding defies existing evidence in our sample. However, the observed association may be due to the interplay of other factors not considered in this study. This finding disagrees with the previous studies that reported an association between previous psychiatric history and the risk of developing depressive symptoms during pregnancy and postnatal period [
2‐
5,
8,
14,
23,
54]. Our interaction terms result showed that having a secondary education by a monthly income within #10,000 - #20,000 was significantly associated with the comorbid depressive and anxiety symptomatology in postpartum women. This is consistent with previous studies suggesting that low socioeconomic factors constitute strong risk factors for psychiatric conditions in women [
23,
55,
57‐
60]. Given the attendant adverse effects of comorbid postpartum depressive and anxiety symptoms on maternal health and infant development, health care workers should intensity their efforts at identifying this sub-category of women, mitigating the risk factors and providing prompt care, referral, and/or treatment at subsidized or affordable costs.
In the MLR, attendance at postnatal care services in PHCs (AOR = 0.32,
p = .05) was significantly associated with anxiety symptoms. This finding is consistent with other studies [
8,
23,
52] who reported an association between mothers’ sociodemographic factors and anxiety disorders. A mother’s level of education and place of residence were not significantly associated with anxiety symptoms. This finding is inconsistent with other studies [
8,
23,
52] who reported an association between mothers’ sociodemographic factors and anxiety disorders. Although not significant, mother’s income level, and being a younger mother were more likely to predict depressive symptoms in the studied postnatal women. This finding is concordance with other studies that reported the association between low income level, younger mother’s age and depressive symptoms [
3,
5,
8,
52]. We found no association between a mother’s level of education, health facility type attended, place of residence and depressive symptoms. This finding is inconsistent with other studies [
5,
8,
23,
52] who reported an association between mothers’ sociodemographic factors and depressive symptoms. However, this finding is consistent with Fiala et al. [
54], who reported no significant association between PND and maternal education.
We found a significant association between attendance at postnatal care services in PHCs, and co-occurrence of anxiety and depressive symptoms in our sample. Although not significant, a mother’s income level and parity have the potential to increase the likelihood of anxious-depressed. In the same way, women aged 15–29 years were 2 times more likely to develop comorbidity of anxiety and depressive symptoms compared to older women. Similarly, women with primary education were 4 times more likely to develop comorbidity of anxiety and depressive symptoms as compared with women with tertiary education. Also, mothers with NFE and secondary education were 2 times more likely to develop comorbidity of anxiety and depressive symptoms compared with women with tertiary education. These findings are consistent with other studies [
23,
48,
57‐
59]. This finding further corroborates the assertion of Miller et al. [
2] that this sub-group of women should be identified and given appropriate interventions. However, such intervention should be context specific, culturally-sensitive, patient-centered, accessible and affordable. Provision of incentives such as free postnatal care, free transport, voucher schemes, elimination of mistreatment and disrespect [
1,
61] have been identified to be effective in facilitating equity of access to quality postnatal care [
1].
In the model for interaction terms, we found a significant association between having a secondary education by attendance at a secondary health facility and the development of depressive symptoms in women. We also found a significant association between having a primary education by attendance at primary health centers (PHCs) and the development of comorbid anxiety and depressive symptomatology in women. This finding is consistent with previous studies who reported a significant association between psychiatry disorders and low socioeconomic status in women [
23,
57‐
60]. This association may be attributed to low literacy among rural women, mistreatment or abuse of women by health workers during childbirth and barriers to quality maternal care at both the PHCs and secondary health facilities in Nigeria [
20‐
22,
55]. More research is needed to examine their relationship to one another in the Nigerian context. Such research could substantially improve our understanding of this relationship with the aim of improving the literacy level of women, fostering positive attitudes towards women by healthcare workers, scaling up conditions of maternal care services across the levels of health facilities in Nigeria. In addition, having a monthly income of < #10,000 or #10,000 - #20,000 or ≥ #20,000 and a history of depression are protective factors against the development of depressive symptoms in women. This finding is inconsistent with previous studies [
2‐
5,
8,
14,
23,
62]. The plausible explanation for this finding could be that women with previous history of depression regardless of income level might have developed resilience to the onset of new depressive symptoms, especially if properly treated using effective interventions [
62]. Furthermore, in Nigerian cultural context, older women in the family and members of women’s associations either in the religious or traditional settings provide support for postnatal women with psychiatric disorders. For instance, such measures may include offering counseling and prayer sessions, provision of material and financial incentives for the affected women. These measures strengthen the capacity of women to cope effectively with psychiatric disorders and enhance their resilience traits. This could explain the null association between income, history of depression and the development of depressive symptoms in our sample.
Strength and limitation
The conclusion of this study was based on primary data via rigorous descriptive and analytic data analysis. Furthermore, the majority of our findings are consistent with those available in the literature. This is a correlational study and only provides information on significant associations between sociodemographic factors and depressive and anxiety symptoms in women, and it cannot be used to infer causality of these events, which would require clinical trials and longitudinal studies. The higher prevalence of depressive and anxiety symptoms and their comorbidity in the studied postnatal women could be attributed to other variables not examined in this study. For instance, variables such as unintentional pregnancy, living alone/single mother, feelings of sadness about being pregnant, prenatal depression and anxiety, breastfeeding, occupation, parental consanguinity and number of people living in the home have been identified as risk factors for both postnatal depressive and anxiety symptoms [
5,
8,
23,
54]. Another limitation is that the study used both the EPDS and HADS-A scales to detect depressive and anxiety symptoms in women. The EPDS and HADS-A are suboptimal tools for detecting clinically significant depressive syndrome and they cannot replace a systematic Clinical Interview Schedule. The EPDS and HADS contain several items that are non-specific for depression. For instance, both tools can be used to detect anxiety symptoms in clinical and non-clinical settings [
31‐
33,
37‐
41,
49,
50]. Also, there is a general lack of agreement on the use of specific cut-off for depression detection on the EPDS. However, there seems to be a consensus on the use of cut-off point + 8 on HADS-A for identifying women at risk for anxiety symptoms. The original studies recommend using an optimal cut-off of 10 points or higher on the EPDS [
35] and an optimal cut-off point + 8 on the HADS-A based on ICD-9 and DSM-III [
38,
45,
46], which are still adopted by many researchers. Some authors recommended that an EPDS optimal cut-off of 12 points or higher is an accepted cut-off for recognizing patients at risk of PND [
35,
40]. Some authors consider a threshold of 13 points [
40,
63] while other studies recommended a lower cut-off of nine points [
2]. In this study, we used the EPDS with an optimal cut-off point of ≥13 [
40,
63] and an optimal cut-off point of + 8 on the HADS-A [
42,
43]. The Hosmer-Lemeshow (HL) test
p-values for the models on associated factors of anxiety symptoms in this study appear to be low. However, the p-values (factors associated with PNA,
p = .063; PNA interaction terms,
p = .217) are greater than 0.05, which are indicative of good fit. However, the HL test has some limitations. For instance, HL test does not account over-fitting of the model and have a tendency to have low power. In addition, HL test offers very little direction on how to select the number of subgroups. The selection of the groups (g) is subject to the researcher’s discretion. Small values for g give the HL test less chance to detect mis-specifications. Larger values imply that the number of items in each subgroup may be too small to find differences between observed and expected values. Thus, in many cases, the selection for g is often unclear and may be arbitrary. Also, we used a small sample in this study, and evidence suggests that in a small sample, a high
p-value from the HL test may simply be due to the test having lower power to detect mis-specification, rather than being indicative of good fit [
64]. We can reasonably conclude that our models demonstrated a good fit for the data used in this study. Another limitation of this study is data collection bias or interviewer bias. Interviewers were briefed on data collection to assure consistency within and between observers. This approach may not be enough to avoid interviewer bias. However, to minimize the interviewer bias, we used well validated scales for data collection. Also, interviewers are experts who have acquired quality training on data collection process. Thus, we believe these measures can minimize or mitigate the inter-observer variability when many personnel are involved in data collection and entry [
65]. Also, in this study, data were collected retrospectively, this might introduce recall bias. To mitigate the recall bias effect, women who gave a live birth in 2017 only were recruited for the study and well validated scales were used for data collection. Our analyses were adjusted for PND and PNA potential predictors such as domestic violence, birth type, family support and parity, however, there were no relationships between the factors and occurrence of PND and PND in the postnatal women. Besides, the self-reported nature of the EPDS, HADS-A and WHO’s DVE questionnaire permit response bias. This may be of particular concern for sensitive items especially in the EPDS, and WHO’s DVE questionnaire, which may lead to underreporting and therefore, may likely bias the estimated associations to null. The small sample size is another limitation of this study. Single population proportion formula was used to derive optimum sample size for this study. However, sample size calculated through single population formula was considered adequate enough to identify associated factors of depressive and anxiety symptoms in women. In this study we conducted a post hoc power analysis and obtained a power (1-β err prob) of 0.9999 which showed that the maximum permissible sample size was used for the study. Thus, we concluded that the power is sufficient for the study. Although, Schulz and Grimes [
66] argued about the irrelevance of post hoc power calculations, perhaps, not calculating the statistical power at all seems to be undesirable. Our study sample consisted of only the postpartum women visiting the health facilities for postnatal care services. The sample does not include postpartum women using alternative postnatal care services provided by traditional births attendants (TBAs) and traditional healers, thus limiting the generalizability of the findings. Future studies should consider using a larger, randomized and more representative sample size.